Changeset - b0801a2b8a2f
[Not reviewed]
0 3 0
Bradley Kuhn (bkuhn) - 9 years ago 2014-12-20 00:09:01
bkuhn@ebb.org
Clarify sentence.

Make this sentence a bit clearer.
3 files changed with 3 insertions and 3 deletions:
0 comments (0 inline, 0 general)
compliance-guide.tex
Show inline comments
 
% compliance-guide.tex                            -*- LaTeX -*-
 

	
 
\part{A Practical Guide to GPL Compliance}
 
\label{gpl-compliance-guide}
 

	
 
{\parindent 0in
 
This part is: \\
 
\begin{tabbing}
 
Copyright \= \copyright{} 2008, 2014 \= \hspace{.2in} Bradley M. Kuhn. \\
 
Copyright \= \copyright{} 2014 \> \hspace{.2in} Free Software Foundation, Inc. \\
 
Copyright \> \copyright{} 2008, 2014 \> \hspace{.2in} Software Freedom Law Center. \\
 
\end{tabbing}
 

	
 
\vspace{.1in}
 

	
 
\begin{center}
 
The copyright holders of this part hereby grant the freedom to copy, modify,
 
convey, Adapt, and/or redistribute this work under the terms of the Creative
 
Commons Attribution Share Alike 4.0 International License.  A copy of that
 
license is available at
 
\url{https://creativecommons.org/licenses/by-sa/4.0/legalcode}.
 

	
 
\vfill
 

	
 
This part includes material from many sources, including the following
 
This part includes material from many sources, including some material from the following
 
CC-By-SA-licensed published works: \\
 

	
 
\begin{itemize}
 
\item \hrefnofollow{http://www.softwarefreedom.org/resources/2008/compliance-guide.html}{\textit{The Practical Guide GPL Compliance}}, by Bradley M. Kuhn, Aaron
 
Williamson and Karen Sandler, first published on 2008-08-20. \\
 
\item \hrefnofollow{http://www.softwarefreedom.org/resources/2014/SFLC-Guide_to_GPL_Compliance_2d_ed.html}{\textit{Software Freedom Law Center Guide to GPL Compliance, 2nd
 
  Edition}} by Eben Moglen and Mishi Choudhary, first published on 2014-10-31. \\
 
\end{itemize}
 

	
 
However, this work is primarily composed of the many contributions it
 
receives as a public, collaborative project.  Please
 
\href{https://gitorious.org/copyleft-org/tutorial/history/master:compliance-guide.tex}{review
 
  its Git logs} for full documentation of all contributions.
 

	
 
\end{center}
 
}
 

	
 
\pagebreak
 

	
 
\chapter*{Executive Summary}
 

	
 
This is a guide to effective compliance with the GNU General Public
 
License (GPL) and related licenses.  Copyleft advocates
 
usually seek to assist the community with
 
GPL compliance cooperatively.   This guide focuses on complying from the
 
start, so that readers can learn to avoid enforcement actions entirely, or, at
 
least, minimize  the negative impact when enforcement actions occur.
 
This guide  introduces and explains basic legal concepts related to the GPL and its
 
enforcement by copyright holders. It also outlines business practices and
 
methods that lead to better GPL compliance.  Finally, it recommends proper
 
post-violation responses to the concerns of copyright holders.
 

	
 
\chapter{Background}
 

	
 
Copyright law grants exclusive rights to authors.  Authors who chose copyleft
 
seek to protect the freedom of users and developers to copy, share, modify
 
and redistribute the software.  However, copyleft is ultimately implemented
 
through copyright, and the GPL is primarily and by default a copyright
 
license.  (See \S~\ref{explaining-copyright} for more about the interaction
 
between copyright and copyleft.)  Copyright law grants an unnatural exclusive
 
control to copyright holders regarding copyright-controlled permissions
 
related to the work.  Therefore, copyright holders (or their agents) are the
 
ultimately the sole authorities to enforce copyleft and protect the rights of
 
users.  Actions for copyright infringement are the ultimate legal mechanism
 
for enforcement.  Therefore, copyright holders, or collaborative groups of
 
copyright holders, have historically been the actors in GPL enforcement.
 

	
 
The earliest of these efforts began soon after the GPL was written by
 
Richard M.~Stallman (RMS) in 1989, and consisted of informal community efforts,
 
often in public Usenet discussions.\footnote{One example is the public
 
  outcry over NeXT's attempt to make the Objective-C front-end to GCC
 
  proprietary.  RMS, in fact, handled this enforcement action personally and
 
  the Objective-C front-end is still part of upstream GCC today.}  Over the next decade, the Free Software Foundation (FSF),
 
which holds copyrights in many GNU programs, was the only visible entity
 
actively enforcing its GPL'd copyrights on behalf of the software freedom
 
community.
 
FSF's enforcement
 
was generally a private process; the FSF contacted violators
 
confidentially and helped them to comply with the license.  Most
 
violations were pursued this way until the early 2000's.
 

	
 
By that time, Linux-based systems such as GNU/Linux and BusyBox/Linux had become very common, particularly in
 
embedded devices such as wireless routers.  During this period, public
 
ridicule of violators in the press and on Internet fora supplemented
 
ongoing private enforcement and increased pressure on businesses to
 
comply.  In 2003, the FSF formalized its efforts into the GPL Compliance
 
Lab, increased the volume of enforcement, and built community coalitions
 
to encourage copyright holders to together settle amicably with violators.
 
Beginning in 2004, Harald Welte took a more organized public enforcement
 
approach and launched \href{http://gpl-violations.org/}{gpl-violations.org}, a website and mailing
 
list for collecting reports of GPL violations.  On the basis of these
 
reports, Welte successfully pursued many enforcement actions in Europe, including
 
formal legal action.  Harald earns the permanent fame as the first copyright
 
holder to bring legal action in a court regarding GPL compliance.
 

	
 
In 2007, two copyright holders in BusyBox, in conjunction with the
 
Software Freedom Conservancy (``Conservancy''), filed the first copyright infringement lawsuit
 
based on a violation of the GPL\@ in the USA. While  lawsuits are of course
 
quite public, the vast majority of Conservancy's enforcement actions 
 
are resolved privately via
 
cooperative communications with violators.  As both FSF and Conservancy have worked to bring
 
individual companies into compliance, both organizations have encountered numerous
 
violations resulting from preventable problems such as inadequate
 
attention to licensing of upstream software, misconceptions about the
 
GPL's terms, and poor communication between software developers and their
 
management.  This document highlights these problems and describe
 
best practices to encourage corporate Free Software users to reevaluate their
 
approach to GPL'd software and avoid future violations.
 

	
 
Both FSF and Conservancy continue GPL enforcement and compliance efforts
 
for software under the GPL, the GNU Lesser
 
Public License (LGPL) and other copyleft licenses.  In doing so, both organizations have
 
found that most violations stem from a few common, avoidable mistakes.  All copyleft advocates  hope to educate the community of
 
commercial distributors, redistributors, and resellers on how to avoid
 
violations in the first place, and to respond adequately and appropriately
 
when a violation occurs.
 

	
 
\section{Who Has Compliance Obligations?}
 

	
 
All distributors of modified or unmodified versions of copylefted works
 
unmodified versions of the works have compliance obligations.  Common methods
 
of modifying the works include innumerable common acts, such as:
 

	
 
\begin{itemize}
 

	
 
  \item embedding those works as executable copies
 
    into a device,
 

	
 
  \item transferring a digital copy of executable copies to someone else,
 

	
 
  \item posting a patch to the copylefted software to a public mailing list.
 

	
 
\end{itemize}
 

	
 
Such distributors have obligations to (at least) the users to whom they (or
 
intermediary parties) distribute those copies.  In some cases, distributors
 
have obligations to third parties not directly receiving their distribution
 
of the works (depending on the distributors chosen licensing options, as
 
described later in \S~\ref{binary-distribution-permission}).  In addition,
 
distributors have compliance obligations to upstream parties, such as
 
preservation of reasonable legal notices embedded in the code, and
 
appropriate labeling of modified versions.
 

	
 
Online service providers and distributors alike have other compliance
 
obligations.  In general, they must refrain from imposing any additional
 
restrictions on downstream parties. Most typically, such compliance problems
 
arise from ``umbrella licenses:'' EULAs, or sublicenses that restrict
 
downstream users' rights under copyleft. (See \S~\ref{GPLv2s6} and
 
\S~\ref{GPLv3s10}).
 

	
 
Patent holders having claims reading on GPL'd works they distribute must
 
refrain from enforcing those claims against parties to whom they distribute.
 
Furthermore, patent holders holding copyrights on GPLv3'd works must further
 
grant an explicit patent license for any patent claims reading on the version
 
they distributed, and therefore cannot enforce those specific patent claims
 
against anyone making, using or selling a work based on their distributed
 
version.  All parties must refrain from acting as a provider of services or
 
distributor of licensed works if they have accepted, or had imposed on them
 
by judicial action, any legal conditions that would prevent them from meeting
 
any obligation under GPL\@.  (See \S~\ref{GPLv2s7}, \S~\ref{GPLv3s11} and
 
\S~\ref{GPLv3s12}.
 

	
 
\section{What Are The Risks of Non-Compliance?}
 

	
 
Copyleft experts have for decades observed a significant mismatch between the
 
assumptions most businesses make about copyleft compliance and the realities.
 
Possibly due to excessive marketing of proprietary tools and services from
 
the for-profit compliance industry, businesses perennially focus on the wrong
 
concerns.  This tutorial seeks to educate those businesses about what
 
actually goes wrong, what causes disputes, and how to resolve those disputes.
 

	
 
Many businesses currently invest undue resources to avoid unlikely risks that
 
have low historical incidence of occurrence and low cost of remediation,
 
while leaving unmanaged the risks that have historically resulted in all the
 
litigation and other adverse outcomes.  For example, some ``compliance
 
industry''\footnote{``Compliance industry'' refers to third-party for-profit
 
  companies that market proprietary software tools and/or consulting services
 
  that purport to aid businesses with their Free Software license compliance
 
  obligations, such as those found in GPL and other copyleft licenses.  This
 
  tutorial leaves the term in quotes throughout, primarily to communicate the
 
  skepticism most of this tutorial's authors feel regarding the mere
 
  existence of this industry.  Not only do copyleft advocates object on
 
  principle to proprietary software tools in general, and to their ironic use
 
  specifically to comply with copyleft, but also to the ``compliance
 
  industry'' vendors' marketing messaging, which some copyleft advocates
 
  claim as a cause in the risk misassessments discussed herein.  Bradley
 
  M.~Kuhn, specifically, regularly uses the term ``compliance industrial
 
  complex''
 
  \href{http://en.wikipedia.org/wiki/Military-industrial_complex}{to
 
    analogize the types of problems in this industry to those warned against
 
    in the phrase of origin}.} vendors insist that great effort must be
 
expended to carefully list, in the menus or manuals of embedded electronics
 
products, copyright notices for every last copyright holder that contributed
 
to the Free Software included in the product.  While nearly all Free Software
 
licenses, including copylefts like GPL, require preservation and display of
 
copyright notices, failure to meet this specific requirement is trivially
 
remedied.  Therefore, businesses should spend just reasonable efforts to
 
properly display copyright notices, and note that failure to do so is simply
 
remedied: add the missing copyright notice!
 

	
 
\section{Understanding Who's Enforcing}
 
\label{compliance-understanding-whos-enforcing}
 

	
 
The mismatch between actual compliance risk and compliance risk management
 
typically results from a misunderstanding of licensor intentions.  For-profit
 
businesses often err by assuming other actors have kindred motivations.  The
 
primary enforcers of the GPL, however, have goals that for-profit businesses
 
will find strange and perhaps downright alien.
 

	
 
Specifically, community-oriented GPL enforcement organizations (called
 
``COGEOs'' throughout the remainder of this tutorial) are typically
 
non-profit charities (such as the FSF and Software Freedom Conservancy) who
 
declare, as part of their charitable mission, advancement of software freedom
 
for all users.  In the USA, these COGEOs are all classified as charitable
 
under the IRS's 501(c)(3) designation, which is reserved for organizations
 
that have a mission to enhance the public good.
 

	
 
As such, these COGEOs enforce GPL primarily to pursue the policy goals and
 
motivations discussed throughout this tutorial: to spread software freedom
 
further.  As such, COGEOs are unified in their primary goal to bring the
 
violator back into compliance as quickly as possible, and redress the damage
 
caused by the violation.  COGEOs are steadfast in their position in a
 
violation negotiation: comply with the license and respect freedom.
 

	
 
Certainly, other entities do not share the full ethos of software freedom as
 
institutionalized by COGEOs, and those entities pursue GPL violations
 
differently.  Oracle, a company that produces the GPL'd MySQL database, upon
 
discovering GPL violations typically negotiates a proprietary software
 
license separately for a fee.  While this practice is not one a COGEO would
 
undertaking nor endorsing, a copyleft license technically permits this
 
behavior.  To put a finer point on this practice already discussed
 
in~\S~\ref{Proprietary Relicensing}, copyleft advocates usually find copyleft
 
enforcement efforts focused on extract alternative proprietary licenses
 
distasteful at best, and a corrupt manipulation of copyleft at worst.  Much
 
to the advocates' chagrin, such for-profit enforcement efforts seem to
 
increase rather than decrease.
 

	
 
Thus, unsurprisingly, for-profit adopters of GPL'd software often incorrectly
 
assume that all copyright holders seek royalties.  Businesses therefore focus
 
on the risk of so-called ``accidental'' (typically as the result of
 
unsupervised activity by individual programmers) infringe copyright by
 
incorporating ``snippets'' of copylefted code into their own proprietary
 
computer program.  ``Compliance industry'' flagship products, therefore,
 
focus on ``code scanning'' services that purport to detect accidental
 
inclusions.  Such effort focuses on proprietary software development and view
 
Free Software as a foreign interloper.  Such approach not only ignores
 
current reality that many companies build their products directly on major
 
copylefted projects (e.g., Android vendor's use of the kernel named Linux),
 
but also creates a culture of fear among developers, leading them into a
 
downward spiral of further hiding their necessary reliance on copylefted
 
software in the company's products.
 

	
 
Fortunately, COGEOs regard GPL compliance failures as an opportunity to
 
improve compliance.  Every compliance failure downstream represents a loss of
 
rights by their users. The COGEOs are the guardian of its users' and
 
developers' rights.  Their activity seeks to restore those rights, and
 
to protect the project's contributors' intentions in the making of their
 
software. 
 

	
 
\chapter{Best Practices to Avoid Common Violations}
 
\label{best-practices}
 

	
 
Unlike highly permissive licenses (such as the ISC license), which
 
typically only require preservation of copyright notices, licensees face many
 
important requirements from the GPL.  These requirements are
 
carefully designed to uphold certain values and standards of the software
 
freedom community.  While the GPL's requirements may initially appear
 
counter-intuitive to those more familiar with proprietary software
 
licenses, by comparison, its terms are in fact clear and quite favorable to
 
licensees.  Indeed, the GPL's terms actually simplify compliance when
 
violations occur.
 

	
 
GPL violations occur (or, are compounded) most often when companies lack sound
 
practices for the incorporation of GPL'd components into their
 
internal development environment.  This section introduces some best
 
practices for software tool selection, integration and distribution,
 
inspired by and congruent with software freedom methodologies.  Companies should
 
establish such practices before building a product based on GPL'd
 
software.\footnote{This document addresses compliance with GPLv2,
 
  GPLv3, LGPLv2, and LGPLv3.  Advice on avoiding the most common
 
  errors differs little for compliance with these four licenses.
 
  \S~\ref{lgpl} discusses the key differences between GPL and LGPL
 
  compliance.}
 

	
 
\section{Evaluate License Applicability}
 
\label{derivative-works}
 
Political discussion about the GPL often centers around determining the
 
``work'' that must be licensed under GPL, or in other words, ``what is the
 
derivative and/or combined work that was created''.  Nearly ever esoteric
 
question asked by lawyers seek to consider that question
 
\footnote{\tutorialpartsplit{In fact, a companion work, \textit{Detailed Analysis of the GNU GPL and Related
 
      Licenses} contains an entire section discussing derivative works}{This tutorial in fact
 
  also addresses the issue at length in~\S~\ref{derivative-works}}.} (perhaps because
 
that question explores exciting legal issues while the majority of the GPL
 
deals with much more mundane ones).
 
Of course, GPL was designed
 
primarily to embody the licensing feature of copyleft.
 

	
 
However, most companies who add
 
complex features to and make combinations with GPL'd software
 
are already well aware of their
 
more complex obligations under the license that require complex legal
 
analysis.  And, there are few companies overall that engage in such
 
activities. Thus,  in practical reality, this issue is not relevant to the vast
 
majority of companies distributing GPL'd software.
 

	
 
Thus, experienced  GPL enforcers find that few redistributors'
 
compliance challenges relate directly to combined work issues in copyleft.
 
Instead, the distributions of GPL'd
 
systems most often encountered typically consist of a full operating system
 
including components under the GPL (e.g., Linux, BusyBox) and components
 
under the LGPL (e.g., the GNU C Library).  Sometimes, these programs have
 
been patched or slightly improved by direct modification of their sources,
 
and thus the result is unequivocally a modified version.  Alongside these programs,
 
companies often distribute fully independent, proprietary programs,
 
developed from scratch, which are designed to run on the Free Software operating
 
system but do not combine with, link to, modify, derive from, or otherwise
 
create a combined work with
 
the GPL'd components.\footnote{However, these programs do often combine
 
  with LGPL'd libraries. This is discussed in detail in \S~\ref{lgpl}.}
 
In the latter case, where the work is unquestionably a separate work of
 
creative expression, no copyleft provisions are invoked.
 
The core compliance issue faced, thus, in such a situation, is not an discussion of what is or is not a
 
combined, derivative, and/or modified version of the work, but rather, issues related to distribution and
 
conveyance of binary works based on GPL'd source, but without Complete,
 
Corresponding Source.
 

	
 
As such, issues of software delivery are the primary frustration for GPL
 
enforcers. In particular, the following short list accounts for at least 95\%
 
of the GPL violations ever encountered:
 

	
 
\begin{itemize}
 

	
 
\item The violator fails to provide required information about the presence
 
  of copylefted programs and their applicable license terms in the product
 
  they have purchased.
 

	
 
\item The violator fails to reliably deliver \hyperref[CCS
 
  Definition]{complete, corresponding source} (CCS) for copylefted programs
 
  the violator knew were included (i.e., the CCS is either delivered but
 
  incomplete, or is not delivered at all).
 

	
 
\item Requestors are ignored when they communicate with violator's published
 
  addresses requesting fulfillment of businesses' obligations.
 
\end{itemize}
 

	
 
This tutorial therefore focuses primarily on these issue.
 
Admittedly, a tiny
 
minority of compliance situations relate to question of derivative,
 
combined, or modified versions of the work.  Those
 
situations are so rare, and the details from situation to situation differ
 
greatly.  Thus, such situations require a highly
 
fact-dependent analysis and cannot be addressed in a general-purpose
 
document such as this one.
 

	
 
\medskip
 

	
 
Most companies accused of violations lack a basic understanding
 
of how to comply even in the straightforward scenario.  This document
 
provides those companies with the fundamental and generally applicable prerequisite knowledge.
 
For answers to rarer and more complicated legal questions, such as whether
 
your software is a derivative or combined work of some copylefted software, consult
 
with an attorney.\footnote{If you would like more information on the
 
  application of derivative works doctrine to software, a detailed legal
 
  discussion is presented in our colleague Dan Ravicher's article,
 
  \textit{Software Derivative Work: A Circuit Dependent Determination} and in
 
  \tutorialpartsplit{\textit{Detailed Analysis of the GNU GPL and Related
 
      Licenses}'s Section on derivative works}{\S~\ref{derivative-works} of
 
    this tutorial}.}
 

	
 
This discussion thus assumes that you have already identified the
 
``work'' covered by the license, and that any components not under the GPL
 
(e.g., applications written entirely by your developers that merely happen
 
to run on a Linux-based operating system) distributed in conjunction with
 
those works are separate works within the meaning of copyright law and the GPL\@.  In
 
such a case, the GPL requires you to provide complete corresponding
 
source (CCS)\footnote{For more on CCS,  see
 
\tutorialpartsplit{\textit{Detailed Analysis of the GNU GPL and Related
 
      Licenses}'s Section on GPLv2~\S2 and GPLv3~\S1.}{\S~\ref{GPLv2s2} and \S~\ref{GPLv3s1} of
 
    this tutorial}.}
 
for the GPL'd components and your modifications thereto, but not
 
for independent proprietary applications.  The procedures described in
 
this document address this typical scenario.
 

	
 

	
 
\section{Monitor Software Acquisition}
 

	
 
Software engineers deserve the freedom to innovate and import useful
 
software components to improve products.  However, along with that
 
freedom should come rules and reporting procedures to make sure that you
 
are aware of what software that you include with your product.
 

	
 
The most typical response to an initial enforcement action is: ``We
 
didn't know there was GPL'd stuff in there''.  This answer indicates
 
failure in the software acquisition and procurement process.  Integration
 
of third-party proprietary software typically requires a formal
 
arrangement and management/legal oversight before the developers
 
incorporate the software.  By contrast, developers often obtain and
 
integrate Free Software without intervention nor oversight. That ease of acquisition, however,
 
does not mean the oversight is any less necessary.  Just as your legal
 
and/or management team negotiates terms for inclusion of any proprietary
 
software, they should gently facilitate all decisions to bring Free Software into your
 
product.
 

	
 
Simple, engineering-oriented rules help provide a stable foundation for
 
Free Software integration.  For example, simply ask your software developers to send an email to a
 
standard place describing each new Free Software component they add to the system,
 
and have them include a brief description of how they will incorporate it
 
into the product.  Further, make sure developers use a revision control
 
system (such as Git or Mercurial), and
 
store the upstream versions of all software in a ``vendor branch'' or
 
similar mechanism, whereby they can easily track and find the main version
 
of the software and, separately, any local changes.
 

	
 
Such procedures are best instituted at your project's launch.  Once 
 
chaotic and poorly-sourced development processes begin, cataloging the
 
presence of GPL'd components  becomes challenging.
 

	
 
Such a situation often requires use of a tool to ``catch up'' your knowledge
 
about what software your product includes.  Most commonly, companies choose
 
some software licensing scanning tool to inspect the codebase.  However,
 
there are few tools that are themselves Free Software.  Thus, GPL enforcers
 
usually recommend the GPL'd
 
\href{http://fossology.org/}{FOSSology system}, which analyzes a
 
source code base and produces a list of Free Software licenses that may apply to
 
the code.  FOSSology can help you build a catalog of the sources you have
 
already used to build your product.  You can then expand that into a more
 
structured inventory and process.
 

	
 
\section{Track Your Changes and Releases}
 

	
 
As explained in further detail below, the most important component of GPL
 
compliance is the one most often ignored: proper inclusion of CCS in all
 
distributions  of GPL'd
 
software.  To comply with GPL's CCS requirements, the distributor
 
\textit{must} always know precisely what sources generated a given binary
 
distribution.
 

	
 
In an unfortunately large number of our enforcement cases, the violating
 
company's engineering team had difficulty reconstructing the CCS
 
for binaries distributed by the company.  Here are three simple rules to
 
follow to decrease the likelihood of this occurrence:
 

	
 
\begin{itemize}
 

	
 
\item Ensure that your
 
developers are using revision control systems properly.
 

	
 
\item Have developers mark or ``tag'' the full source tree corresponding to
 
  builds distributed to customers.
 

	
 
\item Check that your developers store all parts of the software
 
development in the revision control system, including {\sc readme}s, build
 
scripts, engineers' notes, and documentation.
 
\end{itemize}
 

	
 
Your developers will benefit anyway from these rules.  Developers will be
 
happier in their jobs if their tools already track the precise version of
 
source that corresponds to any deployed binary.
 

	
 
\section{Avoid the ``Build Guru''}
 

	
 
Too many software projects rely on only one or a very few team members who
 
know how to build and assemble the final released product.  Such knowledge
 
centralization not only creates engineering redundancy issues, but also
 
thwarts GPL compliance.  Specifically, CCS does not just require source code,
 
but scripts and other material that explain how to control compilation and
 
installation of the executable and object code.
 

	
 
Thus, avoid relying on a ``build guru'', a single developer who is the only one
 
who knows how to produce your final product. Make sure the build process
 
is well defined.  Train every developer on the build process for the final
 
binary distribution, including (in the case of embedded software)
 
generating a final firmware image suitable for distribution to the
 
customer.  Require developers to use revision control for build processes.
 
Make a rule that adding new components to the system without adequate
 
build instructions (or better yet, scripts) is unacceptable engineering
 
practice.
 

	
 
\chapter{Details of Compliant Distribution}
 

	
 
Distribution of GPL'd works has requirements; copyleft will not function
 
without placing requirements on redistribution.  However, some requirements
 
are more likely to cause compliance difficult than others.  This
 
chapter\footnote{Note that this chapter refers heavily to specific provisions
 
  and language in
 
  \hyperref[GPLv2s3-full-text]{GPLv2\S3}
 
  and \hyperref[GPLv3s6-full-text]{GPLv3\S6}.
 
  It may be helpful  to review \S~\ref{GPLv2s3} and \S~\ref{GPLv3s6} first,
 
  and then have a copy of each license open while reading this
 
  section.}  explains some the specific requirements placed upon
 
distributors of GPL'd software that redistributors are most likely to
 
overlook, yielding compliance problems.
 

	
 
First, \hyperref[GPLv2s1]{GPLv2\S1} and \hyperref[GPLv2s4]{GPLv2\S4} require
 
that the full license text must accompany every distribution (either in
 
source or binary form) of each licensed work.  Strangely, this requirement is
 
responsible for a surprisingly significant fraction of compliance errors; too
 
often, physical products lack required information about the presence of
 
GPL'd programs and the applicable license terms.  Automated build processes
 
can and should carry a copy of the license from the the source distribution
 
into the final binary firmware package for embedded products.  Such
 
automation usually achieves compliance regarding license inclusion
 
requirements\footnote{At least one COGEO recommends the
 
  \href{https://www.yoctoproject.org/}{Yocto Project}, since its engineers
 
  have designed such features into it build process.}
 

	
 
\section{Binary Distribution Permission}
 
\label{binary-distribution-permission}
 

	
 
% be careful below, you cannot refill the \if section, so don't refill
 
% this paragraph without care.
 

	
 
The various versions of the GPL are copyright licenses that grant
 
permission to make certain uses of software that are otherwise restricted
 
by copyright law.  This permission is conditioned upon compliance with the
 
GPL's requirements.
 

	
 
This section walks through the requirements (of both GPLv2 and GPLv3) that
 
apply when you distribute GPL'd programs in binary (i.e., executable or
 
object code) form, which is typical for embedded applications.  Because a
 
binary application derives from a program's original sources, you need
 
permission from the copyright holder to distribute it.  \S~3 of GPLv2 and
 
\S~6 of GPLv3 contain the permissions and conditions related to binary
 
distributions of GPL'd programs.\footnote{These sections cannot be fully
 
  understood in isolation; read the entire license thoroughly before
 
  focusing on any particular provision.  However, once you have read and
 
  understood the entire license, look to these sections to guide
 
  compliance for binary distributions.}  Failure to provide or offer CCS is the
 
single largest failure mode leading to compliance disputes.
 

	
 

	
 

	
 
GPL's binary distribution sections offer a choice of compliance methods,
 
each of which we consider in turn.  Each option refers to the
 
``Corresponding Source'' code for the binary distribution, which includes
 
the source code from which the binary was produced.  This abbreviated and
 
simplified definition is sufficient for the binary distribution discussion
 
in this section, but you may wish to refer back to this section after
 
reading the thorough discussion of ``Corresponding Source'' that appears
 
in \S~\ref{corresponding-source}.
 

	
 
\subsection{Option (a): Source Alongside Binary}
 

	
 
GPLv2~\S~3(a) and v3~\S~6(a) embody the easiest option for providing
 
source code: including Corresponding Source with every binary
 
distribution.  While other options appear initially less onerous, this
 
option invariably minimizes potential compliance problems, because when
 
you distribute Corresponding Source with the binary, \emph{your GPL
 
  obligations are satisfied at the time of distribution}.  This is not
 
true of other options, and for this reason, we urge you to seriously
 
consider this option.  If you do not, you may extend the duration of your
 
obligations far beyond your last binary distribution.
 

	
 
Compliance under this option is straightforward.  If you ship a product
 
that includes binary copies of GPL'd software (e.g., in firmware, or on a
 
hard drive, CD, or other permanent storage medium), you can store the
 
Corresponding Source alongside the binaries.  Alternatively, you can
 
include the source on a CD or other removable storage medium in the box
 
containing the product.
 

	
 
GPLv2 refers to the various storage mechanisms as ``medi[a] customarily
 
used for software interchange''.  While the Internet has attained primacy
 
as a means of software distribution where super-fast Internet connections
 
are available, GPLv2 was written at a time when downloading software was
 
not practical (and was often impossible).  For much of the world, this
 
condition has not changed since GPLv2's publication, and the Internet
 
still cannot be considered ``a medium customary for software
 
interchange''.  GPLv3 clarifies this matter, requiring that source be
 
``fixed on a durable physical medium customarily used for software
 
interchange''.  This language affirms that option (a) requires binary
 
redistributors to provide source on a physical medium.
 

	
 
Please note that while selection of option (a) requires distribution on a
 
physical medium, voluntary distribution via the Internet is very useful.  This
 
is discussed in detail in \S~\ref{offer-with-internet}.
 

	
 
\subsection{Option (b): The Offer}
 
\label{offer-for-source}
 

	
 
Many distributors prefer to ship only an offer for source with the binary
 
distribution, rather than the complete source package.  This
 
option has value when the cost of source distribution is a true
 
per-unit cost.  For example, this option might be a good choice for
 
embedded products with permanent storage too small to fit the source, and
 
which are not otherwise shipped with a CD but \emph{are} shipped with a
 
manual or other printed material.
 

	
 
However, this option increases the duration of your obligations
 
dramatically.  An offer for source must be good for three full years from
 
your last binary distribution (under GPLv2), or your last binary or spare
 
part distribution (under GPLv3).  Your source code request and
 
provisioning system must be designed to last much longer than your product
 
life cycle. Thus, it also increases your compliance costs in the long
 
run.
 

	
 
In addition, if you are required to comply with the terms of GPLv2, you
 
{\bf cannot} use a network service to provide the source code.  For GPLv2,
 
the source code offer is fulfilled only with physical media.  This usually
 
means that you must continue to produce an up-to-date ``source code CD''
 
for years after the product's end-of-life.
 

	
 
\label{offer-with-internet}
 

	
 
Under GPLv2, it is acceptable and advisable for your offer for source code
 
to include an Internet link for downloadable source \emph{in addition} to
 
offering source on a physical medium.  This practice enables those with
 
fast network connections to get the source more quickly, and typically
 
decreases the number of physical media fulfillment requests.
 
(GPLv3~\S~6(b) permits provision of source with a public
 
network-accessible distribution only and no physical media.  We discuss
 
this in detail at the end of this section.)
 

	
 
The following is a suggested compliant offer for source under GPLv2 (and
 
is also acceptable for GPLv3) that you would include in your printed
 
materials accompanying each binary distribution:
 

	
 
\begin{quote}
 
The software included in this product contains copyrighted software that
 
is licensed under the GPL\@.  A copy of that license is included in this
 
document on page $X$\@.  You may obtain the complete Corresponding Source
 
code from us for a period of three years after our last shipment of this
 
product, which will be no earlier than 2011-08-01, by sending a money
 
order or check for \$5 to: \\
 
GPL Compliance Division \\
 
Our Company \\
 
Any Town, US 99999 \\
 
\\
 
Please write ``source for product $Y$'' in the memo line of your
 
payment.
 

	
 
You may also find a copy of the source at
 
\url{http://www.example.com/sources/Y/}.
 

	
 
This offer is valid to anyone in receipt of this information.
 
\end{quote}
 

	
 
There are a few important details about this offer.  First, it requires a
 
copying fee.  GPLv2 permits ``a charge no more than your cost of
 
physically performing source distribution''.  This fee must be reasonable.
 
If your cost of copying and mailing a CD is more than around \$10, you
 
should perhaps find a cheaper CD stock and shipment method.  It is simply
 
not in your interest to try to overcharge the community.  Abuse of this
 
provision in order to make a for-profit enterprise of source code
 
provision will likely trigger enforcement action.
 

	
 
Second, note that the last line makes the offer valid to anyone who
 
requests the source.  This is because v2~\S~3(b) requires that offers be
 
``to give any third party'' a copy of the Corresponding Source.  GPLv3 has
 
a similar requirement, stating that an offer must be valid for ``anyone
 
who possesses the object code''.  These requirements indicated in
 
v2~\S~3(c) and v3~\S~6(c) are so that noncommercial redistributors may
 
pass these offers along with their distributions.  Therefore, the offers
 
must be valid not only to your customers, but also to anyone who received
 
a copy of the binaries from them.  Many distributors overlook this
 
requirement and assume that they are only required to fulfill a request
 
from their direct customers.
 

	
 
The option to provide an offer for source rather than direct source
 
distribution is a special benefit to companies equipped to handle a
 
fulfillment process.  GPLv2~\S~3(c) and GPLv3~\S~6(c) avoid burdening
 
noncommercial, occasional redistributors with fulfillment request
 
obligations by allowing them to pass along the offer for source as they
 
received it.
 

	
 
Note that commercial redistributors cannot avail themselves of the option
 
(c) exception, and so while your offer for source must be good to anyone
 
who receives the offer (under v2) or the object code (under v3), it
 
\emph{cannot} extinguish the obligations of anyone who commercially
 
redistributes your product.  The license terms apply to anyone who
 
distributes GPL'd software, regardless of whether they are the original
 
distributor.  Take the example of Vendor $V$, who develops a software
 
platform from GPL'd sources for use in embedded devices.  Manufacturer $M$
 
contracts with $V$ to install the software as firmware in $M$'s device.
 
$V$ provides the software to $M$, along with a compliant offer for source.
 
In this situation, $M$ cannot simply pass $V$'s offer for source along to
 
its customers.  $M$ also distributes the GPL'd software commercially, so
 
$M$ too must comply with the GPL and provide source (or $M$'s \emph{own}
 
offer for source) to $M$'s customers.
 

	
 
This situation illustrates that the offer for source is often a poor
 
choice for products that your customers will likely redistribute.  If you
 
include the source itself with the products, then your distribution to
 
your customers is compliant, and their (unmodified) distribution to their
 
customers is likewise compliant, because both include source.  If you
 
include only an offer for source, your distribution is compliant but your
 
customer's distribution does not ``inherit'' that compliance, because they
 
have not made their own offer to accompany their distribution.
 

	
 
The terms related to the offer for source are quite different if you
 
distribute under GPLv3.  Under v3, you may make source available only over
 
a network server, as long as it is available to the general public and
 
remains active for three years from the last distribution of your product
 
or related spare part.  Accordingly, you may satisfy your fulfillment
 
obligations via Internet-only distribution.  This makes the ``offer for
 
source'' option less troublesome for v3-only distributions, easing
 
compliance for commercial redistributors.  However, before you switch to a
 
purely Internet-based fulfillment process, you must first confirm that you
 
can actually distribute \emph{all} of the software under GPLv3.  Some
 
programs are indeed licensed under ``GPLv2, \emph{or any later version}''
 
(often abbreviated ``GPLv2-or-later'').  Such licensing gives you the
 
option to redistribute under GPLv3.  However, a few popular programs are
 
only licensed under GPLv2 and not ``or any later version''
 
(``GPLv2-only'').  You cannot provide only Internet-based source request
 
fulfillment for the latter programs.
 

	
 
If you determine that all GPL'd works in your whole product allow upgrade
 
to GPLv3 (or were already GPLv3'd to start), your offer for source may be
 
as simple as this:
 

	
 
\begin{quote}
 
The software included in this product contains copyrighted software that
 
is licensed under the GPLv3\@.  A copy of that license is included in this
 
document on page $X$\@.  You may obtain the complete Corresponding Source
 
code from us for a period of three years after our last shipment of this
 
product and/or spare parts therefor, which will be no earlier than
 
2011-08-01, on our website at
 
\url{http://www.example.com/sources/productnum/}.
 
\end{quote}
 

	
 
\medskip
 

	
 
Under both GPLv2 and GPLv3, source offers must be accompanied by a copy of
 
the license itself, either electronically or in print, with every
 
distribution.
 
 
 
Finally, it is unacceptable to use option (b) merely because you do not have
 
Corresponding Source ready.  We find that some companies choose this option
 
because writing an offer is easy, but producing a source distribution as
 
an afterthought to a hasty development process is difficult.  The offer
 
for source does not exist as a stop-gap solution for companies rushing to
 
market with an out-of-compliance product.  If you ship an offer for source
 
with your product but cannot actually deliver \emph{immediately} on that
 
offer when your customers request it, you should expect an enforcement
 
action.
 

	
 
\subsection{Option (c): Noncommercial Offers}
 

	
 
As discussed in the last section, GPLv2~\S~3(c) and GPLv3~\S~6(c) apply
 
only to noncommercial use.  These options are not available to businesses
 
distributing GPL'd software.  Consequently, companies that redistribute
 
software packaged for them by an upstream vendor cannot merely pass along
 
the offer they received from the vendor; they must provide their own offer
 
or corresponding source to their distributees.  We talk in detail about
 
upstream software providers in \S~\ref{upstream}.
 

	
 
\subsection{Option 6(d) in GPLv3: Internet Distribution}
 

	
 
Under GPLv2, your formal provisioning options for Corresponding Source
 
ended with \S~3(c).  But even under GPLv2, pure Internet source
 
distribution was a common practice and generally considered to be
 
compliant.  GPLv2 mentions Internet-only distribution almost as aside in
 
the language, in text at the end of the section after the three
 
provisioning options are listed.  To quote that part of GPLv2~\S~3:
 
\begin{quote}
 
If distribution of executable or object code is made by offering access to
 
copy from a designated place, then offering equivalent access to copy the
 
source code from the same place counts as distribution of the source code,
 
even though third parties are not compelled to copy the source along with
 
the object code.
 
\end{quote}
 

	
 
When that was written in 1991, Internet distribution of software was the
 
exception, not the rule.  Some FTP sites existed, but generally software
 
was sent on magnetic tape or CDs.  GPLv2 therefore mostly assumed that
 
binary distribution happened on some physical media.  By contrast,
 
GPLv3~\S~6(d) explicitly gives an option for this practice that the
 
community has historically considered GPLv2-compliant.
 

	
 
Thus, you may fulfill your source-provision obligations by providing the
 
source code in the same way and from the same location.  When exercising
 
this option, you are not obligated to ensure that users download the
 
source when they download the binary, and you may use separate servers as
 
needed to fulfill the requests as long as you make the source as
 
accessible as the binary.  However, you must ensure that users can easily
 
find the source code at the time they download the binary. GPLv3~\S~6(d)
 
thus clarifies a point that has caused confusion about source provision in
 
v2.  Indeed, many such important clarifications are included in v3 which
 
together provide a compelling reason for authors and redistributors alike
 
to adopt GPLv3.
 

	
 
\subsection{Option 6(e) in GPLv3: Software Torrents}
 

	
 
Peer-to-peer file sharing arose well after GPLv2 was written, and does not
 
easily fit any of the v2 source provision options.  GPLv3~\S~6(e)
 
addresses this issue, explicitly allowing for distribution of source and
 
binary together on a peer-to-peer file sharing network.  If you distribute
 
solely via peer-to-peer networks, you can exercise this option.  However,
 
peer-to-peer source distribution \emph{cannot} fulfill your source
 
provision obligations for non-peer-to-peer binary distributions.  Finally,
 
you should ensure that binaries and source are equally seeded upon initial
 
peer-to-peer distribution.
 

	
 
\section{Preparing Corresponding Source}
 
\label{corresponding-source}
 

	
 
Most enforcement cases involve companies that have unfortunately not
 
implemented procedures like our \S~\ref{best-practices} recommendations
 
and have no source distribution arranged at all.  These companies must
 
work backwards from a binary distribution to come into compliance.  Our
 
recommendations in \S~\ref{best-practices} are designed to make it easy to
 
construct a complete and Corresponding Source release from the outset.  If
 
you have followed those principles in your development, you can meet the
 
following requirements with ease.  If you have not, you may have
 
substantial reconstruction work to do.
 

	
 
\subsection{Assemble the Sources}
 

	
 
For every binary that you produce, you should collect and maintain a copy
 
of the sources from which it was built.  A large system, such as an
 
embedded firmware, will probably contain many GPL'd and LGPL'd components
 
for which you will have to provide source.  The binary distribution may
 
also contain proprietary components which are separate and independent
 
works that are covered by neither the GPL nor LGPL\@.
 

	
 
The best way to separate out your sources is to have a subdirectory for
 
each component in your system.  You can then easily mark some of them as
 
required for your Corresponding Source releases.  Collecting
 
subdirectories of GPL'd and LGPL'd components is the first step toward
 
preparing your release.
 

	
 
\subsection{Building the Sources}
 

	
 
Few distributors, particularly of embedded systems, take care to read the
 
actual definition of Corresponding Source in the GPL\@.  Consider
 
carefully the definition, from GPLv3:
 
\begin{quote}
 
  The ``Corresponding Source'' for a work in object code form means all
 
  the source code needed to generate, install, and (for an executable
 
  work) run the object code and to modify the work, including scripts to
 
  control those activities.
 
\end{quote}
 

	
 
and the definition from GPLv2:
 
\begin{quote}
 
The source code for a work means the preferred form of the work for making
 
modifications to it.  For an executable work, complete source code means
 
all the source code for all modules it contains, plus any associated
 
interface definition files, plus the scripts used to control compilation
 
and installation of the executable.
 
\end{quote}
 

	
 
Note that you must include ``scripts used to control compilation and
 
installation of the executable'' and/or anything ``needed to generate,
 
install, and (for an executable work) run the object code and to modify
 
the work, including scripts to control those activities''.  These phrases
 
are written to cover different types of build environments and systems.
 
Therefore, the details of what you need to provide with regard to scripts
 
and installation instructions vary depending on the software details.  You
 
must provide all information necessary such that someone generally skilled
 
with computer systems could produce a binary similar to the one provided.
 

	
 
Take as an example an embedded wireless device.  Usually, a company
 
distributes a firmware, which includes a binary copy of
 
Linux\footnote{``Linux'' refers only to the kernel, not the larger system
 
  as a whole.} and a filesystem.  That filesystem contains various binary
 
programs, including some GPL'd binaries, alongside some proprietary
 
binaries that are separate works (i.e., not derived from, nor based on
 
freely-licensed sources).  Consider what, in this case, constitutes adequate
 
``scripts to control compilation and installation'' or items ``needed to
 
generate, install and run'' the GPL'd programs.
 

	
 
Most importantly, you must provide some sort of roadmap that allows
 
technically sophisticated users to build your software.  This can be
 
complicated in an embedded environment.  If your developers use scripts to
 
control the entire compilation and installation procedure, then you can
 
simply provide those scripts to users along with the sources they act
 
upon.  Sometimes, however, scripts were never written (e.g., the
 
information on how to build the binaries is locked up in the mind of your
 
``build guru'').  In that case, we recommend that you write out build
 
instructions in a natural language as a detailed, step-by-step {\sc
 
  readme}.
 

	
 
No matter what you offer, you need to give those who receive source a
 
clear path from your sources to binaries similar to the ones you ship.  If
 
you ship a firmware (kernel plus filesystem), and the filesystem contains
 
binaries of GPL'd programs, then you should provide whatever is necessary
 
to enable a reasonably skilled user to build any given GPL'd source
 
program (and modified versions thereof), and replace the given binary in
 
your filesystem.  If the kernel is Linux, then the users must have the
 
instructions to do the same with the kernel.  The best way to achieve this
 
is to make available to your users whatever scripts or process your
 
engineers would use to do the same.
 

	
 
These are the general details for how installation instructions work.
 
Details about what differs when the work is licensed under LGPL is
 
discussed in \S~\ref{lgpl}, and specific details that are unique to
 
GPLv3's installation instructions are in \S~\ref{user-products}.
 

	
 
\subsection{What About the Compiler?}
 

	
 
The GPL contains no provision that requires distribution of the compiler
 
used to build the software.  While companies are encouraged to make it as
 
easy as possible for their users to build the sources, inclusion of the
 
compiler itself is not normally considered mandatory.  The Corresponding
 
Source definition -- both in GPLv2 and GPLv3 -- has not been typically
 
read to include the compiler itself, but rather things like makefiles,
 
build scripts, and packaging scripts.
 

	
 
Nonetheless, in the interest of goodwill and the spirit of the GPL, most
 
companies do provide the compiler itself when they are able, particularly
 
when the compiler is based on GCC\@ or another copylefted compiler.  If you have
 
a GCC-based system, it is your prerogative to redistribute that GCC
 
version (binaries plus sources) to your customers.  We in the software freedom
 
community encourage you to do this, since it often makes it easier for
 
users to exercise their software freedom.  However, if you chose to take
 
this recommendation, ensure that your GCC distribution is itself
 
compliant.
 

	
 
If you have used a proprietary, third-party compiler to build the
 
software, then you probably cannot ship it to your customers.  We consider
 
the name of the compiler, its exact version number, and where it can be
 
acquired as information that \emph{must} be provided as part of the
 
Corresponding Source.  This information is essential to anyone who wishes
 
to produce a binary.  It is not the intent of the GPL to require you to
 
distribute third-party software tools to your customer (provided the tools
 
themselves are not based on the GPL'd software shipped), but we do believe
 
it requires that you give the user all the essential non-proprietary facts
 
that you had at your disposal to build the software.  Therefore, if you
 
choose not to distribute the compiler, you should include a {\sc readme}
 
about where you got it, what version it was, and who to contact to acquire
 
it, regardless of whether your compiler is Free Software, proprietary, or
 
internally developed.
 

	
 
\section{Best Practices and Corresponding Source}
 

	
 
\S~\ref{best-practices} and \S~\ref{corresponding-source} above are
 
closely related.  If you follow the best practices outlined above, you
 
will find that preparing your Corresponding Source release is an easier
 
task, perhaps even a trivial one.
 

	
 
Indeed, the enforcement process itself has historically been useful to
 
software development teams.  Development on a deadline can lead
 
organizations to cut corners in a way that negatively impacts its
 
development processes.  We have frequently been told by violators that
 
they experience difficulty when determining the exact source for a binary
 
in production (in some cases because their ``build guru'' quit during the
 
release cycle).  When management rushes a development team to ship a
 
release, they are less likely to keep release sources tagged and build
 
systems well documented.
 

	
 
We suggest that, if contacted about a violation, product builders use GPL
 
enforcement as an opportunity to improve their development practices.  No
 
developer would argue that their system is better for having a mysterious
 
build system and no source tracking.  Address these issues by installing a
 
revision system, telling your developers to use it, and requiring your
 
build guru to document his or her work!
 

	
 

	
 
\section{Non-Technical Compliance Issues}
 

	
 
Certainly, the overwhelming majority of compliance issues are, in fact,
 
either procedural or technical.  Thus, the primary material in this chapter
 
so far has covered those issues.  However, a few compliance issues do require
 
more direct consideration of a legal situation.  This portion guide does not
 
consider those in detail, as a careful reading of the earlier chapters of
 
Part~\ref{gpl-lgpl-part} shows various places where legal considerations are
 
necessary for considering compliance activity.
 

	
 
For example, specific compliance issues related to
 
\hyperref[GPLv2s7]{GPLv2\S7}, \hyperref[GPLv3s7]{GPLv3\S7}, and
 
\hyperref[GPLv3s7]{GPLv3\S11} demand a more traditional approach to legal
 
license compliance.  Of course, such analysis and consideration can be
 
complicated, and some are considered in the enforcement case studies that
 
follow in the next part.  However, compliance issues related to such sections
 
are not rare, and, as is typical, no specific training is available for
 
dealing with extremely rare occurrences.
 

	
 
\section{Self-Assessment of Compliance}
 

	
 
Most companies that adopt copylefted software believe they have complied.
 
Humans usually have difficult admitting their own mistakes, particularly
 
systematic ones.  Therefore, perhaps the most important necessary step to
 
stay in compliance is a company's regular evaluation of their own compliance.
 

	
 
First, exercise a request CCS for all copylefted works from all your upstream
 
providers of software and of components embedding software.  Then, perform
 
your own CCS check on this material first, and verify that it meets the
 
requirements.  This tutorial presents later a case study of a COGEO's CCS
 
check in \S~\ref{pristine-example}, which you can emulate when examining
 
their own CCS\@.
 

	
 
Second, measure all copyleft compliance from the position of the
 
users\footnote{Realizing of course that user very well may not be your own
 
  customer.} downstream from you exercising their rights under GPL\@.  Have
 
those users received notice of the copylefted software included in your
 
product?  Is CCS available to the users easily (preferably by automated
 
means)?  Ask yourself these questions frequently.  If you cannot answer these
 
questions with certainty in the positive, dig deeper and modify your process.
 

	
 
Avoid ``compliance industry'' marketing distractions and concentrate on the
 
copylefted software you already know is in your product.  Historically, the
 
risk from a copylefted code snippet that some programmer dropped in your
 
proprietary product careless of the consequences is a problem far more
 
infrequent and less difficult to resolve.  Efficient management of the risks
 
of higher concern lies in making sure you can provide, for example, precisely
 
CCS for a copy of Coreboot, the kernel named Linux, BusyBox, or GNU tar that
 
you included in a product your company shipped two years ago than in the risk
 
of 10 lines of GPL'd Java code an engineer accidentally pasted into the
 
source of your ERP system.
 

	
 
Thus, reject the ``compliance industry'' suggestions that code scanners find
 
and help solve fundamental compliance problems.  Consider how COGEO's tend to
 
use code scanners.  FOSSology is indeed an important part of a violation
 
investigation, but such is the last step and catches only some (usually
 
minor) licensing notice problems.  Thus, code scanners can help solve minor
 
compliance problems once you have resolved the major ones.  Code scanners
 
do not manage risk.
 

	
 
\chapter{When The Letter Comes}
 

	
 
Unfortunately, many GPL violators ignore their obligations until they are
 
contacted by a copyright holder or the lawyer of a copyright holder.  You
 
should certainly contact your own lawyer if you have received a letter
 
alleging that you have infringed copyrights that were licensed to you
 
under the GPL\@.  This section outlines a typical enforcement case and
 
provides some guidelines for response.  These discussions are
 
generalizations and do not all apply to every alleged violation.  However,
 
COGEO's in particular universally follow the processes described herein.
 

	
 
\section{Communication Is Key}
 

	
 
GPL violations are typically only escalated when a company ignores the
 
copyright holder's initial communication or fails to work toward timely
 
compliance.  Accused violators should respond very promptly to the
 
initial request.  As the process continues, violators should follow up weekly with the
 
copyright holders to make sure everyone agrees on targets and deadlines
 
for resolving the situation.
 

	
 
Ensure that any staff who might receive communications regarding alleged
 
GPL violations understands how to channel the communication appropriately
 
within your organization.  Often, initial contact is addressed for general
 
correspondence (e.g., by mail to corporate headquarters or by e-mail to
 
general informational or support-related addresses).  Train the staff that
 
processes such communications to escalate them to someone with authority
 
to take action.  An uninformed response to such an inquiry (e.g., from
 
a first-level technical support person) can cause negotiations to fail
 
prematurely.
 

	
 
Answer promptly by multiple means (paper letter, telephone call, and
 
email), even if your response merely notifies the sender that you are
 
investigating the situation and will respond by a certain date.  Do not
 
let the conversation lapse until the situation is fully resolved.
 
Proactively follow up with synchronous communication means to be sure
 
communications sent by non-reliable means (such as email) were received.
 

	
 
Remember that the software freedom community generally values open communication and
 
cooperation, and these values extend to GPL enforcement.  You will
 
generally find that software freedom developers and their lawyers are willing to
 
have a reasonable dialogue and will work with you to resolve a violation
 
once you open the channels of communication in a friendly way.
 

	
 
Furthermore, if the complaint comes from a COGEO, assume they are
 
well-prepared.  COGEO's fully investigate compliance issues before raising
 
the issue.  The claims and concerns will be substantiated, and immediate
 
denials will likely lead the COGEO to suspect malice rather than honest
 
mistake.
 

	
 
However, the biggest and most perennial mistake that all COGEOs see during
 
enforcement is this: failure to include the violators' software development
 
teams in the enforcement discussions and negotiations.  As described above,
 
CCS verification and approval is the most time-consuming and difficult part
 
of resolving most compliance matters.  Without direct contact between
 
software developers on both sides, the resolution of the technical issues
 
involved in demonstrating that the binary distributed was built from the
 
source provided is likely to be tortuous, expensive, and tense. Your lawyers
 
will certainly be understandably reluctant to expose your employees to direct
 
inquiry from potentially adverse parties.  However, facilitated exchanges of
 
information among software engineers communicating on technical subjects
 
shortens the time to resolution, substantially reduces the cost of reaching
 
resolution, and prevents unnecessary escalation due to mutual
 
misunderstanding.  Furthermore, such frank technical discussion will often be
 
the only way to avoid compliance litigation once a violation has occurred.
 

	
 
Fortunately, these frank discussions will improve your company's
 
relationships.  Free Software development communities improve software to
 
benefit everyone, which includes you and your company.  When you use
 
copylefted community software in your products, you are part of that
 
community.  Therefore, resolving a compliance matter is an occasion to
 
strengthen your relationship to the community, by increasing communication
 
between your developers and the project whose work you use for business
 
benefit.
 

	
 
\section{Termination}
 

	
 
Many redistributors overlook the GPL's termination provision (GPLv2~\S~4 and
 
GPLv3~\S~8).  Under v2, violators forfeit their rights to redistribute and
 
modify the GPL'd software until those rights are explicitly reinstated by
 
the copyright holder.  In contrast, v3 allows violators to rapidly resolve
 
some violations without consequence.
 

	
 
If you have redistributed an application under GPLv2\footnote{This applies
 
  to all programs licensed to you under only GPLv2 (``GPLv2-only'').
 
  However, most so-called GPLv2 programs are actually distributed with
 
  permission to redistribute under GPLv2 \emph{or any later version of the
 
    GPL} (``GPLv2-or-later'').  In the latter cases, the redistributor can
 
  choose to redistribute under GPLv2, GPLv3, GPLv2-or-later or even
 
  GPLv3-or-later.  Where the redistributor has chosen v2 explicitly, the
 
  v2 termination provision will always apply.  If the redistributor has
 
  chosen v3, the v3 termination provision will always apply.  If the
 
  redistributor has chosen GPLv2-or-later, then the redistributor may want
 
  to narrow to GPLv3-only upon violation, to take advantage of the
 
  termination provisions in v3.}, but have violated the terms of GPLv2,
 
you must request a reinstatement of rights from the copyright holders
 
before making further distributions, or else cease distribution and
 
modification of the software forever.  Different copyright holders
 
condition reinstatement upon different requirements, and these
 
requirements can be (and often are) wholly independent of the GPL\@.  The
 
terms of your reinstatement will depend upon what you negotiate with the
 
copyright holder of the GPL'd program.
 

	
 
Since your rights under GPLv2 terminate automatically upon your initial
 
violation, \emph{all your subsequent distributions} are violations and
 
infringements of copyright.  Therefore, even if you resolve a violation on
 
your own, you must still seek a reinstatement of rights from the copyright
 
holders whose licenses you violated, lest you remain liable for
 
infringement for even compliant distributions made subsequent to the
 
initial violation.
 

	
 
GPLv3 is more lenient.  If you have distributed only v3-licensed programs,
 
you may be eligible under v3~\S~8 for automatic reinstatement of rights.
 
You are eligible for automatic reinstatement when:
 
\begin{itemize}
 
\item you correct the violation and are not contacted by a copyright
 
  holder about the violation within sixty days after the correction, or
 

	
 
\item you receive, from a copyright holder, your first-ever contact
 
  regarding a GPL violation, and you correct that violation within thirty
 
  days of receipt of copyright holder's notice.
 
\end{itemize}
 

	
 
In addition to these permanent reinstatements provided under v3, violators
 
who voluntarily correct their violation also receive provisional
 
permission to continue distributing until they receive contact from the
 
copyright holder.  If sixty days pass without contact, that reinstatement
 
becomes permanent.  Nonetheless, you should be prepared to cease
 
distribution during those initial sixty days should you receive a
 
termination notice from the copyright holder.
 

	
 
Given that much discussion of v3 has focused on its so-called more
 
complicated requirements, it should be noted that v3 is, in this regard,
 
more favorable to violators than v2.
 

	
 
However, note that most Linux-based systems typically include some software
 
licensed under GPLv2-only, and thus the copyright holders have withheld
 
permission to redistribute under terms of GPLv3.  In larger aggregate
 
distributions which include GPLv2-only works (such as the kernel named
 
Linux), redistributors must operate as if termination is immediate and
 
permanent, since the technological remove of GPLv2-only works from the larger
 
distribution requires much more engineering work than the negotiation
 
required to seek restoration of rights for distribution under GPLv2-only
 
after permanent termination.
 

	
 
\chapter{Standard Requests}
 
\label{enforcement-standard-requests}
 

	
 
As we noted above, different copyright holders have different requirements
 
for reinstating a violator's distribution rights.  Upon violation, you no
 
longer have a license under the GPL\@.  Copyright holders can therefore
 
set their own requirements outside the license before reinstatement of
 
rights.  We have collected below a list of reinstatement demands that
 
copyright holders often require.
 

	
 
\begin{itemize}
 

	
 
\item {\bf Compliance on all Free Software copyrights}.  Copyright holders of Free Software
 
  often want a company to demonstrate compliance for all GPL'd software in
 
  a distribution, not just their own.  A copyright holder may refuse to
 
  reinstate your right to distribute one program unless and until you
 
  comply with the licenses of all Free Software in your distribution.
 
 
 
\item {\bf Notification to past recipients}.  Users to whom you previously
 
  distributed non-compliant software should receive a communication
 
  (email, letter, bill insert, etc.) indicating the violation, describing
 
  their rights under the GPL, and informing them how to obtain a gratis source
 
  distribution.  If a customer list does not exist (such as in reseller
 
  situations), an alternative form of notice may be required (such as a
 
  magazine advertisement).
 

	
 
\item {\bf Appointment of a GPL Compliance Officer.}  The software freedom community
 
  values personal accountability when things go wrong.  Copyright holders
 
  often require that you name someone within the violating company
 
  officially responsible for Free Software license compliance, and that this
 
  individual serve as the key public contact for the community when
 
  compliance concerns arise.
 

	
 
\item {\bf Periodic Compliance Reports.}  Many copyright holders wish to
 
  monitor future compliance for some period of time after the violation.
 
  For some period, your company may be required to send regular reports on
 
  how many distributions of binary and source have occurred.
 
\end{itemize}
 

	
 
These are just a few possible requirements for reinstatement.  In the
 
context of a GPL violation, and particularly under v2's termination
 
provision, the copyright holder may have a range of requests in exchange
 
for reinstatement of rights.  These software developers are talented
 
professionals from whose work your company has benefited.  Indeed, you are
 
unlikely to find a better value or more generous license terms for similar
 
software elsewhere.  Treat the copyright holders with the same respect you
 
treat your corporate partners and collaborators.
 

	
 
\chapter{Special Topics in Compliance}
 

	
 
There are several other issues that are less common, but also relevant in
 
a GPL compliance situation.  To those who face them, they tend to be of
 
particular interest.
 

	
 
\section{LGPL Compliance}
 
\label{lgpl}
 

	
 
GPL compliance and LGPL compliance mostly involve the same issues.  As we
 
discussed in \S~\ref{derivative-works}, questions of modified versions of
 
software are highly fact-dependent and cannot be easily addressed in any
 
overview document.  The LGPL adds some additional complexity to the
 
analysis.  Namely, the various LGPL versions permit proprietary licensing
 
of certain types of modified versions.  These issues are discussed in greater
 
detail in Chapter~\ref{LGPLv2} and~\ref{LGPLv3}.  However, as a rule of thumb, once you have determined
 
(in accordance with LGPLv3) what part of the work is the ``Application''
 
and what portions of the source are ``Minimal Corresponding Source'', then
 
you can usually proceed to follow the GPL compliance rules that
 
discussed above, replacing our discussion of ``Corresponding Source'' with
 
``Minimal Corresponding Source''.
 

	
 
LGPL also requires that you provide a mechanism to combine the Application
 
with a modified version of the library, and outlines some options for
 
this.  Also, the license of the whole work must permit ``reverse
 
engineering for debugging such modifications'' to the library.  Therefore,
 
you should take care that the EULA used for the Application does not
 
contradict this permission.
 

	
 
Thus, under the terms of LGPL, you must refrain from license terms on works
 
based on the licensed work that prohibit replacement of the licensed
 
components of the larger non-LGPL'd work, or prohibit decompilation or
 
reverse engineering in order to enhance or fix bugs in the LGPL'd components.
 

	
 
LGPLv3 is not surprisingly easier to understand and examine from a compliance
 
lens, since the FSF was influenced in LGPLv3's drafting by questions and
 
comments on LGPLv2.1 over a period of years.  Admittedly, LGPLv2.1 is still
 
in wide use, and thus compliance with LGPLv2.1 remains a frequent topic you
 
may encounter.  The best advice there is careful study of
 
Chapter~\ref{LGPLv2}.
 

	
 
However, to repeat a key point here made within that chapter: Note though
 
that, since the LGPLv2.1 can be easily upgraded to GPLv2-or-later, in the
 
worst case you simply need to comply as if the software was licensed under
 
GPLv2.  The only reason you must consider the question of whether you have a
 
``work that uses the library'' or a ``work based on the library'' is when you
 
wish to take advantage of the ``weak copyleft'' effect of the Lesser GPL\@.
 
If GPLv2-or-later is an acceptable license (i.e., if you plan to copyleft the
 
entire work anyway), you may find this an easier option.
 

	
 
\section{Upstream Providers}
 
\label{upstream}
 

	
 
With ever-increasing frequency, software development (particularly for
 
embedded devices) is outsourced to third parties.  If you rely on an
 
upstream provider for your software, note that you \emph{cannot ignore
 
  your GPL compliance requirements} simply because someone else packaged
 
the software that you distribute.  If you redistribute GPL'd software
 
(which you do, whenever you ship a device with your upstream's software in
 
it), you are bound by the terms of the GPL\@.  No distribution (including
 
redistribution) is permissible absent adherence to the license terms.
 

	
 
Therefore, you should introduce a due diligence process into your software
 
acquisition plans.  This is much like the software-oriented
 
recommendations we make in \S~\ref{best-practices}.  Implementing
 
practices to ensure that you are aware of what software is in your devices
 
can only improve your general business processes.  You should ask a clear
 
list of questions of all your upstream providers and make sure the answers
 
are complete and accurate.  The following are examples of questions you
 
should ask:
 
\begin{itemize}
 

	
 
\item What are all the licenses that cover the software in this device?
 

	
 
\item From which upstream vendors, be they companies or individuals, did
 
  \emph{you} receive your software before distributing it to us?
 

	
 
\item What are your GPL compliance procedures?
 

	
 
\item If there is GPL'd software in your distribution, we will be
 
  redistributors of this GPL'd software.  What mechanisms do you have in
 
  place to aid us with compliance?
 

	
 
\item If we follow your recommended compliance procedures, will you
 
  formally indemnify us in case we are nonetheless found to be in
 
  violation of the GPL?
 

	
 
\end{itemize}
 

	
 
This last point is particularly important.  Many GPL enforcement actions are
 
escalated because of petty finger-pointing between the distributor and its
 
upstream.  In our experience, agreements regarding GPL compliance issues
 
and procedures are rarely negotiated up front.  However, when they are,
 
violations are resolved much more smoothly (at least from the point of
 
view of the redistributor).
 

	
 
Consider the cost of potential violations in your acquisition process.
 
Using Free Software allows software vendors to reduce costs significantly, but be
 
wary of vendors who have done so without regard for the licenses.  If your
 
vendor's costs seem ``too good to be true,'' you may ultimately bear the
 
burden of the vendor's inattention to GPL compliance.  Ask the right
 
questions, demand an account of your vendors' compliance procedures, and
 
seek indemnity from them.
 

	
 
In particular, any time your vendor incorporates copylefted software, you
 
\textit{must} exercise your own rights as a user to request CCS for all the
 
copylefted programs that your suppliers provided to you.  Furthermore, you
 
must ensure that CCS is correct and adequate yourself.  Good vendors should
 
help you do this, and make it easy.  If those vendors cannot, pick a
 
different vendor before proceeding with the product. 
 

	
 
\section{Mergers and Acquisitions}
 

	
 
Often, larger companies often encounter copyleft licensing during a Mergers
 
and Acquisitions (M\&A) process.  Ultimately, a merger or acquisition causes
 
all of the other company's problems to become yours.  Therefore, for most
 
concerns, the acquirer ``simply'' must apply the compliance analysis and
 
methodologies discussed earlier across the acquired company's entire product
 
line.  Of course, this is not so simple, as such effort may be substantial,
 
but a well-defined process for compliance investigation means the required
 
work, while voluminous, is likely rote.
 

	
 
A few sections of GPL require careful attention and legal analysis to
 
determine the risk of acquisitions.  Those handling M\&A issues should pay
 
particular attention to the requirements of GPLv2~\S7 and GPLv3~\S10--12 ---
 
focusing on how they relate to the acquired assets may be of particular
 
importance.
 

	
 
For example, GPLv3\S10 clarifies that in business acquisitions, whether by
 
sale of assets or transfers of control, the acquiring party is downstream
 
from the party acquired.  This results in new automatic downstream licenses
 
from upstream copyright holders, licenses to all modifications made by the
 
acquired business, and rights to source code provisioning for the
 
now-downstream purchaser.  However, despite this aid given by explicit
 
language in GPLv3, acquirers must still confirm compliance by the acquired
 
(even if GPLv3\S10 does assert the the acquirers rights under GPL, that does
 
not help if the acquired is out of compliance altogether).  Furthermore, for
 
fear of later reprisal by the acquirer if a GPL violation is later discovered
 
in the acquired's product line, the acquired may need to seek a waiver and
 
release of from additional damages beyond a requirement to comply fully (and
 
a promise of rights restoration) if a GPL violation by the acquired is later
 
uncovered during completion of the acquisition or thereafter.
 

	
 
Finally, other advice available regarding handling of GPL compliance in an
 
M\&A situation tends to ignore the most important issue: most essential
 
copylefted software is not wholly copyrighted by the entities involved in the
 
M\&A transaction.  Therefore, copyleft obligations likely reach out to the
 
customers of all entities involved, as well as to the original copyright
 
holders of the copylefted work.  As such, notwithstanding the two paragraphs
 
in GPLv3\S10, the entities involved in M\&A should read the copyleft licenses
 
through the lens of third parties whose software freedom rights under those
 
licenses are of equal importance to then entities inside the transaction.
 

	
 
\section{User Products and Installation Information}
 
\label{user-products}
 

	
 
GPLv3 requires you to provide ``Installation Information'' when v3
 
software is distributed in a ``User Product.''  During the drafting of v3,
 
the debate over this requirement was contentious.  However, the provision
 
as it appears in the final license is reasonable and easy to understand.
 

	
 
If you put GPLv3'd software into a User Product (as defined by the
 
license) and \emph{you} have the ability to install modified versions onto
 
that device, you must provide information that makes it possible for the
 
user to install functioning, modified versions of the software.  Note that
 
if no one, including you, can install a modified version, this provision
 
does not apply.  For example, if the software is burned onto an
 
non-field-upgradable ROM chip, and the only way that chip can be upgraded
 
is by producing a new one via a hardware factory process, then it is
 
acceptable that the users cannot electronically upgrade the software
 
themselves.
 

	
 
Furthermore, you are permitted to refuse support service, warranties, and
 
software updates to a user who has installed a modified version.  You may
 
even forbid network access to devices that behave out of specification due
 
to such modifications.  Indeed, this permission fits clearly with usual
 
industry practice.  While it is impossible to provide a device that is
 
completely unmodifiable\footnote{Consider that the iPhone, a device
 
  designed primarily to restrict users' freedom to modify it, was unlocked
 
  and modified within 48 hours of its release.}, users are generally on
 
notice that they risk voiding their warranties and losing their update and
 
support services when they make modifications.\footnote{A popular t-shirt
 
  in the software freedom community reads: ``I void warranties.''.  Our community is
 
  well-known for modifying products with full knowledge of the
 
  consequences.  GPLv3's ``Installation Instructions'' section merely
 
  confirms that reality, and makes sure GPL rights can be fully exercised,
 
  even if users exercise those rights at their own peril.}
 

	
 
GPLv3 is in many ways better for distributors who seek some degree of
 
device lock-down.  Technical processes are always found for subverting any
 
lock-down; pursuing it is a losing battle regardless.  With GPLv3, unlike
 
with GPLv2, the license gives you clear provisions that you can rely on
 
when you are forced to cut off support, service or warranty for a customer
 
who has chosen to modify.
 

	
 
% FIXME-soon: write a full section on Javascript compliance.  Here's a
 
%             potentially useful one-sentence introduction for such a
 
%             section.
 

	
 
% Non-compliance with GPLv3 in the
 
% distribution of Javascript on the Web is becoming more frequent
 
%FIXME-soon: END
 

	
 
\section{Beware The Consultant in Enforcers' Clothing}
 

	
 
There are admittedly portions of the GPL enforcement community that function
 
somewhat like the
 
\href{http://en.wikipedia.org/wiki/Hacker_%28computer_security%29#Classifications}{computer
 
  security and network penetration testing hacker community}.  By analogy,
 
most COGEO's consider themselves
 
\href{http://en.wikipedia.org/wiki/White_hat_%28computer_security%29}{white hats},
 
while some might appropriately call
 
\hyperref[Proprietary Relicensing]{proprietary relicensing} by the name ``\href{http://en.wikipedia.org/wiki/Hacker_%28computer_security%29#Black_hat}{black hats}''.
 
And, to finalize the analogy, there are indeed few
 
\href{http://en.wikipedia.org/wiki/Grey_hat}{grey hat} GPL enforcers.
 

	
 
Grey hat GPL enforcers usually have done some community-oriented GPL
 
enforcement themselves, typically working as a volunteer for a COGEO, but make
 
their living as a ``hired gun'' consultant to find GPL violations and offer
 
to ``fix them'' for companies.  Other such operators hold copyrights in some
 
key piece of copylefted software and enforce as a mechanism to find out who
 
is most likely to fund improvements on the software.
 

	
 
A few companies report that they have formed beneficial consulting or
 
employment relationships with developers they first encountered through
 
enforcement.  In some such cases, companies have worked with such consultants
 
to alter the mode of use of the project's code in the company's products.
 
More often in these cases, the communication channels opened in the course of
 
the inquiry served other consulting purposes later.
 

	
 
Feelings and opinions about this behavior are mixed within the larger
 
copyleft community.  Some see it as a reasonable business model and others
 
renounce it as corrupt behavior.  Regardless, a GPL
 
violator should always immediately determine the motivations of the
 
enforcer via documented, verifiable facts.  For example, COGEOs such as the FSF and Conservancy have made substantial
 
public commitments to enforce in a way that is uniform, transparent, and
 
publicly documented.  Furthermore, since these specific organizations are
 
public charities in the USA, they
 
are accountable to the IRS (and the public at large) in their annual Form 990
 
filings.   Everyone may examine their revenue models and scrutinize their
 
work.
 

	
 
However, entities and individuals who do GPL enforcement centered primarily
 
around a profit motive are likely the most dangerous enforcement entities for
 
one simple reason: an agreement to comply fully with the GPL for past and
 
future products --- always the paramount goal to COGEOs --- may not suffice as
 
adequate resolution for a proprietary relicensing company or grey hat GPL
 
enforcer.  Therefore, violators must consider carefully who has
 
made the enforcement inquiry and ask when and where the enforcer made public
 
commitments and reports regarding their enforcement work and perhaps even ask
 
the enforcer to directly mimic CEOGEO's detailed public disclosures and
 
follow the \hyperref[enforcement-standard-requests]{standard requests for
 
  resolution} found in this document.
 

	
 
\chapter{Conclusion}
 

	
 
GPL compliance need not be an onerous process.  Historically, struggles
 
have been the result of poor development methodologies and communications,
 
rather than any unexpected application of the GPL's source code disclosure
 
requirements.
 

	
 
Compliance is straightforward when the entirety of your enterprise is
 
well-informed and well-coordinated.  The receptionists should know how to
 
route a GPL source request or accusation of infringement.  The lawyers
 
should know the basic provisions of Free Software licenses and your source
 
disclosure requirements, and should explain those details to the software
 
developers.  The software developers should use a version control system
 
that allows them to associate versions of source with distributed
 
binaries, have a well-documented build process that anyone skilled in the
 
art can understand, and inform the lawyers when they bring in new
 
software.  Managers should build systems and procedures that keep everyone
 
on target.  With these practices in place, any organization can comply
 
with the GPL without serious effort, and receive the substantial benefits
 
of good citizenship in the software freedom community, and lots of great code
 
ready-made for their products.
 

	
 
\vfill
 

	
 
% LocalWords:  redistributors NeXT's Slashdot Welte gpl ISC embedders BusyBox
 
% LocalWords:  someone's downloadable subdirectory subdirectories filesystem
 
% LocalWords:  roadmap README upstream's Ravicher's FOSSology readme CDs iPhone
 
% LocalWords:  makefiles violator's Michlmayr Stallman RMS GPL'd Harald LGPL
 
%%  LocalWords:  GPL's resellers copylefted sublicenses GPLv unmanaged MySQL
 
%%  LocalWords:  misassessments licensor COGEOs COGEO LGPLv CCS Requestors
 
%%  LocalWords:  codebase Yocto distributees COGEO's Coreboot ERP reseller
 
%%  LocalWords:  redistributor reinstatements decompilation acquired's grey
 
%%  LocalWords:  upgradable unmodifiable Relicensing relicensing
enforcement-case-studies.tex
Show inline comments
 
%      Tutorial Text for the Detailed Study and Analysis of GPL and LGPL course
 

	
 
% License: CC-By-SA-4.0
 

	
 
% The copyright holders hereby grant the freedom to copy, modify, convey,
 
% Adapt, and/or redistribute this work under the terms of the Creative
 
% Commons Attribution Share Alike 4.0 International License.
 

	
 
% This text is distributed in the hope that it will be useful, but
 
% WITHOUT ANY WARRANTY; without even the implied warranty of
 
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 

	
 
% You should have received a copy of the license with this document in
 
% a file called 'CC-By-SA-4.0.txt'.  If not, please visit
 
% https://creativecommons.org/licenses/by-sa/4.0/legalcode to receive
 
% the license text.
 

	
 

	
 
\part{Case Studies in GPL Enforcement}
 

	
 
{\parindent 0in
 
This part is: \\
 
\begin{tabbing}
 
Copyright \= \copyright{} 2003, 2004, 2014 \hspace{1mm} \= \hspace{1.mm} \=  \kill
 
Copyright \> \copyright{} 2014 \>  Bradley M. Kuhn. \\
 
Copyright \> \copyright{} 2014 \>  Denver Gingerich \\
 
Copyright \> \copyright{} 2003, 2004, 2014 \> Free Software Foundation, Inc. \\
 
\end{tabbing}
 

	
 
\vspace{.2in}
 

	
 
\begin{center}
 

	
 
The copyright holders hereby grant the freedom to copy, modify, convey,
 
Adapt, and/or redistribute this work under the terms of the Creative Commons
 
Attribution Share Alike 4.0 International License.  A copy of that license is
 
available at \url{https://creativecommons.org/licenses/by-sa/4.0/legalcode}.
 
\end{center}
 

	
 
\vfill
 

	
 
This part includes material from many sources, including the following
 
This part includes material from many sources, including some material from the following
 
CC-By-SA-licensed published works: \\
 

	
 
\begin{itemize}
 
\item \textit{Enforcement Case Studies}, written by Bradley M. Kuhn and published by the Free
 
  Software Foundation for its CLE courses  on 2004-01-20, 2004-08-24, and 2014-03-24.
 
\end{itemize}
 

	
 
However, this work is primarily composed of the many contributions it
 
receives as a public, collaborative project.  Please
 
\href{https://gitorious.org/copyleft-org/tutorial/history/master:enforcement-case-studies.tex}{review
 
  its Git logs} for full documentation of all contributions.
 

	
 

	
 
}
 
% =====================================================================
 
% START OF SECOND DAY SEMINAR SECTION
 
% =====================================================================
 

	
 
\chapter*{Preface}
 

	
 
This one-day course presents the details of five different GPL
 
compliance cases handled by FSF's GPL Compliance Laboratory. Each case
 
offers unique insights into problems that can arise when the terms of
 
the GPL are not properly followed, and how diplomatic negotiation between
 
the violator and the copyright holder can yield positive results for
 
both parties.
 

	
 
Attendees should have successfully completely the course, a ``Detailed
 
Study and Analysis of the GPL and LGPL,'' as the material from that
 
course forms the building blocks for this material.
 

	
 
This course is of most interest to lawyers who have clients or
 
employers that deal with Free Software on a regular basis. However,
 
technical managers and executives whose businesses use or distribute
 
Free Software will also find the course very helpful.
 

	
 
\bigskip
 

	
 
These course materials are merely a summary of the highlights of the
 
course presented. Please be aware that during the actual GPL course, class
 
discussion supplements this printed curriculum. Simply reading it is
 
not equivalent to attending the course.
 

	
 
%FIXME-LATER: write these
 

	
 
%\chapter{Not All GPL Enforcement is Created Equal}
 

	
 
%\section{For-Profit Enforcement}
 

	
 
%\section{Community and Non-Profit Enforcement}
 

	
 
\chapter{Overview of Community Enforcement}
 

	
 
The GPL is a Free Software license with legal teeth. Unlike licenses like
 
the X11-style or various BSD licenses, the GPL (and by extension, the LGPL) is
 
designed to defend as well as grant freedom. We saw in the last course
 
that the GPL uses copyright law as a mechanism to grant all the key freedoms
 
essential in Free Software, but also to ensure that those freedoms
 
propagate throughout the distribution chain of the software.
 

	
 
\section{Termination Begins Enforcement}
 

	
 
As we have learned, the assurance that Free Software under the GPL remains
 
Free Software is accomplished through various terms of the GPL: \S 3 ensures
 
that binaries are always accompanied with source; \S 2 ensures that the
 
sources are adequate, complete and usable; \S 6 and \S 7 ensure that the
 
license of the software is always the GPL for everyone, and that no other
 
legal agreements or licenses trump the GPL. It is \S 4, however, that ensures
 
that the GPL can be enforced.
 

	
 
Thus, \S 4 is where we begin our discussion of GPL enforcement. This
 
clause is where the legal teeth of the license are rooted. As a copyright
 
license, the GPL governs only the activities governed by copyright law ---
 
copying, modifying and redistributing computer software. Unlike most
 
copyright licenses, the GPL gives wide grants of permission for engaging with
 
these activities. Such permissions continue, and all parties may exercise
 
them until such time as one party violates the terms of the GPL\@. At the
 
moment of such a violation (i.e., the engaging of copying, modifying or
 
redistributing in ways not permitted by the GPL) \S 4 is invoked. While other
 
parties may continue to operate under the GPL, the violating party loses their
 
rights.
 

	
 
Specifically, \S 4 terminates the violators' rights to continue
 
engaging in the permissions that are otherwise granted by the GPL\@.
 
Effectively, their rights revert to the copyright defaults ---
 
no permission is granted to copy, modify, nor redistribute the work.
 
Meanwhile, \S 5 points out that if the violator has no rights under
 
the GPL, they are prohibited by copyright law from engaging in the
 
activities of copying, modifying and distributing. They have lost
 
these rights because they have violated the GPL, and no other license
 
gives them permission to engage in these activities governed by copyright law.
 

	
 
\section{Ongoing Violations}
 

	
 
In conjunction with \S 4's termination of violators' rights, there is
 
one final industry fact added to the mix: rarely does one engage in a
 
single, solitary act of copying, distributing or modifying software.
 
Almost always, a violator will have legitimately acquired a copy of a
 
GPL'd program, either making modifications or not, and then begun
 
distributing that work. For example, the violator may have put the
 
software in boxes and sold them at stores. Or perhaps the software
 
was put up for download on the Internet. Regardless of the delivery
 
mechanism, violators almost always are engaged in {\em ongoing\/}
 
violation of the GPL\@.
 

	
 
In fact, when we discover a GPL violation that occurred only once --- for
 
example, a user group who distributed copies of a GNU/Linux system without
 
source at one meeting --- we rarely pursue it with a high degree of
 
tenacity. In our minds, such a violation is an educational problem, and
 
unless the user group becomes a repeat offender (as it turns out, they
 
never do), we simply forward along a FAQ entry that best explains how user
 
groups can most easily comply with the GPL, and send them on their merry way.
 

	
 
It is only the cases of {\em ongoing\/} GPL violation that warrant our
 
active attention. We vehemently pursue those cases where dozens, hundreds
 
or thousands of customers are receiving software that is out of
 
compliance, and where the company continually offers for sale (or
 
distributes gratis as a demo) software distributions that include GPL'd
 
components out of compliance. Our goal is to maximize the impact of
 
enforcement and educate industries who are making such a mistake on a
 
large scale.
 

	
 
In addition, such ongoing violation shows that a particular company is
 
committed to a GPL'd product line. We are thrilled to learn that someone
 
is benefiting from Free Software, and we understand that sometimes they
 
become confused about the rules of the road. Rather than merely
 
giving us a postmortem to perform on a past mistake, an ongoing violation
 
gives us an active opportunity to educate a new contributor to the GPL'd
 
commons about proper procedures to contribute to the community.
 

	
 
Our central goal is not, in fact, to merely clear up a particular violation.
 
In fact, over time, we hope that our compliance lab will be out of
 
business. We seek to educate the businesses that engage in commerce
 
related to GPL'd software to obey the rules of the road and allow them to
 
operate freely under them. Just as a traffic officer would not revel in
 
reminding people which side of the road to drive on, so we do not revel in
 
violations. By contrast, we revel in the successes of educating an
 
ongoing violator about the GPL so that GPL compliance becomes a second-nature
 
matter, allowing that company to join the GPL ecosystem as a contributor.
 

	
 
\section{How are Violations Discovered?}
 

	
 
Our enforcement of the GPL is not a fund-raising effort; in fact, FSF's GPL
 
Compliance Lab runs at a loss (in other words, it is subsided by our
 
donors). Our violation reports come from volunteers, who have encountered,
 
in their business or personal life, a device or software product that
 
appears to contain GPL'd software. These reports are almost always sent
 
via email to $<$license-violation@fsf.org$>$.
 

	
 
Our first order of business, upon receiving such a report, is to seek
 
independent confirmation. When possible, we get a copy of the software
 
product. For example, if it is an offering that is downloadable from a
 
Web site, we download it and investigate ourselves. When it is not
 
possible for us to actually get a copy of the software, we ask the
 
reporter to go through the same process we would use in examining the
 
software.
 

	
 
By rough estimation, about 95\% of violations at this stage can be
 
confirmed by simple commands. Almost all violators have merely made an
 
error and have no nefarious intentions. They have made no attempt to
 
remove our copyright notices from the software. Thus, given the
 
third-party binary, {\tt tpb}, usually, a simple command (on a GNU/Linux
 
system) such as the following will find a Free Software copyright notice
 
and GPL reference:
 
\begin{quotation}
 
{\tt strings tpb | grep Copyright}
 
\end{quotation}
 
In other words, it is usually more than trivial to confirm that GPL'd
 
software is included.
 

	
 
Once we have confirmed that a violation has indeed occurred, we must then
 
determine whose copyright has been violated. Contrary to popular belief,
 
FSF does not have the power to enforce the GPL in all cases. Since the GPL
 
operates under copyright law, the powers of enforcement --- to seek
 
redress once \S 4 has been invoked --- lie with the copyright holder of
 
the software. FSF is one of the largest copyright holders in the world of
 
GPL'd software, but we are by no means the only one. Thus, we sometimes
 
discover that while GPL'd code is present in the software, there is no
 
software copyrighted by FSF present.
 

	
 
In cases where FSF does not hold copyright interest in the software, but
 
we have confirmed a violation, we contact the copyright holders of the
 
software, and encourage them to enforce the GPL\@. We offer our good offices
 
to help negotiate compliance on their behalf, and many times, we help as a
 
third party to settle such GPL violations. However, what we will describe
 
primarily in this course is FSF's first-hand experience enforcing its own
 
copyrights and the GPL\@.
 

	
 
\section{First Contact}
 

	
 
The Free Software community is built on a structure of voluntary
 
cooperation and mutual help. Our community has learned that cooperation
 
works best when you assume the best of others, and only change policy,
 
procedures and attitudes when some specific event or occurrence indicates
 
that a change is necessary. We treat the process of GPL enforcement in
 
the same way. Our goal is to encourage violators to join the cooperative
 
community of software sharing, so we want to open our hand in friendship.
 

	
 
Therefore, once we have confirmed a violation, our first assumption is
 
that the violation is an oversight or otherwise a mistake due to confusion
 
about the terms of the license. We reach out to the violator and ask them
 
to work with us in a collaborative way to bring the product into
 
compliance. We have received the gamut of possible reactions to such
 
requests, and in this course, we examine four specific examples of such
 
compliance work.
 

	
 
% FIXME: make this section properly TeX-formatted
 
\chapter{ThinkPenguin Wireless Router: Excellent CCS}
 
\label{pristine-example}
 

	
 
Too often, case studies examine failure and mistakes.  Indeed, most of the
 
chapters that follow herein will consider the myriad difficulties discovered
 
in community-oriented GPL enforcement for the last two decades.  However, to
 
begin, this is a case study in how copyleft compliance can indeed be done correctly.
 

	
 
This example is, in fact, more than ten years in the making.  Since almost
 
the inception of for-profit corporate adoption of Free Software, companies
 
have requested a clear example of a model citizen to emulate.  Sadly, while
 
community-oriented enforcers have vetted uncounted thousands of ``Complete,
 
Corresponding Source'' (CCS) candidates from hundreds of companies, this
 
particular CCS release described herein is the first ever declared a ``pristine
 
example''.
 

	
 
% FIXME (above): link to a further discussion of CCS in the compliance guide
 
% when a good spot exists, then (below) link to a ``CCS iteration''
 
% discussion in compliance-guide.tex when one exists.  (the ``iteration
 
% process'' is discussed in~\ref{} of this guide)
 

	
 
Of course, most CCS examined for the last decade has (eventually) complied
 
with the GPL, perhaps after many iterations of review by the enforcer.
 
However, in the experience of the two primary community-oriented enforcers
 
(Conservancy and the FSF), such CCS results routinely 
 
``barely comply with GPL's requirements''.  To use an academic analogy:
 
while a ``C'' is certainly a passing grade, any instructor prefers to
 
disseminate to the class an exemplar sample that earned an ``A''.
 

	
 
Fortunately, thanks in large part to the FSF's
 
``Respects Your Freedom'' (RYF) certification campaign\footnote{\href{http://www.fsf.org/resources/hw/endorsement/respects-your-freedom}{RYF is
 
    a campaign by FSF to certify products that truly meet the principles of
 
    software freedom}.  Products must meet
 
  \href{http://www.fsf.org/resources/hw/endorsement/criteria}{strict
 
    standards for RYF certification}, and among them is a pristine example of
 
  CCS\@.}, a few electronics products on the market meet
 
a higher standard of copyleft compliance.  As such, for the first
 
time in the history of copyleft, CCS experts have pristine examples to study
 
and present as exemplars worthy of emulation.
 

	
 
This case study therefore examines the entire life-cycle of a GPL compliance
 
investigation: from product purchase, to source request, to CCS review, and concluding
 
in a final compliance determination.
 
Specifically, this chapter discusses the purchase, CCS provision, and a
 
step-by-step build and installation analysis of a specific, physical,
 
embedded electronics product:
 
\href{https://www.thinkpenguin.com/gnu-linux/free-software-wireless-n-broadband-router-gnu-linux-tpe-nwifirouter}{the
 
  ``TPE-NWIFIROUTER'' wireless router by ThinkPenguin}.\footnote{The FSF of
 
  course performed a thorough CCS check as part of its certification process.
 
  The analysis discussed herein was independently performed by Software
 
  Freedom Conservancy without reviewing the FSF's findings.  Thus, this
 
  analysis is ``true to form'', and explains the typical procedures Conservancy
 
  uses when investigating a potential GPL violation.  In this case, obviously, no
 
  violation was uncovered.}
 

	
 
\section{Consumer Purchase and Unboxing}
 

	
 
The process for copyleft compliance investigation, when properly conducted,
 
determines whether users inclined to exercise their rights under a copyleft
 
license will be successful in their attempt.  Therefore, at every stage, the
 
investigator seeks to take actions that reasonably technically knowledgeable
 
users would during the ordinary course of their acquisition and use of
 
copyleft-covered products.  As such, the investigator typically purchases the device on the
 
open market to verify that distribution of the copylefted software therein
 
complies with binary distribution requirements (such as those
 
\tutorialpartsplit{discussed in \textit{Detailed Analysis of the GNU GPL and
 
    Related Licenses}}{discussed earlier in \S~\ref{GPLv2s3} and
 
  \S~\ref{GPLv3s6}}).
 

	
 
% FIXME: Above is my only use of \tutorialpartsplit in this chapter.  I just
 
% got lazy and that should be fixed by someone.
 

	
 
\label{thinkpenguin-included-ccs}
 

	
 
Therefore, the investigator first purchased the TPE-NWIFIROUTER through an
 
online order, and when the package arrived, examined the contents of the box.
 
The investigator immediately discovered that ThinkPenguin had taken advice
 
from \S~\ref{offer-for-source}, and exercised
 
\hyperref[GPLv2s3a]{GPLv2\S3(a)} and \hyperref[GPLv3s6]{GPLv3\S6}, rather than
 
using the \hyperref[offer-for-source]{problematic offer for source
 
  provisions}.  This choice not only accelerated the investigation (since there
 
was no CCS offer to ``test''), but also simplified the compliance requirements for
 
ThinkPenguin.
 

	
 
\section{Root Filesystem and Kernel Compilation}
 

	
 
The CD found in the box was labeled ``libreCMC v1.2.1 source code'', and
 
contained 407 megabytes of data.  The investigator copied this ISO to a
 
desktop GNU/Linux system and
 
examined its contents.  Upon doing so, the investigator immediately found a
 
file called ``README'' at the top-level directory:
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}[language=bash]
 
$ dd if=/dev/cdrom of=libreCMC_v1.2.1_SRC.iso
 
$ mkdir libCMC
 
$ sudo mount -o loop ./libreCMC_v1.2.1_SRC.iso libCMC
 
mount: block device /path/to/libreCMC_v1.2.1_SRC.iso
 
       is write-protected, mounting read-only
 
$ ls -1 libCMC
 
bin
 
librecmc-u-boot.tar.bz2
 
librecmc-v1.2.1.tar.bz2
 
README
 
u-boot_reflash
 
$ cat libCMC/README
 
...
 
\end{lstlisting}
 
\label{thinkpenguin-toplevel-readme}
 
The investigator therefore knew immediately to begin the CCS check should
 
begin with a study of the contents of ``README''.  Indeed, that file contained the appropriate
 
details to start the build:
 
\begin{quotation}
 

	
 
In order to build firmware images for your router, the following needs to be
 
installed:
 

	
 
gcc, binutils, bzip2, flex, python, perl, make, find, grep, diff, unzip,
 
gawk, getopt, libz-dev and libc headers.
 

	
 
Please use ``make menuconfig'' to configure your appreciated configuration
 
for the toolchain and firmware. Please note that the default configuration is
 
what was used to build the firmware image for your router. It is advised that
 
you use this configuration.
 

	
 
Simply running ``make'' will build your firmware.  The build system will
 
download all sources, build the cross-compile toolchain, the kernel and all
 
chosen applications.
 

	
 
To build your own firmware you need to have access to a GNU/Linux system
 
(case-sensitive filesystem required).
 
\end{quotation}
 

	
 
In other words, the first ``script'' that investigator ``ran'' was the above.
 
This was not a software script, rather the processor for the script was the investigator's own
 
brain --- like a script of a play.  Less glibly stated: instructions written in
 
English are usually necessary for the build and installation operations
 
that demand actual intelligence.
 
In this case, the investigator ascertained the host system requirements
 
for construction of this embedded firmware.
 

	
 
GPL does not give specific guidance on the form or location of
 
``scripts used to control compilation and installation of the executable''
 
and/or ``Installation Information''.  Community-oriented GPL enforcers apply a
 
``reasonableness standard'' to evaluate such instructions.  If an investigator of
 
average skill in embedded firmware construction can surmise the proper
 
procedures to build and install a replacement firmware, the instructions are
 
likely sufficient to meet GPL's requirements.  Fortunately, in this case, the
 
instructions are more abundant and give extra detail.
 

	
 
Nevertheless, these instructions offer more options than the reader
 
typically sees in other CCS candidates.  More typically, top-level build
 
instructions name an exact host distribution to use, such as
 
``Debian 7 installed on an amd64 system with the following packages
 
installed''.  Of course, if the build will fail on any other system,
 
instructions \textit{should} include such details.  However, this CCS builds
 
on a wide range of distributions, and thus it was appropriate (and preferred)
 
that the build instructions do not  specify a specific distribution.
 

	
 
\label{thinkpenguin-specific-host-system}
 

	
 
In this specific case, the developers of the libreCMC project (a Free
 
Software project that forms the base system for the TPE-NWIFIROUTER) have
 
clearly made an effort to ensure the CCS builds on a variety of host systems.
 
The investigator was in fact dubious upon seeing these instructions, since
 
finicky embedded build processes usually require a very specific host system.
 
Fortunately, it seems such doubts were generally unfounded (although the
 
investigator did find
 
\hyperref[thinkpenguin-glibc-214-issue]{a minor annoyance that could be
 
  resolved with more detailed instructions}).
 

	
 
Anyway, since these instructions did not specify a specific host system, the
 
investigator simply used his own amd64 Debian GNU/Linux 6 desktop system.  Before
 
beginning, the investigator used the following command:
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}[language=bash]
 
$ dpkg --list | egrep '^iii' | less
 
\end{lstlisting}
 

	
 
to verify that the required packages listed in the README were
 
installed\footnote{The ``dpkg'' command is a Debian-specific way of
 
  finding installed packages.}.
 

	
 

	
 
Next, the investigator then extracted the primary source package with the
 
following command:
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}[language=bash]
 
$ tar --posix -jxpf libCMC/librecmc-v1.2.1.tar.bz2
 
\end{lstlisting}
 

	
 
The investigator did notice an additional source release, entitled
 
``librecmc-u-boot.tar.bz2''.  The investigator concluded upon simple
 
inspection that the instructions found in ``u-boot\verb0_0reflash'' were
 
specific instructions for that part of the CCS\@.  This was a minor
 
annoyance, and ideally the ``README'' would so-state, but the CCS filesystem
 
layout met the reasonableness standard; the skilled investigator determine the correct
 
course of action with a few moments of study.
 

	
 
The investigator then noted the additional step offered by the ``README'',
 
which read:
 
\begin{quotation}
 
Please use ``make menuconfig'' to configure your appreciated configuration
 
for the toolchain and firmware. Please note that the default configuration is
 
what was used to build the firmware image for your router. It is advised that
 
you use this configuration.
 
\end{quotation}
 

	
 
This instruction actually exceeds GPL's requirements.
 
Specifically, the instruction guides users in their first step toward
 
exercising the freedom to modify the software.  While the GPL does contain
 
requirements that facilitate the freedom to modify (such as ensuring the CCS is
 
in the ``preferred form \ldots for making modifications to it''), GPL
 
does not require specific instructions explaining how to undertake
 
modifications.  This specific instruction therefore exemplifies
 
the exceptional quality of this particular CCS\@.
 

	
 
%FIXME: add a \hyperref to some ``preferred for for modification'' stuff above.
 

	
 
However, for purposes of the CCS verification process, typically the
 
investigator avoids any unnecessary changes to the source code during the
 
build process, lest the investigator err and cause the build to fail through
 
his own modification, and thus incorrectly identify the CCS as inadequate.
 
Therefore, the investigator proceeded to simply run:
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}[language=bash]
 
$ cd libCMC
 
$ make
 
\end{lstlisting}
 

	
 
and waited approximately 40 minutes for the build to complete\footnote{Build
 
  times will likely vary widely on various host systems.}.  The investigator
 
kept a
 
\href{https://gitorious.org/copyleft-org/tutorial/source/master:enforcement-case-studies_log-output/thinkpenguin_librecmc-complete.log}{full
 
  log of the build}, which is not included herein due its size (approximately
 
7.2K of text).
 
\label{thinkpenguin-main-build}
 

	
 
Upon completion of the ``make'' process, the investigator immediately found
 
(almost to his surprise) several large firmware files in the ``bin/ar71xx''
 
directory.  Typically, this step in the CCS verification process is
 
harrowing.  In most cases, the ``make'' step will fail due to a missing
 
package or because toolchain paths are not setup correctly.
 

	
 
In light of such experiences, the investigator speculated that ThinkPenguin's engineers did
 
the most important step in self-CCS verification: test one's own instructions
 
on a clean system.  Ideally, an employee with similar skills but
 
unfamiliar with the specific product can most easily verify CCS  and identify
 
problems before a violation occurs.
 

	
 
% FIXME: Is there stuff about the above in the compliance guide?  If so, link
 
% to it.  If not, write it, then link to it. :)
 

	
 
However, upon completing the ``make'', the investigator was unclear which
 
filesystem and kernel images to install on the TPE-NWIFIROUTER hardware.
 
Ideally, the original ``README'' would indicate which image is appropriate
 
for the included hardware.  However, this was ultimately an annoyance rather
 
than a compliance issue.  Fortunately,
 
the web UI (see next section) on the TPE-NWIFIROUTER performs firmware image
 
installation.  Additionally, the router's version number was specified on the
 
bottom of the device, which indicated which of the differently-versioned images
 
we should install.  The investigator would prefer instructions such as
 
those found at
 
\url{http://librecmc.org/librecmc/wiki?name=Tp+MR3020}{instructions similar
 
  to these} in the README itself; however, application of the reasonableness
 
standard here again indicates compliance, since a knowledgeable user can easily
 
determine the proper course of action.
 

	
 

	
 
\section{U-Boot Compilation}
 

	
 
%FIXME: link to u-boot reflash, maybe put it in log-output dir?
 

	
 
The investigator then turned his attention to the file,
 
``u-boot\verb0_0reflash''.  These instructions explained how to
 
build and install the bootloader for the device.
 

	
 
The investigator followed the instructions for compiling U-Boot, and found
 
them quite straight-forward.  The investigator discovered two minor
 
annoyances, however, while building U-Boot:
 

	
 
\begin{itemize}
 

	
 
 \item The variable \verb0$U-BOOT_SRC0 was used as a placeholder for the name
 
   of the extracted source directory.  This was easy to surmise and was not a
 
   compliance issue (per the reasonableness standard), but explicitly stating
 
   that fact at the top of the instructions would be helpful.
 

	
 
\label{thinkpenguin-glibc-214-issue}
 
\item Toolchain binaries were included and used by default by the build
 
  process.  These binaries were not the appropriate ones for the
 
  investigator's host system, and the build failed with the following error:
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}
 
mips-librecmc-linux-uclibc-gcc.bin: /lib/libc.so.6:
 
   version `GLIBC`_2.14' not found
 
   (required by mips-librecmc-linux-uclibc-gcc.bin)
 
\end{lstlisting}
 

	
 
   (The
 
\href{https://gitorious.org/copyleft-org/tutorial/source/master:enforcement-case-studies_log-output/thinkpenguin_u-boot-build_fail.log}{complete
 
  log output from the failure} is too lengthy to include herein.)
 

	
 
   This issue is an annoyance, not a compliance problem.  It was clear from
 
   context that these binaries were simply for a different host architecture, and
 
   the investigator simply removed ``toolchain/bin'' and created a symlink to
 
   utilize the toolchain already built earlier (during the compilation
 
   discussed in \S~\ref{thinkpenguin-main-build}):
 

	
 
\lstset{tabsize=2}
 
\begin{lstlisting}
 
$ ln -s \
 
  ../../staging_dir/toolchain-mips_34kc_gcc-4.6-linaro_uClibc-0.9.33.2/bin \
 
  toolchain/bin
 
\end{lstlisting}
 

	
 

	
 
   After this change, the U-Boot build completed successfully.
 
\end{itemize}
 

	
 
The
 
\href{https://gitorious.org/copyleft-org/tutorial/source/master:enforcement-case-studies_log-output/thinkpenguin_u-boot-finish_build.log}{full
 
  log of the build} is not included herein due its size (approximately 3.8K
 
of text).  After that, the investigator found a new U-Boot image in the
 
``bin'' directory.
 

	
 
\section{Root Filesystem and Kernel Installation}
 

	
 
The investigator next tested installation of the firmware.  In particular,
 
the investigator connected the TPE-NWIFIROUTER to a local network, and
 
visited \url{http://192.168.10.1/}, logged in, and chose the option sequence:
 
``System $\Rightarrow$ Backup / Flash Firmware''.
 

	
 
From there, the investigator chose the ``Flash new firmware image'' section
 
and selected the
 
``librecmc-ar71xx-generic-tl-wr841n-v8-squashfs-sysupgrade.bin'' image from
 
the ``bin/ar71xx'' directory.  The investigator chose the ``v8'' image upon
 
verifying the physical router read ``v8.2'' on its bottom.  The investigator
 
chose the ``sysupgrade'' version of the image because this was clearly a
 
system upgrade (as a firmware already came preinstalled on the
 
TPE-NWIFIROUTER).
 

	
 
Upon clicking ``Flash image\ldots'', the web interface prompted the
 
investigator to confirm the MD5 hash of the image to flash.  The investigator
 
did so, and then clicked ``Proceed'' to flash the image.  The process took
 
about one minute, at which point the web page refreshed to the login screen.
 
Upon logging in, the investigator was able to confirm in the ``Kernel Log''
 
section of the web interface that the newly built copy of Linux had indeed been
 
installed.
 

	
 
The investigator confirmed that a new version of ``busybox'' had also been
 
installed by using SSH to connect to the router and ran the command
 
``busybox'', which showed the newly-compiled version (via its date of
 
compilation).
 

	
 
%FIXME: dg: can you get me  a screen shot for the Kernel Log above, and paste
 
%in the output of running busybox ?
 
%FIXME: bkuhn: the screen shot for the Kernel Log is in the log output dir at
 
%thinkpenguin_librecmc-built-kernel_log.png and the BusyBox output is in the
 
%same directory at thinkpenguin_librecmc-built-busybox_output.log - you may want
 
%to only use part of the BusyBox output (maybe even just the login) for brevity
 

	
 
\section{U-Boot Installation}
 

	
 
The U-Boot installation process is substantially more complicated than the
 
firmware update.  The investigator purchased the optional serial cable
 
along with the TPE-NWIFIROUTER, in order to complete the U-Boot installation
 
per the instructions in ``u-boot\verb0_0reflash'' in its section ``Installing
 
u-boot to your router'', which reads:
 

	
 
\begin{quotation}
 
  \begin{enumerate}
 

	
 
    \item Install and configure any TFTP server on your PC (tftp-hpa).
 
       Set a fixed IP address on your PC \ldots and connect it to the router,
 
       using RJ45 network cable \ldots
 

	
 
 \item Connect USB to UART adapter to the router and start any application to
 
   communicate with it, like PuTTY. \ldots
 

	
 
   \item Power on the router, wait for a line like one of the following and
 
     interrupt the process of loading a kernel:
 
\begin{verbatim}
 
 Autobooting in 1 seconds
 
      (for most TP-Link routers, you should enter tpl at this point)
 
Hit ESC key to stop autoboot: 1 (for 8devices Carambola 2, use ESC key)
 
 Hit any key to stop autoboot: 1 (for D-Link DIR-505, use any key)
 
\end{verbatim}
 
\item   Set ipaddr and serverip environment variables:
 
\lstset{tabsize=2}
 
\begin{lstlisting}
 
    hornet> setenv ipaddr 192.168.1.1
 
    hornet> setenv serverip 192.168.1.2
 
\end{lstlisting}
 

	
 
  \end{enumerate}
 
\end{quotation}
 

	
 

	
 
%FIXME: image of the serial cable available anywhere to put here:
 
%  https://www.adafruit.com/images/970x728/954-02.jpg
 

	
 
The investigator used the purchased serial cable, which was a USB serial
 
adapter with a male USB type A connector to 4 female jumper wires.  The
 
female jumper wires were red, black, white, and green.
 

	
 
The instructions here were slightly incomplete, since they did not specify
 
how to connect the wires to the router.  However, the investigator found
 
general information available online at
 
\url{http://wiki.openwrt.org/toh/tp-link/tl-wr841nd#serial.console} which
 
described the proper procedure.  While the ``power'' and ``ground'' cables
 
were obvious, some trial and error was necessary to find the RX and TX
 
cables, but this was easily determined since miswiring TX and RX yields no
 
I/O and proper wiring yields the output as expected.  Using the pin gender
 
changer included with the adapter, the investigator was able to stably wire
 
the pins for use once the proper RX and TX connections were determined.
 

	
 
The investigator then used the recommended 115200 8N1 for serial console
 
settings, leaving all flow control off, and tested both with the
 
\verb0minicom0 and \verb0screen0 commands.  The investigator found that if
 
all 4 wires were connected on the USB serial adapter that the router would
 
start without additional power and the console would receive the startup
 
messages.  The investigator could replicate the same behavior by omitting the
 
power cable from the USB serial adapter (red wire) and connecting the main
 
power adapter to the router instead.
 

	
 
At this point, the on-screen messages as described in the installation
 
instructions appeared, but the investigator found that no key events sent via
 
the serial port appeared to reach the U-Boot console.  In other words, while
 
the investigator saw both U-Boot and kernel boot messages in the serial
 
console, the investigator was unable to interrupt the boot process as
 
instructed by ``u-boot\verb0_0reflash''.  Hitting a key simply did
 
\textit{not} interrupt the boot process and yield the \verb0hornet>0 prompt.
 

	
 
After additional trial and error over a period of hours, the investigator had
 
finally to consider this question for the first time during the process:
 
``Has ThinkPenguin violated the GPL?'' More specifically, the immediate
 
question was: ``Given this failure, has the distributor met
 
\hyperref[GPLv2s3-build-scripts]{the requirements for `scripts used to
 
  control \ldots installation of the executable' (GPLv2)} and
 
\hyperref[GPLv3-installation-information]{necessary `Installation
 
  Information' (GPLv3)}?''
 

	
 
The appropriate answer to the question (at this specific stage) is ``possibly,
 
but more information is needed''.  Embedded installation and configuration is
 
a tricky and complex technical process.  While the GPL requires documentation
 
and clear instructions for this process, the investigator did not immediately blame the distributor
 
for what may be an honest, correctable mistake, or an issue legitimately explained by
 
feasible alternative theories.
 

	
 
In this case, upon remembering the issues of wiring, the investigator wonder
 
if perhaps the power pin was accidentally connected for a moment to the RX
 
pin while live.  Such action could easily fry the RX pin, and yield the
 
observed behavior.  Since the investigator could not rule out the possibility
 
of accidental connection of power to the RX pin mentioned, the investigator
 
had to assume the instructions would work properly if he had not made that
 
error.
 

	
 
That conclusion, while correct, also left the investigator with only two
 
option to complete the final verification of the CCS:
 

	
 
\begin{itemize}
 

	
 
   \item Purchase a new router and cable anew, and reattempt the installation
 
     process while taking extra care not to miswire any cables.
 

	
 
   \item Seek assistance from the libreCMC community to find an alternative
 
     method of installation.
 

	
 
\end{itemize}
 

	
 
The investigator chose the latter and then contacted a libreCMC developer
 
familiar with the product.  That developer, who  agreed the the RX pin was
 
likely ruined, described an alternative method for console access using the
 
{\tt netcat}.  The libreCMC developer described the process as follows:
 

	
 
\begin{quotation}
 

	
 
  \begin{itemize}
 

	
 
  \item Change the IP address of the router to 192.168.1.1.
 

	
 
  \item Change the IP address of a desktop GNU/Linux system to 192.168.1.2.
 

	
 
  \item Power on the router while holding the reset button for 7 seconds.
 

	
 
  \item Use the {\tt netcat} command (as below) on the desktop, and press
 
    enter to receive U-Boot's prompt:
 
    
 
\lstset{tabsize=2}
 
\begin{lstlisting}[language=bash]
 
$ nc -u -p 6666 192.168.1.1 6666
 
uboot>
 
\end{lstlisting}
 
  \end{itemize}
 
\end{quotation}
 

	
 
Upon following this procedure, the investigator was able to confirm the
 
(original) shipped version of U-Boot was still installed:
 
\begin{lstlisting}[language=bash]
 
$ nc -u -p 6666 192.168.1.1 6666
 
uboot> version
 
U-Boot 1.1.4 (Jul 28 2014)
 
\end{lstlisting}
 

	
 
Thereafter, the investigator followed the instructions from
 
``u-boot\verb0_0reflash''.  Specifically, the investigator configured a TFTP server
 
and placed the newly built firmware into \texttt{/srv/tftp}.  The investigator
 
also followed the remaining instructions in ``u-boot\verb0_0reflash'', but
 
used the \texttt{netcat} console rather than the serial console, and
 
used U-Boot's \texttt{reset} command to reboot the router.
 

	
 
Upon reboot, the serial console (still connect with working output) showed
 
the message \texttt{U-Boot 1.1.4  (Oct 17 2014)}, and thus confirmed a
 
successful reflash of the U-Boot image built by the investigator.
 

	
 
\section{Firmware Comparison}
 

	
 
Next, to ensure the CCS did indeed correspond to the firmware original
 
installed on the TPE-NWIFIROUTER, the investigator compared the built
 
firmware image with the filesystem originally found on the device itself.
 
The comparison steps were as follows:
 

	
 
\begin{enumerate}
 
  
 
\item Extract the filesystem from the image we built by running
 
  \href{https://gitorious.org/copyleft-org/gpl-compliance-scripts/source/master:find-firmware.pl}{find-firmware.pl}
 
  on ``bin/ar71xx/librecmc-ar71xx-generic-tl-wr841n-v8-squashfs-factory.bin''
 
  and then running
 
  \href{http://www.binaryanalysis.org/en/content/show/download}{bat-extratools}'
 
  ``squashfs4.2/squashfs-tools/bat-unsquashfs42'' on the resulting
 
  morx0.squash, using the filesystem in the new squashfs-root directory for
 
  comparison.
 

	
 
\item Login to the router's web interface (at \url{http://192.168.10.1/ }) from a computer
 
  connected to the router.
 
  
 
\item Set a password using the provided link at the top (since the router's
 
  UI warns that no password is set and asks the user to change it).
 
  
 
\item Logged into the router via SSH, using the root user with the
 
  aforementioned password.
 
  
 
\item Compared representative directory listings and binaries to ensure the set of
 
  included files (on the router) is similar to those found in the firmware
 
  image that the investigator created (whose contents are now in the local squashfs-root directory).  In
 
  particular, the investigator did the following comparisons:
 

	
 
  \begin{enumerate}
 
  \item Listed the /bin folder (``ls -l /bin'') and confirm the list of files is the same
 
    and that the file sizes are similar.
 
    
 
  \item Checked the ``strings'' output of ``/bin/busybox'' to confirm it is similar in both
 
   places (similar number of lines and content of lines).  (One cannot directly
 
   compare the binaries because the slight compilation variations will cause
 
   some bits to be different.)
 
 \item Repeated the above two steps for ``/lib/modules'', ``/usr/bin'', and other directories with
 
   a significant number of binaries.
 
   
 
 \item Checked that the kernel was sufficiently similar.  The investigator
 
   compared the ``dmesg'' output both before and after flashing the new
 
   firmware.  As the investigator expected, the kernel version string was
 
   similar, but had a different build date and user@host indicator.  (The
 
   kernel binary itself is not easily accessible from an SSH login, but was
 
   retrievable using the U-Boot console (the start address of the kernel in
 
   flash appears to be 0x9F020000, based on the boot messages seen in the
 
   serial console).
 
  \end{enumerate}
 
\end{enumerate}
 

	
 
\section{Minor Annoyances}
 

	
 
As discussed in detail above, there were a few minor annoyances, none of
 
which were GPL violations.  Rather, the annoyances briefly impeded the
 
build and installation.  However, the investigator, as a reasonably skilled
 
build engineer for embedded devices, was able to complete the process with
 
the instructions provided.
 

	
 
To summarize, no GPL compliance issues were found, and the CCS release was
 
one of the best ever reviewed by any investigator at any community-oriented
 
enforcement organization.  However, the following annoyances were discovered:
 

	
 
\begin{itemize}
 
\item Failure to explain how to extract the source tarball and then where to run the
 
  ``make'' command.
 
\item Failure to explain how to install the kernel and root filesystem on the
 
  device; the user must assume the web UI must be used.
 

	
 
\item Including pre-built toolchain binaries that don't work on all systems,
 
  and failure to copy and/or symlink built toolchain binaries in the right location.
 

	
 
\item Failure to include information in the U-Boot installation instructions for
 
  wiring the serial cable.
 

	
 
\item Ideally, the U-Boot installation instructions would also include the
 
  {\tt netcat} method of installation.
 

	
 
\item Finally, the instructions should note that the new U-Boot firmware
 
  should be placed into \texttt{/srv/tftp} when using TFTP on most GNU/Linux
 
  desktops.
 
\end{itemize}
 

	
 
Thus, no CCS is absolutely perfect, but GPL violation investigators always
 
give the distributors the benefit of any doubts and seek to work with the
 
vendors and improve their CCS for the betterment of their users, customers,
 
and the entire software freedom community.
 

	
 
\section{Lessons Learned}
 

	
 
Companies that seek to redistribute copylefted software can benefit greatly
 
from ThinkPenguin's example.  Here are just a few of the many lessons that
 
can be learned here:
 

	
 
\begin{enumerate}
 

	
 
\item Even though copyleft licenses have them,
 
  \hyperref[thinkpenguin-included-ccs]{\bf avoid the offer-for-source
 
    provisions}.  Not only does including the CCS alongside binary
 
  distribution make violation investigation and compliance confirmation
 
  substantially easier, but also (and more importantly) doing so
 
  \hyperref[offer-for-source]{completes the distributor's CCS compliance
 
    obligations at the time of distribution} (provided, of course, that the
 
  distributor is otherwise in compliance with the relevant copyleft license).
 
  
 
\item {\bf Include top-level build instructions in a natural language (such
 
  as English) in a \hyperref[thinkpenguin-toplevel-readme]{clear and
 
    conspicuous place}.}  Copyleft licenses require that someone reasonably
 
  skilled in the art can reproduce the build and installation.  Typically,
 
  instructions written in English are necessary, and often easier than writing
 
  programmed scripts.  The ``script'' included can
 
  certainly be more like the script of a play and less like a Bash script.
 

	
 
\item {\bf Write build/install instructions to the appropriate level of
 
  specificity}.  The upstream engineers
 
  in this case study \hyperref[thinkpenguin-specific-host-system]{clearly did
 
    additional work to ensure functionality on a wide variety of host build
 
    systems}; this is quite rare.  When in doubt, include the maximum level
 
  of detail build engineers can provide with the CCS instructions, but also
 
  double-check to investigate if a more generalized solution (such as other
 
  host systems) work just as well for the build.
 

	
 
\item {\bf Seek to adhere to the spirit of copyleft, not just the letter of
 
  the license}.  Encouragement of users to improve and
 
  make their devices better is one of ThinkPenguin's commercial differentiators.  Copyleft advocates
 
  that other companies have undervalued the large and lucrative
 
  market of
 
  users who seek hackable devices.  By going beyond the
 
  mere minimal requirements of GPL, companies can immediately reap the
 
  benefits in that target market.
 

	
 
  \item Community-oriented enforcement organizations do not play ``gotcha''\footnote{For lack of a better
 
    phrase.} with distributors regarding GPL
 
    violations.  The goal in the GPL enforcement process is to achieve
 
    compliance and correct mistakes and annoyances.  Such organizations
 
    therefore take an ``innocent until proven guilty $\rightarrow$ assume guilty
 
    due to honest error rather than malicious action '' approach.  The goal
 
    is compliance (in direct contrast with
 
    the \hyperref[Proprietary Relicensing]{discussion in \S~\ref*{Proprietary Relicensing} about the
 
      proprietary relicensing} business model).
 
    
 
\end{enumerate}
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Bortez: Modified GCC SDK}
 

	
 
In our first case study, we will consider Bortez, a company that
 
produces software and hardware toolkits to assist OEM vendors, makers
 
of consumer electronic devices.
 

	
 
\section{Facts}
 

	
 
One of Bortez's key products is a Software Development Kit (``SDK'')
 
designed to assist developers building software for a specific class of
 
consumer electronics devices.
 

	
 
FSF received a report that the SDK may be based on the GNU Compiler
 
Collection (which is an FSF-copyrighted collection of tools for software
 
development in C, C++ and other popular languages). FSF investigated the
 
claim, but was unable to confirm the violation. The violation reporter
 
was unresponsive to follow-up requests for more information.
 

	
 
Since FSF was unable to confirm the violation, we did not pursue it any
 
further. Bogus reports do happen, and we do not want to burden companies
 
with specious GPL violation complaints. FSF shelved the matter until
 
more evidence was discovered.
 

	
 
FSF was later able to confirm the violation when two additional reports
 
surfaced from other violation reporters, both of whom had used the SDK
 
professionally and noticed clear similarities to FSF's GNU GCC\@. FSF's
 
Compliance Engineer asked the reporters to run standard tests to confirm
 
the violation, and it was confirmed that Bortez's SDK was indeed a
 
modified version of GCC\@. Bortez had ported to Windows and added a number
 
of features, including support for a specific consumer device chipset and
 
additional features to aid in the linking process (``LP'') for those
 
specific devices. FSF explained the rights that the GPL afforded these
 
customers and pointed out, for example, that Bortez only needed to provide
 
source to those in possession of the binaries, and that the users may need
 
to request that source (if \S 3(b) was exercised). The violators
 
confirmed that such requests were not answered.
 

	
 
FSF brought the matter to the attention of Bortez, who immediately
 
escalated the matter to their attorneys. After a long negotiation,
 
Bortez acknowledged that their SDK was indeed a modified version of
 
GCC\@. Bortez released most of the source, but some disagreement
 
occurred over whether LP was also derivative of GCC\@. After repeated
 
FSF inquiries, Bortez reaudited the source to discover that FSF's
 
analysis was correct. Bortez determined that LP included a number of
 
source files copied from the GCC code-base.
 

	
 
\label{davrik-build-problems}
 
Once the full software release was made available, FSF asked the violation
 
reporters if it addressed the problem. Reports came back that the source
 
did not properly build. FSF asked Bortez to provide better build
 
instructions with the software, and such build instructions were
 
incorporated into the next software release.
 

	
 
At FSF's request as well, Bortez informed customers who had previously
 
purchased the product that the source was now available by announcing
 
the availability on its Web site and via a customer newsletter.
 

	
 
Bortez did have some concerns regarding patents. They wished to include a
 
statement with the software release that made sure they were not granting
 
any patent permission other than what was absolutely required by the GPL\@.
 
They understood that their patent assertions could not trump any rights
 
granted by the GPL\@. The following language was negotiated into the release:
 

	
 
\begin{quotation}
 
Subject to the qualifications stated below, Bortez, on behalf of itself
 
and its Subsidiaries, agrees not to assert the Claims against you for your
 
making, use, offer for sale, sale, or importation of the Bortez's GNU
 
Utilities or derivative works of the Bortez's GNU Utilities
 
(``Derivatives''), but only to the extent that any such Derivatives are
 
licensed by you under the terms of the GNU General Public License. The
 
Claims are the claims of patents that Bortez or its Subsidiaries have
 
standing to enforce that are directly infringed by the making, use, or
 
sale of an Bortez Distributed GNU Utilities in the form it was distributed
 
by Bortez and that do not include any limitation that reads on hardware;
 
the Claims do not include any additional patent claims held by Bortez that
 
cover any modifications of, derivative works based on or combinations with
 
the Bortez's GNU Utilities, even if such a claim is disclosed in the same
 
patent as a Claim. Subsidiaries are entities that are wholly owned by
 
Bortez.
 

	
 
This statement does not negate, limit or restrict any rights you already
 
have under the GNU General Public License version 2.
 
\end{quotation}
 

	
 
This quelled Bortez's concerns about other patent licensing they sought to
 
do outside of the GPL'd software, and satisfied FSF's concerns that Bortez
 
give proper permissions to exercise teachings of patents that were
 
exercised in their GPL'd software release.
 

	
 
Finally, a GPL Compliance Officer inside Bortez was appointed to take
 
responsibility for all matters of GPL compliance inside the company.
 
Bortez is responsible for informing FSF if the position is given to
 
someone else inside the company, and making sure that FSF has direct
 
contact with Bortez's Compliance Officer.
 

	
 
\section{Lessons}
 

	
 
This case introduces a number of concepts regarding GPL enforcement.
 

	
 
\begin{enumerate}
 

	
 
\item {\bf Enforcement should not begin until the evidence is confirmed.}
 
  Most companies that distribute GPL'd software do so in compliance, and at
 
  times, violation reports are mistaken. Even with extensive efforts in
 
  GPL education, many users do not fully understand their rights and the
 
  obligations that companies have. By working through the investigation
 
  with reporters, the violation can be properly confirmed, and {\bf the
 
    user of the software can be educated about what to expect with GPL'd
 
    software}. When users and customers of GPL'd products know their
 
  rights, what to expect, and how to properly exercise their rights
 
  (particularly under \S 3(b)), it reduces the chances for user
 
  frustration and inappropriate community outcry about an alleged GPL
 
  violation.
 

	
 
\item {\bf GPL compliance requires friendly negotiation and cooperation.}
 
  Often, attorneys and managers are legitimately surprised to find out
 
  GPL'd software is included in their company's products. Engineers
 
  sometimes include GPL'd software without understanding the requirements.
 
  This does not excuse companies from their obligations under the license,
 
  but it does mean that care and patience are essential for reaching GPL
 
  compliance. We want companies to understand that participating and
 
  benefiting from a collaborative Free Software community is not a burden,
 
  so we strive to make the process of coming into compliance as smooth as
 
  possible.
 

	
 
\item {\bf Confirming compliance is a community effort.}  The whole point
 
  of making sure that software distributors respect the terms of the GPL is to
 
  allow a thriving software sharing community to benefit and improve the
 
  work. FSF is not the expert on how a compiler for consumer electronic
 
  devices should work. We therefore inform the community who originally
 
  brought the violation to our attention and ask them to assist in
 
  evaluation and confirmation of the product's compliance. Of course, FSF
 
  coordinates and oversees the process, but we do not want compliance for
 
  compliance's sake; rather, we wish to foster a cooperating community of
 
  development around the Free Software in question, and encourage the
 
  once-violator to begin participating in that community.
 

	
 
\item {\bf Informing the harmed community is part of compliance.} FSF asks
 
  violators to make some attempt --- such as via newsletters and the
 
  company's Web site --- to inform those who already have the products as
 
  to their rights under the GPL\@. One of the key thrusts of the GPL's \S 1 and
 
  \S 3 is to {\em make sure the user knows she has these rights\/}. If a
 
  product was received out of compliance by a customer, she may never
 
  actually discover that she has such rights. Informing customers, in a
 
  way that is not burdensome but has a high probability of successfully
 
  reaching those who would seek to exercise their freedoms, is essential
 
  to properly remedy the mistake.
 

	
 
\item {\bf Lines between various copyright, patent, and other legal
 
  mechanisms must be precisely defined and considered.}  The most
 
  difficult negotiation point of the Bortez case was drafting language
 
  that simultaneously protected Bortez's patent rights outside of the
 
  GPL'd source, but was consistent with the implicit patent grant in
 
  the GPL\@. As we discussed in the first course of this series, there is
 
  indeed an implicit patent grant with the GPL, thanks to \S 6 and \S 7.
 
  However, many companies become nervous and wish to make the grant
 
  explicit to assure themselves that the grant is sufficiently narrow for
 
  their needs. We understand that there is no reasonable way to determine
 
  what patent claims read on a company's GPL holdings and which do not, so
 
  we do not object to general language that explicitly narrows the patent
 
  grant to only those patents that were, in fact, exercised by the GPL'd
 
  software as released by the company.
 

	
 
\end{enumerate}
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Bracken: a Minor Violation in a GNU/Linux Distribution}
 

	
 
In this case study, we consider a minor violation made by a company whose
 
knowledge of the Free Software community and its functions is deep.
 

	
 
\section{The Facts} 
 

	
 
Bracken produces a GNU/Linux operating system product that is sold
 
primarily to OEM vendors to be placed in appliance devices used for a
 
single purpose, such as an Internet-browsing-only device. The product
 
is almost 100\% Free Software, mostly licensed under the GPL and related
 
Free Software licenses.
 

	
 
FSF found out about this violation through a report first posted on a
 
  Slashdot\footnote{Slashdot is a popular news and discussion site for
 
  technical readers.} comment, and then it was brought to our attention again
 
  by another Free Software copyright holder who had discovered the
 
  same violation.
 

	
 
Bracken's GNU/Linux product is delivered directly from their Web site.
 
This allowed FSF engineers to directly download and confirm the
 
violation quickly. Two primary problems were discovered with the
 
online distribution:
 

	
 
\begin{itemize}
 

	
 
\item No source code nor offer for source code was provided for a number
 
  of components for the distributed GNU/Linux system; only binaries were
 
  available
 

	
 
\item An End User License Agreement (``EULA'') was included that
 
  contradicted the permissions granted by the GPL\@
 

	
 
\end{itemize}
 

	
 
FSF contacted Bracken and gave them the details of the violation. Bracken
 
immediately ceased distribution of the product temporarily and set forth
 
a plan to bring themselves back into compliance. This plan included the
 
following steps:
 

	
 
\begin{itemize}
 

	
 
\item Bracken attorneys would rewrite the EULA to comply with the GPL and
 
  would vet the new EULA through FSF before use
 

	
 
\item Bracken engineers would provide source side-by-side with the
 
  binaries for the GNU/Linux distribution on the site (and on CD's, if
 
  ever they distributed that way)
 

	
 
\item Bracken attorneys would run an internal seminar for its engineers
 
  regarding proper GPL compliance to help ensure that such oversights
 
  regarding source releases would not occur in the future
 

	
 
\item Bracken would resume distribution of the product only after FSF
 
  formally restored Bracken's distribution rights
 
\end{itemize}
 

	
 
This case was completed in about a month. FSF approved the new EULA
 
text. The key portion in the EULA relating to the GPL read as follows:
 

	
 
\begin{quotation}
 
Many of the Software Programs included in Bracken Software are distributed
 
under the terms of agreements with Third Parties (``Third Party
 
Agreements'') which may expand or limit the Licensee's rights to use
 
certain Software Programs as set forth in [this EULA]. Certain Software
 
Programs may be licensed (or sublicensed) to Licensee under the GNU
 
General Public License and other similar license agreements listed in part
 
in this section which, among other rights, permit the Licensee to copy,
 
modify and redistribute certain Software Programs, or portions thereof,
 
and have access to the source code of certain Software Programs, or
 
portions thereof. In addition, certain Software Programs, or portions
 
thereof, may be licensed (or sublicensed) to Licensee under terms stricter
 
than those set forth in [this EULA]. The Licensee must review the
 
electronic documentation that accompanies certain Software Programs, or
 
portions thereof, for the applicable Third Party Agreements. To the
 
extent any Third Party Agreements require that Bracken provide rights to
 
use, copy or modify a Software Program that are broader than the rights
 
granted to the Licensee in [this EULA], then such rights shall take
 
precedence over the rights and restrictions granted in this Agreement
 
solely for such Software Programs.
 
\end{quotation}
 

	
 
FSF restored Bracken's distribution rights shortly after the work was
 
completed as described.
 

	
 
\section{Lessons Learned}
 

	
 
This case was probably the most quickly and easily resolved of all GPL
 
violations in the history of FSF's Compliance Lab. The ease with which
 
the problem was resolved shows a number of cultural factors that play a
 
role in GPL compliance.
 

	
 
\begin{enumerate}
 

	
 
\item {\bf Companies that understand Free Software culture better have an
 
  easier time with compliance.}  Bracken's products were designed and
 
  built around the GNU/Linux system and Free Software components. Their
 
  engineers were deeply familiar with the Free Software ecosystem, and
 
  their lawyers had seen and reviewed the GPL before. The violation was
 
  completely an honest mistake. Since the culture inside the company had
 
  already adapted to the cooperative style of resolution in the Free
 
  Software world, there was very little work for either party to bring the
 
  product into compliance.
 

	
 
\item {\bf When people in key positions understand the Free Software
 
  nature of their software products, compliance concerns are as
 
  mundane as minor software bugs.}  Even the most functional system or
 
  structure has its problems, and successful business often depends on
 
  agile response to the problems that do come up; avoiding problems
 
  altogether is a pipe dream. Minor GPL violations can and do happen
 
  even with well-informed redistributors. However, resolution is
 
  reached quickly when the company --- and in particular, the lawyers,
 
  managers, and engineers working on the Free Software product lines
 
  --- have adapted to Free Software culture that the lower-level
 
  engineer already understood
 

	
 
\item {\bf Legally, distribution must stop when a violation is
 
  identified.}  In our opinion, Bracken went above and beyond the call of
 
  duty by ceasing distribution while the violation was being resolved.
 
  Under GPL \S 4, the redistributor loses the right to distribute the
 
  software, and thus they are in ongoing violation of copyright law if
 
  they distribute before rights are restored. It is FSF's policy to
 
  temporarily allow distribution while compliance negotiations are ongoing
 
  and only in the most extreme cases (where the other party appears to be
 
  negotiating in bad faith) does FSF even threaten an injunction on
 
  copyright grounds. However, Bracken --- as a good Free Software citizen
 
  --- chose to be on the safe side and do the legally correct thing while
 
  the violation case was pending. From start to finish, it took less
 
  than a month to resolve. This lapse in distribution did not, to FSF's
 
  knowledge, impact Bracken's business in any way.
 

	
 
\item {\bf EULAs are a common area for GPL problems.}  Often, EULAs
 
  are drafted from boilerplate text that a company uses for all its
 
  products. Even the most diligent attorneys forget or simply do not
 
  know that a product contains software licensed under the GPL and other
 
  Free Software licenses. Drafting a EULA that accounts for such
 
  licenses is straightforward; the text quoted above works just fine.
 
  The EULA must be designed so that it does not trump rights and
 
  permissions already granted by the GPL\@. The EULA must clearly state
 
  that if there is a conflict between it and the GPL, with regard to GPL'd
 
  code, the GPL is the overriding license.
 

	
 
\item {\bf Compliance Officers are rarely necessary when companies are
 
  educated about GPL compliance.}  As we saw in the Bortez case, FSF asks
 
  that a formal ``GPL Compliance Officer'' be appointed inside a
 
  previously violating organization to shepherd the organization to a
 
  cooperative approach to GPL compliance. However, when FSF
 
  sees that an organization already has such an approach, there is no
 
  need to request that such an officer be appointed.
 

	
 
\end{enumerate}
 

	
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Vigorien: Security, Export Controls, and GPL Compliance}
 

	
 
This case study introduces how concerns of ``security through obscurity''
 
and regulatory problems can impact GPL compliance matters.
 

	
 
\section{The Facts}
 

	
 
Vigorien distributes a back-up solution product that allows system
 
administrators to create encrypted backups of file-systems on
 
Unix-like computers. The product is based on GNU tar, a backup utility
 
that replaces the standard Unix utility simply called tar, but has
 
additional features.
 

	
 
Vigorien's backup solution added cryptographic features to GNU tar, and
 
included a suite of utilities and graphical user interfaces surrounding
 
GNU tar to make backups convenient.
 

	
 
FSF discovered the violation from a user report, and determined that the
 
cryptographic features were the only part of the product that constituted
 
a derivative work of GNU tar; the extraneous utilities merely made
 
shell calls out to GNU tar. FSF requested that Vigorien come into
 
compliance with the GPL by releasing the source of GNU tar, with the
 
cryptographic modifications, to its customers.
 

	
 
Vigorien released the original GNU tar sources, but kept the cryptographic
 
modifications proprietary. They argued that the security of their system
 
depending on keeping the software proprietary and that regardless, USA
 
export restrictions on cryptographic software prohibited such a release.
 
FSF disputed the first claim, pointing out that Vigorien had only one
 
option if they did not want to release the source: they would have to
 
remove GNU tar from the software and not distribute it further. Vigorien
 
rejected this suggestion, since GNU tar was an integral part of the
 
product, and the security changes were useless without GNU tar.
 

	
 
Regarding the export control claims, FSF proposed a number of options,
 
including release of the source from one of Vigorien's divisions overseas
 
where no such restrictions occurred, but Vigorien argued that the problem
 
was insoluble because they operated primarily in the USA\@.
 

	
 
The deadlock on the second issue was resolved when those cryptographic
 
export restrictions were lifted shortly thereafter, and FSF again raised
 
the matter with Vigorien. At that point, they dropped the first claim and
 
agreed to release the remaining source module to their customers. They
 
did so, and the violation was resolved.
 

	
 

	
 
\section{Lessons Learned}
 

	
 
\begin{enumerate}
 

	
 
\item {\bf Removing the GPL'd portion of the product is always an
 
  option.}  Many violators' first response is to simply refuse to
 
  release the source code as the GPL requires. FSF offers the option to
 
  simply remove the GPL'd portions from the product and continue along
 
  without them. Every case where this has been suggested has led to
 
  the same conclusion. Like Vigorien, the violator argues that the
 
  product cannot function without the GPL'd components, and they
 
  cannot effectively replace them.
 

	
 
  Such an outcome is simply further evidence that the combined work in
 
  question is indeed a modified version of the original GPL'd component.
 
  If the other components cannot stand on their own and be useful without
 
  the GPL'd portions, then one cannot effectively argue that the work as a
 
  whole is not a based on the GPL'd portions.
 

	
 
\item {\bf The whole product is not always covered.}  In this case,
 
  Vigorien had additional works aggregated. The backup system was a suite
 
  of utilities, some of which were the GPL and some of which were not. While
 
  the cryptographic routines were tightly coupled with GNU tar and clearly
 
  made a whole new combined work of both components, the various GUI utilities were separate and
 
  independent works merely aggregated with the distribution of the
 
  GNU-tar-based product.
 

	
 

	
 
\item {\bf ``Security'' concerns do not exonerate a distributor from GPL
 
  obligations, and ``security through obscurity'' does not work anyway.}
 
  The argument that ``this is security software, so it cannot be released
 
  in source form'' is not a valid defense for explaining why the terms of
 
  the GPL are ignored. If companies do not want to release source code
 
  for some reason, then they should not base the work on GPL'd software.
 
  No external argument for noncompliance can hold weight if the work as
 
  a whole is indeed a modified version of a GPL'd program.
 

	
 
  The ``security concerns'' argument is often floated as a reason to keep
 
  software proprietary, but the computer security community has on
 
  numerous occasions confirmed that such arguments are entirely specious.
 
  Security experts have found --- since the beginnings of the field of
 
  cryptography in the ancient world --- that sharing results about systems
 
  and having such systems withstand peer review and scrutiny builds the
 
  most secure systems. While full disclosure may help some who wish to
 
  compromise security, it helps those who want to fix problems even more
 
  by identifying them early.
 

	
 
\item {\bf External regulatory problems can be difficult to resolve.}
 
  The GPL, though grounded in copyright law, does not have the power to trump
 
  regulations like export controls. While Vigorien's ``security
 
  concerns'' were specious, their export control concerns were not. It is
 
  indeed a difficult problem that FSF acknowledges. We want compliance
 
  with the GPL and respect for users' freedoms, but we certainly do not expect
 
  companies to commit criminal offenses for the sake of compliance. We
 
  will see more about this issue in our next case study.
 
\end{enumerate}
 

	
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Haxil, Polgara, and Thesulac: Mergers, Upstream Providers and Radio Devices}
 

	
 
This case study considers an ongoing (at the time of writing) violation
 
that has occurred. By the end of the investigation period, three
 
companies were involved and many complex issues arose.
 

	
 
\section{The Facts}
 

	
 
Haxil produced a consumer electronics device which included a mini
 
GNU/Linux distribution to control the device. The device was of interest
 
to many technically-minded consumers, who purchased the device and very
 
quickly discovered that Free Software was included without source.
 
Mailing lists throughout the Free Software community erupted with
 
complaints about the problem, and FSF quickly investigated.
 

	
 
FSF confirmed that FSF-copyrighted GPL'd software was included. In
 
addition, the whole distribution included GPL'd works from hundreds of
 
individual copyright holders, many of whom were, at this point, up in
 
arms about the violation.
 

	
 
Meanwhile, Haxil was in the midst of being acquired by Polgara. Polgara
 
was as surprised as everyone else to discover the product was based on
 
GPL'd software; this fact had not been part of the disclosures made during
 
acquisition. FSF contacted Haxil, Polgara, and the product managers
 
who had transitioned into the ``Haxil division'' of the newly-merged
 
Polgara company. Polgara's General Counsel's office worked with FSF on
 
the matter.
 

	
 
FSF formed a coalition with the other primary copyright holders
 
to pursue the enforcement effort on their behalf. FSF communicated
 
directly with Polgara's representatives to begin working through the
 
issues on behalf of itself and the Free Software community at large.
 

	
 
Polgara pointed out that the software distribution they used was mostly
 
contributed by an upstream provider, Thesulac, and Haxil's changes to that
 
code base were minimal. Polgara negotiated with Thesulac to obtain the
 
source, although the issue moved very slowly in the channels between
 
Polgara and Thesulac.
 

	
 
FSF encouraged a round-table meeting so that high bandwidth communication
 
could occur between FSF, Polgara and Thesulac. Polgara and Thesulac
 
agreed, and that discussion began. Thesulac provided nearly complete
 
sources to Polgara, and Polgara made a full software release on their
 
Web site. At the time of writing, that software still has some build
 
problems (similar to those that occurred with Bortez, as described in
 
Section~\ref{davrik-build-problems}). FSF continues to negotiate with
 
Polgara and Thesulac to resolve these problems, which have a clear path to
 
a solution and are expected to resolve.
 

	
 
Similar to the Vigorien case, Thesulac has regulatory concerns. In this
 
case, it is not export controls --- an issue that has since been resolved
 
--- but radio spectrum regulation. Since this consumer electronic device
 
contains a software-programmable radio transmitter, regulations in (at
 
least) the USA and Japan prohibit release of those portions of the code
 
that operate the device. Since this is a low-level programming issue, the
 
changes to operate the device form a single combined work with the kernel named
 
Linux.  A decade later, this situation remains largely unresolved.
 

	
 
\section{Lessons Learned}
 

	
 
\begin{enumerate}
 

	
 
\item {\bf Community outrage, while justified, can often make negotiation
 
  more difficult.}  FSF has a strong policy never to publicize names of
 
  GPL violators if they are negotiating in a friendly way and operating in
 
  good faith toward compliance. Most violations are honest mistakes, and
 
  FSF sees no reason to publicly admonish violators who genuinely want to
 
  come into compliance with the GPL and to work hard staying in compliance.
 

	
 
  This case was so public in the Free Software community that both Haxil's
 
  and Polgara's representatives were nearly shell-shocked by the time FSF
 
  began negotiations. There was much work required to diffuse the
 
  situation. We empathize with our community and their outrage about GPL
 
  violations, but we also want to follow a path that leads expediently
 
  to compliance. In our experience, public outcry works best as a last
 
  resort, not the first.
 

	
 
\item {\bf For software companies, GPL compliance belongs on a corporate
 
  acquisition checklist. }  Polgara was truly amazed that Haxil had used
 
  GPL'd software in a major new product line but never informed Polgara
 
  during the acquisition process. While GPL compliance is not a
 
  particularly difficult matter, it is an additional obligation that comes
 
  along with the product line. When planning mergers and joint ventures,
 
  one should include lists of GPL'd components contained in the products
 
  discussed.
 

	
 
\item {\bf Compliance problems of upstream providers do not excuse a
 
  violation for the downstream distributor.}  To paraphrase \S 6, upstream
 
  providers are not responsible for enforcing compliance of their
 
  downstream, nor are downstream distributors responsible for compliance
 
  problems of upstream providers. However, engaging in distribution of
 
  GPL'd works out of compliance is still just that: a compliance problem.
 
  When FSF carries out enforcement, we are patient and sympathetic when
 
  the problem appears to be upstream. In fact, we urge the violator to
 
  point us to the upstream provider so we may talk to them directly. In
 
  this case, we were happy to begin negotiations with Thesulac. However,
 
  Polgara still has an obligation to bring their product into compliance,
 
  regardless of Thesulac's response.
 

	
 
\item {\bf It behooves upstream providers to advise downstream
 
  distributors about compliance matters.}  FSF has encouraged Thesulac to
 
  distribute a ``good practices for GPL compliance'' document with their
 
  product. Polgara added various software components to Thesulac's
 
  product, and it is conceivable that such additions can introduce
 
  compliance. In FSF's opinion, Thesulac is in no way legally responsible
 
  for such a violation introduced by their customer, but it behooves them
 
  from a marketing standpoint to educate their customers about using the
 
  product. We can argue whether or not it is your coffee vendor's fault
 
  if you burn yourself with their product, but (likely) no one on either
 
  side would dispute the prudence of placing a ``caution: hot'' label on
 
  the cup.
 

	
 
\item {\bf FSF enforcement often avoids redundant enforcement cases from
 
  many parties.}  Most Free Software systems have hundreds of copyright
 
  holders. Some have thousands. FSF is in a unique position as one of
 
  the largest single copyright holders on GPL'd software and as a
 
  respected umpire in the community, neutrally enforcing the rules of the
 
  GPL road. FSF works hard in the community to convince copyright
 
  holders that consolidating GPL claims through FSF is better for them,
 
  and more likely to yield positive compliance results.
 

	
 
  A few copyright holders engage in the ``proprietary relicensing''
 
  business, so they use GPL enforcement as a sales channel for that
 
  business. FSF, as a community-oriented, not-for-profit organization,
 
  seeks only to preserve the freedom of Free Software in its enforcement
 
  efforts. As it turns out, most of the community of copyright holders
 
  of Free Software want the same thing. Share and share alike is a
 
  simple rule to follow, and following that rule to FSF's satisfaction
 
  usually means you are following it to the satisfaction of the entire
 
  Free Software community.
 

	
 
\end{enumerate}
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 

	
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
% COMMENT OUT THIS CHAPTER.
 
% FIXME: is this material moot now that we include the compliance guide?
 
% Either way, it should be merged into compliance guide.
 
%\chapter{Good Practices for Compliance}
 

	
 
Generally, from the experience of GPL enforcement, we glean the following
 
general practices that can help in GPL compliance for organizations that
 
distribute products based on GPL'd software:
 

	
 
\begin{itemize}
 

	
 
\item Talk to your software engineers and ask them where they got the
 
  components they use in the products they build. Find out if GPL'd
 
  components are present.
 

	
 
\item Teach your engineering staff to pay attention to license documents.
 
  Give them easy-to-follow policies to get approval for using a Free
 
  Software component.
 

	
 
\item Build a ``Free Software Licensing'' committee that handles requests
 
  and questions about the GPL and other Free Software licenses.
 

	
 
\item Add ``What parts of your products are under the GPL or other Free
 
  Software licenses?'' to your checklist of questions to ask when you
 
  consider mergers, acquisitions, or joint ventures.
 

	
 
\item Encourage your engineers to participate collaboratively with GPL'd
 
  software development. The more knowledge about the Free Software world
 
  your organization has, the better equipped it is to deal with this
 
  rapidly changing field.
 

	
 
\item When someone points out a potential GPL violation in one of your
 
  products, do not assume the product line is doomed. The GPL is not a virus;
 
  merely having GPL'd code in one part of a product does not necessarily
 
  mean that every related product must also be GPL'd. And, even if some
 
  software needs to be released that was not before, the product will
 
  surely survive. In FSF's enforcement efforts, we have not yet
 
  seen a product line die because source was released to customers in
 
  compliance with the GPL.
 

	
 
\end{itemize}
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
% LocalWords:  proprietarize redistributors sublicense yyyy Gnomovision EULAs
 
% LocalWords:  Yoyodyne FrontPage improvers Berne copyrightable Stallman's GPLs
 
% LocalWords:  Lessig Lessig's UCITA pre PDAs CDs reshifts GPL's Gentoo glibc
 
% LocalWords:  TrollTech administrivia LGPL's MontaVista OpenTV Mitek Arce DVD
 
% LocalWords:  unprotectable protectable Unfreedonia chipset CodeSourcery Iqtel
 
% LocalWords:  impermissibly Bateman faire minimis Borland uncopyrightable Mgmt
 
% LocalWords:  franca downloadable Bortez Bortez's Gingerich Michlmayr LGPL
 
% LocalWords:  Slashdot sublicensed Vigorien Vigorien's Haxil Polgara GPL'd
 
% LocalWords:  Thesulac Polgara's Haxil's Thesulac's SDK CD's tpb CCS RYF cd
 
%%  LocalWords:  ThinkPenguin TPE NWIFIROUTER Unboxing copylefted GPLv mkdir
 
%%  LocalWords:  Filesystem libreCMC README tabsize libCMC sudo reflash gcc
 
%%  LocalWords:  binutils bzip perl getopt libz dev libc menuconfig amd dpkg
 
%%  LocalWords:  toolchain filesystem thinkpenguin egrep posix jxpf ar UI ln
 
%%  LocalWords:  ThinkPenguin's versioned bootloader SRC mips librecmc linux
 
%%  LocalWords:  uclibc symlink tl wr squashfs sysupgrade preinstalled TFTP
 
%%  LocalWords:  busybox tftp hpa IP RJ UART PuTTY ipaddr serverip setenv nc
 
%%  LocalWords:  miswiring minicom startup miswire netcat uboot extratools
 
%%  LocalWords:  morx Login dmesg login ccs toplevel readme differentiators
 
%%  LocalWords:  hackable Relicensing relicensing toolkits OEM reaudited
 
%%  LocalWords:  redistributor cryptographic
gpl-lgpl.tex
Show inline comments
 
% gpl-lgpl.tex                                                  -*- LaTeX -*-
 
%      Tutorial Text for the Detailed Study and Analysis of GPL and LGPL course
 
%
 

	
 
% License: CC-By-SA-4.0
 

	
 
% The copyright holders hereby grant the freedom to copy, modify, convey,
 
% Adapt, and/or redistribute this work under the terms of the Creative
 
% Commons Attribution Share Alike 4.0 International License.
 

	
 
% This text is distributed in the hope that it will be useful, but
 
% WITHOUT ANY WARRANTY; without even the implied warranty of
 
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 

	
 
% You should have received a copy of the license with this document in
 
% a file called 'CC-By-SA-4.0.txt'.  If not, please visit
 
% https://creativecommons.org/licenses/by-sa/4.0/legalcode to receive
 
% the license text.
 

	
 
% FIXME-LATER: I should make a macro like the Texinfo @xref stuff for places
 
%      where I'm saying ``see section X in this tutorial'', so that the extra
 
%      verbiage isn't there in the HTML versions that I'll eventually do.
 
%      Maybe something like that already exists?  In the worst case, I could
 
%      adapt @xref from texinfo.texi for it.
 

	
 
\newcommand{\defn}[1]{\emph{#1}}
 

	
 
\part{Detailed Analysis of the GNU GPL and Related Licenses}
 
\label{gpl-lgpl-part}
 

	
 
{\parindent 0in
 
\tutorialpartsplit{``Detailed Analysis of the GNU GPL and Related Licenses''}{This part} is: \\
 
\begin{tabbing}
 
Copyright \= \copyright{} 2003--2007, 2014 \hspace{.1mm} \=  \kill
 
Copyright \> \copyright{} 2014 \> Bradley M. Kuhn \\
 
Copyright \> \copyright{} 2014 \>  Anthony K. Sebro, Jr. \\
 
Copyright \> \copyright{} 2003--2007, 2014 \>  Free Software Foundation, Inc. \\
 
Copyright \> \copyright{} 2014 \>  Software Freedom Law Center.
 
\end{tabbing}
 

	
 

	
 
\vspace{.2in}
 

	
 
\begin{center}
 

	
 
The copyright holders of \tutorialpartsplit{``Detailed Analysis of the GNU GPL and Related Licenses''}{this part} hereby grant the freedom to copy, modify,
 
convey, Adapt, and/or redistribute this work under the terms of the Creative
 
Commons Attribution Share Alike 4.0 International License.  A copy of that
 
license is available at
 
\verb=https://creativecommons.org/licenses/by-sa/4.0/legalcode=.
 
\end{center}
 

	
 
\vfill
 

	
 
This part includes material from many sources, including the following
 
This part includes material from many sources, including some material from the following
 
CC-By-SA-licensed published works: \\
 

	
 
\begin{itemize}
 
\item \textit{Detailed Analysis of the GNU GPL and Related Licenses}, written by
 
Bradley M. Kuhn, Daniel B.~Ravicher, and John Sullivan and published by the Free Software Foundation for its CLE courses on 2004-01-20,
 
2004-08-24, and 2014-03-24.
 
\item \hrefnofollow{http://gplv3.fsf.org/gpl-rationale-2006-01-16.html}{\textit{GPLv3 First Discussion Draft Rationale}}, written and published by the Free
 
  Software Foundation on 2006-01-16.
 
\item \hrefnofollow{http://gplv3.fsf.org/opinions-draft-2.html}{\textit{GPLv3 Second Discussion Draft Rationale}}, written and published by the Free
 
  Software Foundation circa 2006-07.
 
\item \hrefnofollow{http://gplv3.fsf.org/gpl3-dd3-guide}{\textit{GPLv3 Third Discussion Draft Rationale}}, written and published by the Free
 
  Software Foundation on   2007-03-28.
 
\item \hrefnofollow{http://gplv3.fsf.org/dd3-faq}{\textit{GPLv3  Discussion Draft 3 FAQ}}, written and published by the Free1 Software Foundation on   2007-03-28.
 
\item \hrefnofollow{http://gplv3.fsf.org/gpl3-dd4-guide.html}{\textit{GPLv3 Final Discussion Draft Rationale}} written and published by the Free
 
  Software Foundation onon 2007-05-31.
 
\item \hrefnofollow{http://www.gnu.org/licences/gpl3-final-rationale.pdf}{\textit{GPLv3 Final Rationale}}, written and published by the Free
 
  Software Foundation on 2007-06-29.
 
  
 
\end{itemize}
 

	
 
However, this work is primarily composed of the many contributions it
 
receives as a public, collaborative project.  Please
 
\href{https://gitorious.org/copyleft-org/tutorial/history/master:gpl-lgpl.tex}{review
 
  its Git logs} for full documentation of all contributions.
 
}
 
\pagebreak
 

	
 
\tutorialpartsplit{This tutorial}{This part of the tutorial} gives a
 
comprehensive explanation of the most popular Free Software copyright
 
license, the GNU General Public License (``GNU GPL'', or sometimes just
 
``GPL'') -- both version 2 (``GPLv2'') and version 3 (``GPLv3'') -- and
 
teaches lawyers, software developers, managers and business people how to use
 
the GPL (and GPL'd software) successfully both as a community-building
 
``Constitution'' for a software project, and to incorporate copylefted
 
software into a new Free Software business and in existing, successful
 
enterprises.
 

	
 
To benefit from this part of the tutorial, readers should
 
have a general familiarity with software development processes.  A basic
 
understanding of how copyright law applies to software is also helpful.  The
 
tutorial is of most interest to lawyers, software developers and managers who
 
run or advise software businesses that modify and/or redistribute software
 
under the terms of the GNU GPL (or who wish to do so in the future), and those
 
who wish to make use of existing GPL'd software in their enterprise.
 

	
 
Upon completion of this part of the tutorial, readers can expect
 
to have learned the following:
 

	
 
\begin{itemize}
 

	
 
  \item The freedom-defending purpose of various terms in the GNU GPLv2 and GPLv3.
 

	
 
  \item The differences between GPLv2 and GPLv3.
 

	
 
  \item The redistribution options under the GPLv2 and GPLv3.
 

	
 
  \item The obligations when modifying GPLv2'd or GPLv3'd software.
 

	
 
  \item How to build a plan for proper and successful compliance with the GPL.
 

	
 
  \item The business advantages that the GPL provides.
 

	
 
  \item The most common business models used in conjunction with the GPL.
 

	
 
  \item How existing GPL'd software can be used in existing enterprises.
 

	
 
  \item The basics of LGPLv2.1 and LGPLv3, and how they
 
    differ from the GPLv2 and GPLv3, respectively.
 

	
 
  \item The basics to begin understanding the complexities regarding
 
    derivative and combined works of software.
 
\end{itemize}
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
% END OF ABSTRACTS SECTION
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
% START OF DAY ONE COURSE
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 

	
 
\chapter{What Is Software Freedom?}
 

	
 
Study of the GNU General Public License (herein, abbreviated as \defn{GNU
 
  GPL} or just \defn{GPL}) must begin by first considering the broader world
 
of software freedom. The GPL was not created in a vacuum. Rather, it was
 
created to embody and defend a set of principles that were set forth at the
 
founding of the GNU project and the Free Software Foundation (FSF) -- the
 
preeminent organization that upholds, defends and promotes the philosophy of software
 
freedom. A prerequisite for understanding both of the popular versions
 
of the GPL
 
(GPLv2 and GPLv3) and their terms and conditions is a basic understanding of
 
the principles behind them.  The GPL family of licenses are unlike nearly all
 
other software licenses in that they are designed to defend and uphold these
 
principles.
 

	
 
\section{The Free Software Definition}
 
\label{Free Software Definition}
 

	
 
The Free Software Definition is set forth in full on FSF's website at
 
\verb0http://fsf.org/0 \verb0philosophy/free-sw.html0. This section presents
 
an abbreviated version that will focus on the parts that are most pertinent
 
to the GPL\@.
 

	
 
A particular user has software freedom with respect to a particular program if that
 
user has the following freedoms:
 

	
 
\begin{itemize}
 

	
 

	
 
\item The freedom to run the program, for any purpose.
 

	
 
\item The freedom to study how the program works, and modify it
 

	
 
\item The freedom to redistribute copies.
 

	
 
\item The freedom to distribute copies of  modified versions to others.
 

	
 
\end{itemize}
 

	
 
The focus on ``a particular user'' is particularly pertinent here.  It is not
 
uncommon for a subset of a specific program's user base to have these freedoms, while other
 
users of the same version the program have none or only some of these freedoms.
 
Section~\ref{Proprietary Relicensing} talks in detail about how
 
this can unfortunately happen even if a program is released under the GPL\@.
 

	
 
Many people refer to software with these freedoms as ``Open Source.''
 
Besides having a different political focus than those who call it Free
 
Software,\footnote{The political differences between the Free Software
 
  Movement and the Open Source Movement are documented on FSF's Web site at
 
  \url{http://www.fsf.org/licensing/essays/free-software-for-freedom.html}.}
 
Those who call the software ``Open Source'' are often focused on a side
 
issue.  Specifically, user access to the source code of a program is a
 
prerequisite to make use of the freedom to modify.  However, the important
 
issue is what freedoms are granted in the license of that source code.
 

	
 
Software freedom is only complete when no restrictions are imposed on how
 
these freedoms are exercised.  Specifically, users and programmers can
 
exercise these freedoms noncommercially or commercially.  Licenses that grant
 
these freedoms for noncommercial activities but prohibit them for commercial
 
activities are considered non-free.  Even the Open Source Initiative
 
(\defn{OSI}) (the arbiter of what is considered ``Open Source'') also rules
 
such licenses not in fitting with its ``Open Source Definition''.
 

	
 
In general, software for which any of these freedoms are
 
restricted in any way is called ``non-Free Software.''  Typically, the
 
term ``proprietary software'' is used more or less interchangeably with
 
``non-Free Software.''  Personally, I tend to use the term ``non-Free
 
Software'' to refer to noncommercial software that restricts freedom
 
(such as ``shareware'') and ``proprietary software'' to refer to
 
commercial software that restricts freedom (such as nearly all of
 
Microsoft's and Oracle's offerings).
 

	
 
Keep in mind that none of the terms ``software freedom'', ``open source''
 
and ``free software'' are known to be trademarked or otherwise legally
 
restricted by any organization in
 
any jurisdiction.  As such, it's quite common that these terms are abused and
 
misused by parties who wish to bank on the popularity of software freedom.
 
When one considers using, modifying or redistributing a software package that
 
purports to be Open Source or Free Software, one \textbf{must} verify that
 
the license grants software freedom.
 

	
 
Furthermore, throughout this text, we generally prefer the term ``software
 
freedom'', as this is the least ambiguous term available to describe software
 
that meets the Free Software Definition.  For example, it is well known and
 
often discussed that the adjective ``free'' has two unrelated meanings in
 
English: ``free as in freedom'' and ``free as in price''.  Meanwhile, the
 
term ``open source'' is even more confusing, because it appears to refer only to the
 
``freedom to study'', which is merely a subset of one of the four freedoms.
 

	
 
The remainder of this section considers each of each component of software
 
freedom in detail.
 

	
 
\subsection{The Freedom to Run}
 
\label{freedom-to-run}
 

	
 
The first tenet of software freedom is the user's fully unfettered right to
 
run the program.  The software's license must permit any conceivable use of
 
the software.  Perhaps, for example, the user has discovered an innovative
 
use for a particular program, one that the programmer never could have
 
predicted.  Such a use must not be restricted.
 

	
 
It was once rare that this freedom was restricted by even proprietary
 
software; but such is quite common today. Most End User License Agreements
 
(EULAs) that cover most proprietary software typically restrict some types of
 
uses.  Such restrictions of any kind are an unacceptable restriction on
 
software freedom.
 

	
 
\subsection{The Freedom to Change and Modify}
 

	
 
Perhaps the most useful right of software freedom is the users' right to
 
change, modify and adapt the software to suit their needs.  Access to the
 
source code and related build and installation scripts are an essential part
 
of this freedom.  Without the source code, and the ability to build and
 
install the binary applications from that source, users cannot effectively
 
exercise this freedom.
 

	
 
Programmers directly benefit from this freedom.  However, this freedom
 
remains important to users who are not programmers.  While it may seem
 
counterintuitive at first, non-programmer users often exercise this freedom
 
indirectly in both commercial and noncommercial settings.  For example, users
 
often seek noncommercial help with the software on email lists and in user
 
groups.  To make use of such help they must either have the freedom to
 
recruit programmers who might altruistically assist them to modify their
 
software, or to at least follow rote instructions to make basic modifications
 
themselves.
 

	
 
More commonly, users also exercise this freedom commercially.  Each user, or
 
group of users, may hire anyone they wish in a competitive free market to
 
modify and change the software.  This means that companies have a right to
 
hire anyone they wish to modify their Free Software.  Additionally, such
 
companies may contract with other companies to commission software
 
modifications.
 

	
 
\subsection{The Freedom to Copy and Share}
 

	
 
Users share Free Software in a variety of ways. Software freedom advocates
 
work to eliminate a fundamental ethical dilemma of the software age: choosing
 
between obeying a software license and friendship (by giving away a copy of a
 
program to your friend who likes the software you are using). Licenses that
 
respect software freedom, therefore, permit altruistic sharing of software
 
among friends.
 

	
 
The commercial environment also benefits from this freedom.  Commercial sharing
 
includes selling copies of Free Software: that is, Free Software can
 
be distributed for any monetary
 
price to anyone.  Those who redistribute Free Software commercially also have
 
the freedom to selectively distribute (i.e., you can pick your customers) and
 
to set prices at any level that redistributor sees fit.
 

	
 
Of course, most people get copies of Free Software very cheaply (and
 
sometimes without charge).  The competitive free market of Free Software
 
tends to keep prices low and reasonable.  However, if someone is willing to
 
pay billions of dollars for one copy of the GNU Compiler Collection, such a
 
sale is completely permitted.
 

	
 
Another common instance of commercial sharing is service-oriented
 
distribution.  For example, some distribution vendors provide immediate
 
security and upgrade distribution via a special network service.  Such
 
distribution is not necessarily contradictory with software freedom.
 

	
 
(Section~\ref{Business Models} of this tutorial talks in detail about some
 
common Free Software business models that take advantage of the freedom to
 
share commercially.)
 

	
 
\subsection{The Freedom to Share Improvements}
 

	
 
The freedom to modify and improve is somewhat empty without the freedom to
 
share those improvements.  The software freedom community is built on the
 
pillar of altruistic sharing of improved Free Software. Historically
 
it was typical for a
 
Free Software project to sprout a mailing list where improvements
 
would be shared
 
freely among members of the development community.\footnote{This is still
 
commonly the case, though today there are additional ways of
 
sharing Free Software.}  Such noncommercial
 
sharing is the primary reason that Free Software thrives.
 

	
 
Commercial sharing of modified Free Software is equally important.
 
For commercial support to exist in a competitive free market, all
 
developers -- from single-person contractors to large software
 
companies -- must have the freedom to market their services as
 
augmenters of Free Software.  All forms of such service marketing must
 
be equally available to all.
 

	
 
For example, selling support services for Free Software is fully
 
permitted. Companies and individuals can offer themselves as ``the place
 
to call'' when software fails or does not function properly.  For such a
 
service to be meaningful, the entity offering that service needs the
 
right to modify and improve the software for the customer to correct any
 
problems that are beyond mere user error.
 

	
 
Software freedom licenses also permit any entity to distribute modified
 
versions of Free Software.  Most Free Software programs have a ``standard
 
version'' that is made available from the primary developers of the software.
 
However, all who have the software have the ``freedom to fork'' -- that is,
 
make available nontrivial modified versions of the software on a permanent or
 
semi-permanent basis.  Such freedom is central to vibrant developer and user
 
interaction.
 

	
 
Companies and individuals have the right to make true value-added versions
 
of Free Software.  They may use freedom to share improvements to
 
distribute distinct versions of Free Software with different functionality
 
and features.  Furthermore, this freedom can be exercised to serve a
 
disenfranchised subset of the user community.  If the developers of the
 
standard version refuse to serve the needs of some of the software's
 
users, other entities have the right to create a long- or short-lived fork
 
to serve that sub-community.
 

	
 
\section{How Does Software Become Free?}
 

	
 
The previous section set forth key freedoms and rights that are referred to
 
as ``software freedom''.  This section discusses the licensing mechanisms
 
used to enable software freedom.  These licensing mechanisms were ultimately
 
created as a community-oriented ``answer'' to the existing proprietary
 
software licensing mechanisms.  Thus, first, consider carefully why
 
proprietary software exists in the first place.
 

	
 
\label{explaining-copyright}
 

	
 
The primary legal regime that applies to software is copyright law.
 
Proprietary software exists at all only because copyright law governs
 
software.\footnote{This statement is admittedly an oversimplification. Patents and
 
  trade secrets can cover software and make it effectively non-Free, and one
 
  can contract away their rights and freedoms regarding software, or source
 
  code can be practically obscured in binary-only distribution without
 
  reliance on any legal system.  However, the primary control mechanism for
 
  software is copyright, and therefore this section focuses on how copyright
 
  restrictions make software proprietary.} Copyright law, with respect to
 
software, typically governs copying, modifying, and redistributing that
 
software (For details of this in the USA, see
 
\href{http://www.copyright.gov/title17/92chap1.html#106}{\S~106} and
 
\href{http://www.copyright.gov/title17/92chap1.html#117}{\S~117} of
 
\href{http://www.law.cornell.edu/uscode/text/17}{Title 17} of the
 
\textit{United States Code}).\footnote{Copyright law in general also governs
 
  ``public performance'' of copyrighted works. There is no generally agreed
 
  definition for public performance of software and both GPLv2 and GPLv3 do
 
  not restrict public performance.} By law (in the USA and in most other
 
jurisdictions), the copyright holder (most typically, the author) of the work controls
 
how others may copy, modify and/or distribute the work. For proprietary
 
software, these controls are used to prohibit these activities. In addition,
 
proprietary software distributors further impede modification in a practical
 
sense by distributing only binary code and keeping the source code of the
 
software secret.
 

	
 
Copyright is not a natural state, it is a legal construction. In the USA, the
 
Constitution permits, but does not require, the creation of copyright law as
 
federal legislation.  Software, since it is an ``original work of authorship
 
fixed in any tangible medium of expression ...  from which they can be
 
perceived, reproduced, or otherwise communicated, either directly or with the
 
aid of a machine or device'' (as stated in
 
\href{http://www.law.cornell.edu/uscode/text/17/102}{17 USC \S~102}), is thus
 
covered by the statute, and is copyrighted by default.
 

	
 
However, software, in its natural state without copyright, is Free
 
Software. In an imaginary world with no copyright, the rules would be
 
different. In this world, when you received a copy of a program's source
 
code, there would be no default legal system to restrict you from sharing it
 
with others, making modifications, or redistributing those modified
 
versions.\footnote{Note that this is again an oversimplification; the
 
  complexities with this argument are discussed in
 
  Section~\ref{software-and-non-copyright}.}
 

	
 
Software in the real world is copyrighted by default and is automatically
 
covered by that legal system.  However, it is possible to move software out
 
of the domain of the copyright system.  A copyright holder can often
 
\defn{disclaim} their copyright. (For example, under USA copyright law
 
it is possible for a copyright holder to engage in conduct resulting
 
in abandonment of copyright.)  If copyright is disclaimed, the software is
 
effectively no longer restricted by copyright law.   Software not restricted by copyright is in the
 
``public domain.''
 

	
 
\subsection{Public Domain Software}
 

	
 
In the USA and other countries that
 
are parties to the Berne Convention on Copyright, software is copyrighted
 
automatically by the author when she fixes the software in a tangible
 
medium.  In the software world, this usually means typing the source code
 
of the software into a file.
 

	
 
Imagine if authors could truly disclaim those default controls of copyright
 
law.  If so, the software is in the public domain --- no longer covered by
 
copyright.  Since copyright law is the construction allowing for most
 
restrictions on software (i.e., prohibition of copying, modification, and
 
redistribution), removing the software from the copyright system usually
 
yields software freedom for its users.
 

	
 
Carefully note that software truly in the public domain is \emph{not} licensed
 
in any way.  It is confusing to say software is ``licensed for the
 
public domain,'' or any phrase that implies the copyright holder gave
 
express permission to take actions governed by copyright law.
 

	
 
Copyright holders who state that they are releasing their code into
 
the public domain are effectively renouncing copyright controls on
 
the work.  The law gave the copyright holders exclusive controls over the
 
work, and they chose to waive those controls.  Software that is, in
 
this sense, in the public domain
 
is conceptualized by the developer as having no copyright and thus no license. The software freedoms discussed in
 
Section~\ref{Free Software Definition} are all granted because there is no
 
legal system in play to take them away.
 

	
 
Admittedly, a discussion of public domain software is an oversimplified
 
example.  
 
Because copyright controls are usually automatically granted and because, in
 
some jurisdictions, some copyright controls cannot be waived (see
 
Section~\ref{non-usa-copyright} for further discussion), many copyright
 
holders sometimes incorrectly believe a work has been placed in the public
 
domain.  Second, due to aggressive lobbying by the entertainment industry,
 
the ``exclusive Right'' of copyright, that was supposed to only exist for
 
``Limited Times'' according to the USA Constitution, appears to be infinite:
 
simply purchased on the installment plan rather than in whole.  Thus, we must
 
assume no works of software will fall into the public domain merely due to
 
the passage of time.
 

	
 
Nevertheless, under USA law it is likely that the typical
 
disclaimers of copyright or public domain dedications we see in the
 
Free Software world would be interpreted by courts as copyright
 
abandonment, leading to a situation in which the user effectively receives a
 
maximum grant of copyright freedoms, similar to a maximally-permissive
 
Free Software license.
 

	
 
The best example of software known to truly be in the public domain is software
 
that is published by the USA government.  Under
 
\href{http://www.law.cornell.edu/uscode/text/17/105}{17 USC 101 \S~105}, all
 
works published by the USA Government are not copyrightable in the USA.
 

	
 
\subsection{Why Copyright Free Software?}
 

	
 
If simply disclaiming copyright on software yields Free Software, then it
 
stands to reason that putting software into the public domain is the
 
easiest and most straightforward way to produce Free Software. Indeed,
 
some major Free Software projects have chosen this method for making their
 
software Free. However, most of the Free Software in existence \emph{is}
 
copyrighted. In most cases (particularly in those of FSF and the GNU
 
Project), this was done due to very careful planning.
 

	
 
Software released into the public domain does grant freedom to those users
 
who receive the standard versions on which the original author disclaimed
 
copyright. However, since the work is not copyrighted, any nontrivial
 
modification made to the work is fully copyrightable.
 

	
 
Free Software released into the public domain initially is Free, and
 
perhaps some who modify the software choose to place their work into the
 
public domain as well. However, over time, some entities will choose to
 
proprietarize their modified versions. The public domain body of software
 
feeds the proprietary software. The public commons disappears, because
 
fewer and fewer entities have an incentive to contribute back to the
 
commons. They know that any of their competitors can proprietarize their
 
enhancements. Over time, almost no interesting work is left in the public
 
domain, because nearly all new work is done by proprietarization.
 

	
 
A legal mechanism is needed to redress this problem. FSF was in fact
 
originally created primarily as a legal entity to defend software freedom,
 
and that work of defending software freedom is a substantial part of
 
its work today. Specifically because of this ``embrace, proprietarize and
 
extend'' cycle, FSF made a conscious choice to copyright its Free Software,
 
and then license it under ``copyleft'' terms. Many, including the
 
developers of the kernel named Linux, have chosen to follow this paradigm.
 

	
 
\label{copyleft-definition}
 

	
 
Copyleft is a strategy of utilizing copyright law to pursue the policy goal
 
of fostering and encouraging the equal and inalienable right to copy, share,
 
modify and improve creative works of authorship.  Copyleft (as a general
 
term) describes any method that utilizes the copyright system to achieve the
 
aforementioned goal.  Copyleft as a concept is usually implemented in the
 
details of a specific copyright license, such as the
 
\hyperref[GPLv3-full-text]{GNU General Public License (GPL)} and the Creative
 
Commons Attribution Share Alike License (the latter of which is the license
 
of this work itself).  Copyright holders of creative work can unilaterally
 
implement these licenses for their own works to build communities that
 
collaboratively share and improve those copylefted creative works.
 

	
 
Copyleft uses functional parts of the copyright system to achieve an unusual
 
result (legal protection for free sharing). Copyleft modifies, or ``hacks''
 
copyright law, which is usually employed to strengthen the rights of authors
 
or publishers, to strengthen instead the rights of users.  Thus, Copyleft is
 
a legal strategy and mechanism to defend, uphold and propagate software
 
freedom. The basic technique of copyleft is as follows: copyright the
 
software, license it under terms that give all the software freedoms, but use
 
the copyright law controls to ensure that all who receive a copy of the
 
software have equal rights and freedom. In essence, copyleft grants freedom,
 
but forbids others to forbid that freedom to anyone else along the
 
distribution and modification chains.
 

	
 
Copyleft's ``reciprocity'' or ``share and share alike'' rule protects both
 
developers, who avoid facing a ``prioritized'' competitor of their project,
 
and users, who can be sure that they will have all four software freedoms ---
 
not only in the present version of the program they use, but in all its
 
future improved versions.
 

	
 
Copyleft is a general concept. Much like ideas for what a computer might
 
do must be \emph{implemented} by a program that actually does the job, so
 
too must copyleft be implemented in some concrete legal structure.
 
``Share and share alike'' is a phrase that is used often enough to explain the
 
concept behind copyleft, but to actually make it work in the real world, a
 
true implementation in legal text must exist, written as a ``copyright
 
license''.  The GPL implements the concept of copyleft for software-oriented
 
and other functional works of a technical nature.  The ``CC BY SA'' license
 
implements copyleft for works of textual, musical and visual authorship, such
 
as this tutorial.
 

	
 
Copyleft advocates often distinguish between the concept of a ``strong
 
copyleft'' or a ``weak copyleft''.  However, ``strong vs. weak'' copyleft is
 
not a dichotomy, it's a spectrum.  The strongest copylefts strive to the
 
exclusive rights that copyright  grants to authors as extensively as possible
 
to maximize software freedom.  As a copyleft gets ``weaker'', the copyleft
 
license typically makes ``trade offs'' that might impede software freedom,
 
but reach other tactic goals for the community of users and developers of the
 
work.
 

	
 
In other words, strong copyleft licenses place the more requirements on how
 
``the work'' is licensed.  The unit of copyright law is ``the work''.  In
 
that sense, the ``work'' referenced by the licenses is anything that can be
 
copyrighted or will be subject to the terms of copyright law.  Strong
 
copyleft licenses exercise their scope fully.  Anything which is ``a work''
 
or a ``work based on a work'' licensed under a strong copyleft is subject to
 
its requirements, including the requirement of complete, corresponding source
 
code\footnote{Copyleft communities' use of the term ``strong copyleft'' is
 
  undoubtedly imprecise.  For example, most will call the GNU GPL a ``strong
 
  copyleft'' license, even though the GPL itself has various exceptions, such
 
  as the \hyperref[GPLv3-system-library-exception]{GPLv3's system library
 
    exception} written into the text of the license itself.  Furthermore, the
 
  copyleft community continues to debate where the a license cross the line
 
  from ``strong copyleft'' to ``license that fails to respect software
 
  freedom'', although ultimately these debates are actually regarding whether
 
  the license fits \hyperref[Free Software Definition]{Free Software
 
    definition} at all.}.  Thus, copyleft licenses, particularly strong ones,
 
seek to ensure the same license covers every version of ``work based on the
 
work'', as recognized by local copyright law, and thereby achieve the
 
specific strategic policy aim of ensuring software freedom for all users,
 
developers, authors, and readers who encounter the copylefted work.
 

	
 
\subsection{Software and Non-Copyright Legal Regimes}
 
\label{software-and-non-copyright}
 

	
 
The use, modification and distribution of software, like many endeavors,
 
simultaneously interacts with multiple different legal regimes.  As was noted
 
early via footnotes, copyright is merely the \textit{most common way} to
 
restrict users' rights to copy, share, modify and/or redistribute software.
 
However, proprietary software licenses typically use every mechanism
 
available to subjugate users.  For example:
 

	
 
\begin{itemize}
 

	
 
\item Unfortunately, despite much effort by many in the software freedom
 
  community to end patents that read on software (i.e., patents on
 
  computational ideas), they still exist.  As such, a software
 
  program might otherwise seem to be unrestricted, but a patent might read on
 
  the software and ruin everything for its users.\footnote{See
 
  \S\S~\ref{gpl-implied-patent-grant},~\ref{GPLv2s7},~\ref{GPLv3s11} for more
 
  discussion on how the patent system interacts with copyleft, and read
 
  Richard M.~Stallman's essay,
 
  \href{http://www.wired.com/opinion/2012/11/richard-stallman-software-patents/}{\textit{Let's
 
      Limit the Effect of Software Patents, Since We Can't Eliminate Them}}
 
  for more information on the problems these patents present to society.}
 

	
 
\item Digital Restrictions Management (usually called \defn{DRM}) is often
 
  used to impose technological restrictions on users' ability to exercise
 
  software freedom that they might otherwise be granted.\footnote{See
 
    \S~\ref{GPLv3-drm} for more information on how GPL deals with this issue.}
 
  The simplest (and perhaps oldest) form of DRM, of course, is separating
 
  software source code (read by humans), from their compiled binaries (read
 
  only by computers).  Furthermore,
 
  \href{http://www.law.cornell.edu/uscode/text/17/1201}{17 USC~\S1201} often
 
  prohibits users legally from circumventing some of these DRM systems.
 

	
 
\item Most EULAs also include a contractual agreement that bind users further
 
  by forcing them to agree to a contractual, prohibitive software license
 
  before ever even using the software.
 

	
 
\end{itemize}
 

	
 
Thus, most proprietary software restricts users via multiple interlocking
 
legal and technological means.  Any license that truly respect the software
 
freedom of all users must not only grant appropriate copyright permissions,
 
but also \textit{prevent} restrictions from other legal and technological
 
means like those listed above.
 

	
 
\subsection{Non-USA Copyright Regimes}
 
\label{non-usa-copyright}
 

	
 
Generally speaking, copyright law operates similarly enough in countries that
 
have signed the Berne Convention on Copyright, and software freedom licenses
 
have generally taken advantage of this international standardization of
 
copyright law.  However, copyright law does differ from country to country,
 
and commonly, software freedom licenses like the GPL must be considered under the
 
copyright law in the jurisdiction where any licensing dispute occurs.
 

	
 
Those who are most familiar with the USA's system of copyright often are
 
surprised to learn that there are certain copyright controls that cannot be
 
waived nor disclaimed.  Specifically, many copyright regimes outside the USA
 
recognize a concept of moral rights of authors.  Typically, moral rights are
 
fully compatible with respecting software freedom, as they are usually
 
centered around controls that software freedom licenses generally respect,
 
such as the right of an authors to require proper attribution for their work.
 

	
 
\section{A Community of Equality}
 

	
 
The previous section described the principles of software freedom, a brief
 
introduction to mechanisms that typically block these freedoms, and the
 
simplest ways that copyright holders might grant those freedoms to their
 
users for their copyrighted works of software.  The previous section also
 
introduced the idea of \textit{copyleft}: a licensing mechanism to use
 
copyright to not only grant software freedom to users, but also to uphold
 
those rights against those who might seek to curtail them.
 

	
 
Copyleft, as defined in \S~\ref{copyleft-definition}, is a general term for this
 
mechanism.  The remainder of this text will discuss details of various
 
real-world implementations of copyleft -- most notably, the GPL\@.
 

	
 
This discussion begins first with some general explanation of what the GPL is
 
able to do in software development communities.  After that brief discussion
 
in this section, deeper discussion of how GPL accomplishes this in practice
 
follows in the next chapter.
 

	
 
Simply put, though, the GPL ultimately creates a community of equality for
 
both business and noncommercial users.
 

	
 
\subsection{The Noncommercial Community}
 

	
 
A GPL'd code base becomes a center of a vibrant development and user
 
community.  Traditionally, volunteers, operating noncommercially out of
 
keen interest or ``scratch an itch'' motivations, produce initial versions
 
of a GPL'd system.  Because of the efficient distribution channels of the
 
Internet, any useful GPL'd system is adopted quickly by noncommercial
 
users.
 

	
 
Fundamentally, the early release and quick distribution of the software
 
gives birth to a thriving noncommercial community.  Users and developers
 
begin sharing bug reports and bug fixes across a shared intellectual
 
commons.  Users can trust the developers, because they know that if the
 
developers fail to address their needs or abandon the project, the GPL
 
ensures that someone else has the right to pick up development.
 
Developers know that the users cannot redistribute their software without
 
passing along the rights granted by the GPL, so they are assured that every
 
one of their users is treated equally.
 

	
 
Because of the symmetry and fairness inherent in GPL'd distribution,
 
nearly every GPL'd package in existence has a vibrant noncommercial user
 
and developer base.
 

	
 
\subsection{The Commercial Community}
 

	
 
By the same token, nearly all established GPL'd software systems have a
 
vibrant commercial community.  Nearly every GPL'd system that has gained
 
wide adoption from noncommercial users and developers eventually begins
 
to fuel a commercial system around that software.
 

	
 
For example, consider the Samba file server system that allows Unix-like
 
systems (including GNU/Linux) to serve files to Microsoft Windows systems.
 
Two graduate students originally developed Samba in their spare time and
 
it was deployed noncommercially in academic environments.\footnote{See
 
  \href{http://turtle.ee.ncku.edu.tw/docs/samba/history}{Andrew Tridgell's
 
    ``A bit of history and a bit of fun''}}  However, very
 
soon for-profit companies discovered that the software could work for them
 
as well, and their system administrators began to use it in place of
 
Microsoft Windows NT file-servers.  This served to lower the cost of
 
running such servers by orders of magnitude. There was suddenly room in
 
Windows file-server budgets to hire contractors to improve Samba.  Some of
 
the first people hired to do such work were those same two graduate
 
students who originally developed the software.
 

	
 
The noncommercial users, however, were not concerned when these two
 
fellows began collecting paychecks off of their GPL'd work.  They knew
 
that because of the nature of the GPL that improvements that were
 
distributed in the commercial environment could easily be folded back into
 
the standard version.  Companies are not permitted to proprietarize
 
Samba, so the noncommercial users, and even other commercial users are
 
safe in the knowledge that the software freedom ensured by the GPL will remain
 
protected.
 

	
 
Commercial developers also work in concert with noncommercial
 
developers.  Those two now-long-since graduated students continue to
 
contribute to Samba altruistically, but also get paid work doing it.
 
Priorities change when a client is in the mix, but all the code is
 
contributed back to the standard version.  Meanwhile, many other
 
individuals have gotten involved noncommercially as developers,
 
because they want to ``cut their teeth on Free Software,'' or because
 
the problems interest them.  When they get good at it, perhaps they
 
will move on to another project, or perhaps they will become
 
commercial developers of the software themselves.
 

	
 
No party is a threat to another in the GPL software scenario because
 
everyone is on equal ground.  The GPL protects rights of the commercial
 
and noncommercial contributors and users equally. The GPL creates trust,
 
because it is a level playing field for all.
 

	
 
\subsection{Law Analogy}
 

	
 
In his introduction to Stallman's \emph{Free Software, Free Society},
 
Lawrence Lessig draws an interesting analogy between the law and Free
 
Software. He argues that the laws of a free society must be protected
 
much like the GPL protects software.  So that I might do true justice to
 
Lessig's argument, I quote it verbatim:
 

	
 
\begin{quotation}
 

	
 
A ``free society'' is regulated by law. But there are limits that any free
 
society places on this regulation through law: No society that kept its
 
laws secret could ever be called free.  No government that hid its
 
regulations from the regulated could ever stand in our tradition. Law
 
controls.  But it does so justly only when visibly.  And law is visible
 
only when its terms are knowable and controllable by those it regulates,
 
or by the agents of those it regulates (lawyers, legislatures).
 

	
 
This condition on law extends beyond the work of a legislature.  Think
 
about the practice of law in American courts.  Lawyers are hired by their
 
clients to advance their clients' interests.  Sometimes that interest is
 
advanced through litigation. In the course of this litigation, lawyers
 
write briefs. These briefs in turn affect opinions written by judges.
 
These opinions decide who wins a particular case, or whether a certain law
 
can stand consistently with a constitution.
 

	
 
All the material in this process is free in the sense that Stallman means.
 
Legal briefs are open and free for others to use.  The arguments are
 
transparent (which is different from saying they are good), and the
 
reasoning can be taken without the permission of the original lawyers.
 
The opinions they produce can be quoted in later briefs.  They can be
 
copied and integrated into another brief or opinion.  The ``source code''
 
for American law is by design, and by principle, open and free for anyone
 
to take. And take lawyers do---for it is a measure of a great brief that
 
it achieves its creativity through the reuse of what happened before.  The
 
source is free; creativity and an economy is built upon it.
 

	
 
This economy of free code (and here I mean free legal code) doesn't starve
 
lawyers.  Law firms have enough incentive to produce great briefs even
 
though the stuff they build can be taken and copied by anyone else.  The
 
lawyer is a craftsman; his or her product is public.  Yet the crafting is
 
not charity. Lawyers get paid; the public doesn't demand such work
 
without price.  Instead this economy flourishes, with later work added to
 
the earlier.
 

	
 
We could imagine a legal practice that was different --- briefs and
 
arguments that were kept secret; rulings that announced a result but not
 
the reasoning. Laws that were kept by the police but published to no one
 
else. Regulation that operated without explaining its rule.
 

	
 
We could imagine this society, but we could not imagine calling it
 
``free.''  Whether or not the incentives in such a society would be better
 
or more efficiently allocated, such a society could not be known as free.
 
The ideals of freedom, of life within a free society, demand more than
 
efficient application.  Instead, openness and transparency are the
 
constraints within which a legal system gets built, not options to be
 
added if convenient to the leaders.  Life governed by software code should
 
be no less.
 

	
 
Code writing is not litigation.  It is better, richer, more
 
productive.  But the law is an obvious instance of how creativity and
 
incentives do not depend upon perfect control over the products
 
created.  Like jazz, or novels, or architecture, the law gets built
 
upon the work that went before. This adding and changing is what
 
creativity always is.  And a free society is one that assures that its
 
most important resources remain free in just this sense.\footnote{This
 
quotation is Copyright \copyright{} 2002, Lawrence Lessig. It is
 
licensed under the terms of
 
\href{http://creativecommons.org/licenses/by/1.0/}{the ``Attribution
 
License'' version 1.0} or any later version as published by Creative
 
Commons.}
 
\end{quotation}
 

	
 
In essence, lawyers are paid to service the shared commons of legal
 
infrastructure.  Few citizens defend themselves in court or write their
 
own briefs (even though they are legally permitted to do so) because
 
everyone would prefer to have an expert do that job.
 

	
 
The Free Software economy is a market ripe for experts.  It
 
functions similarly to other well established professional fields like the
 
law. The GPL, in turn, serves as the legal scaffolding that permits the
 
creation of this vibrant commercial and noncommercial Free Software
 
economy.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{A Tale of Two Copyleft Licenses}
 
\label{tale-of-two-copylefts}
 

	
 
While determining the proper methodology and criteria to yield an accurate
 
count remains difficult, the GPL is generally considered one of the most
 
widely used Free Software licenses.  For most of its history --- for 16 years
 
from June 1991 to June 2007 --- there was really only one version of the GPL,
 
version 2.
 

	
 
However, the GPL had both earlier versions before version 2, and, more well
 
known, a revision to version 3. 
 

	
 
\section{Historical Motivations for the General Public License}
 

	
 
The earliest license to grant software freedom was likely the Berkeley
 
Software Distribution (``BSD'') license.  This license is typical of what are
 
often called lax, highly permissive licenses.  Not unlike software in the
 
public domain, these non-copyleft licenses (usually) grant software freedom
 
to users, but they do not go to any effort to uphold that software freedom
 
for users.  The so-called ``downstream'' (those who receive the software and
 
then build new things based on that software) can restrict the software and
 
distribute further.
 

	
 
The GNU's Not Unix (``GNU'') project, which Richard M.~Stallman (``RMS'')
 
founded in 1984 to make a complete Unix-compatible operating system
 
implementation that assured software freedom for all.  However, RMS saw that
 
using a license that gave but did not assure software freedom would be
 
counter to the goals of the GNU project.  RMS invented ``copyleft'' as an
 
answer to that problem, and began using various copyleft licenses for the
 
early GNU project programs.\footnote{RMS writes more fully about this topic in
 
  his essay entitled simply
 
  \href{http://www.gnu.org/gnu/thegnuproject.html}{\textit{The GNU Project}}.
 
    For those who want to hear the story in his own voice,
 
    \href{http://audio-video.gnu.org/audio/}{speech recordings} of his talk,
 
    \textit{The Free Software Movement and the GNU/Linux Operating System}
 
    are also widely available}
 

	
 
\section{Proto-GPLs And Their Impact}
 

	
 
%FIXME-LATER: bad line break:
 
%\href{http://www.free-soft.org/gpl_history/emacs_gpl.html}{The Emacs
 
%  General Public License}
 
The earliest copyleft licenses were specific to various GNU programs.  For
 
example,  The Emacs
 
General Public License was likely the first copyleft license ever
 
published.  Interesting to note that even this earliest copyleft license
 
contains a version of the well-known GPL copyleft clause:
 

	
 
\begin{quotation}
 
You may modify your copy or copies of GNU Emacs \ldots provided that you also
 
\ldots cause the whole of any work that you distribute or publish, that in
 
whole or in part contains or is a derivative of GNU Emacs or any part
 
thereof, to be licensed at no charge to all third parties on terms identical
 
to those contained in this License Agreement.
 
\end{quotation}
 

	
 
This simply stated clause is the fundamental innovation of copyleft.
 
Specifically, copyleft \textit{uses} the copyright holders' controls on
 
permission to modify the work to add a conditional requirement.  Namely,
 
downstream users may only have permission to modify  the work if they pass
 
along the same permissions on the modified version that came originally to
 
them.
 

	
 
These original program-specific proto-GPLs give an interesting window into
 
the central ideas and development of copyleft.  In particular, reviewing them
 
shows how the text of the GPL we know has evolved to address more of the
 
issues discussed earlier in \S~\ref{software-and-non-copyright}.
 

	
 
\section{The GNU General Public License, Version 1}
 
\label{GPLv1}
 

	
 
In January 1989, the FSF announced that the GPL had been converted into a
 
``subroutine'' that could be reused not just for all FSF-copyrighted
 
programs, but also by anyone else.  As the FSF claimed in its announcement of
 
the GPLv1:\footnote{The announcement of GPLv1 was published in the
 
  \href{http://www.gnu.org/bulletins/bull6.html\#SEC8}{GNU's Bulletin, vol 1,
 
    number 6 dated January 1989}.  (Thanks very much to Andy Tai for his
 
  \href{http://www.free-soft.org/gpl_history/}{consolidation of research on
 
    the history of the pre-v1 GPL's}.)}
 
\begin{quotation}
 
To make it easier to copyleft programs, we have been improving on the
 
legalbol architecture of the General Public License to produce a new version
 
that serves as a general-purpose subroutine: it can apply to any program
 
without modification, no matter who is publishing it.
 
\end{quotation}
 

	
 
This, like many inventive ideas, seems somewhat obvious in retrospect.  But,
 
the FSF had some bright people and access to good lawyers when it started.
 
It took almost five years from the first copyleft licenses to get to a
 
generalized, reusable GPLv1.  In the context and mindset of the 1980s, this
 
is not surprising.  The idea of reusable licensing infrastructure was not
 
only uncommon, it was virtually nonexistent!  Even the early BSD licenses
 
were simply copied and rewritten slightly for each new use.\footnote{It
 
  remains an interesting accident of history that the early BSD problematic
 
  ``advertising clause'' (discussion of which is somewhat beyond the scope of
 
  this tutorial) lives on into current day, simply because while the
 
  University of California at Berkeley gave unilateral permission to remove
 
  the clause from \textit{its} copyrighted works, others who adapted the BSD
 
  license with their own names in place of UC-Berkeley's never have.}  The
 
GPLv1's innovation of reusable licensing infrastructure, an obvious fact
 
today, was indeed a novel invention for its day.\footnote{We're all just
 
  grateful that the FSF also opposes business method patents, since the FSF's
 
  patent on a ``method for reusable licensing infrastructure'' would have
 
  not expired until 2006!}
 

	
 
\section{The GNU General Public License, Version 2}
 

	
 
The GPLv2 was released two and a half years after GPLv1, and over the
 
following sixteen years, it became the standard for copyleft licensing until
 
the release of GPLv3 in 2007 (discussed in more detail in the next section).
 

	
 
While this tutorial does not discuss the terms of GPLv1 in detail, it is
 
worth noting below the three key changes that GPLv2 brought:
 

	
 
\begin{itemize}
 

	
 
\item Software patents and their danger are explicitly mentioned, inspiring
 
  (in part) the addition of GPLv2~\S\S5--7.  (These sections are discussed in
 
  detail in \S~\ref{GPLv2s5}, \S~\ref{GPLv2s6} and \S~\ref{GPLv2s7} of this
 
  tutorial.)
 

	
 
\item GPLv2~\S2's copyleft terms are expanded to more explicitly discuss the
 
  issue of combined works.  (GPLv2~\S2 is discussed in detail in
 
  \S~\ref{GPLv2s2} in this tutorial).
 

	
 
\item GPLv2~\S3 includes more detailed requirements, including the phrase
 
 ``the scripts used to control compilation and installation of the
 
  executable'', which is a central component of current GPLv2 enforcement.
 
  (GPLv2~\S3 is discussed in detail in
 
  \S~\ref{GPLv2s3} in this tutorial).
 
\end{itemize}
 

	
 
The next chapter discusses GPLv2 in full detail, and readers who wish to dive
 
into the section-by-section discussion of the GPL should jump ahead now to
 
that chapter.  However, the most interesting fact to note here is how GPLv2
 
was published with little fanfare and limited commentary.  This contrasts
 
greatly with the creation of GPLv3.
 

	
 
\section{The GNU General Public License, Version 3}
 

	
 
RMS began drafting GPLv2.2 in mid-2002, and FSF ran a few discussion groups
 
during that era about new text of that license.  However, rampant violations
 
of the GPL required more immediate attention of FSF's licensing staff, and as
 
such, much of the early 2000's was spent doing GPL enforcement
 
work.\footnote{More on GPL enforcement is discussed in \tutorialpartsplit{a
 
    companion tutorial, \textit{A Practical Guide to GPL
 
      Compliance}}{Part~\ref{gpl-compliance-guide} of this tutorial}.}  In
 
2006, FSF began in earnest drafting work for GPLv3.
 

	
 
The GPLv3 process began in earnest in January 2006.  It became clear that
 
many provisions of the GPL could benefit from modification to fit new
 
circumstances and to reflect what the entire community learned from
 
experience with version 2.  Given the scale of revision it seems proper to
 
approach the work through public discussion in a transparent and accessible
 
manner.
 

	
 
The GPLv3 process continued through June 2007, culminating in publication of
 
GPLv3 and LGPLv3 on 29 June 2007, AGPLv3 on 19 November 2007, and the GCC
 
Runtime Library Exception on 27 January 2009.
 

	
 
All told, four discussion drafts of GPLv3, two discussion drafts of LGPLv3
 
and two discussion drafts of AGPLv3 were published and discussed.
 
Ultimately, FSF remained the final arbiter and publisher of the licenses, and
 
RMS himself their primary author, but input was sought from many parties, and
 
these licenses do admittedly look and read more like legislation as a result.
 
Nevertheless, all of the ``v3'' group are substantially better and improved
 
licenses.
 

	
 
GPLv3 and its terms are discussed in detail in Chapter~\ref{GPLv3}.
 

	
 
\section{The Innovation of Optional ``Or Any Later'' Version}
 

	
 
An interesting fact of all GPL licenses is that there are ultimately multiple
 
choices for use of the license.  The FSF is the primary steward of GPL (as
 
discussed later in \S~\ref{GPLv2s9} and \S~\ref{GPLv3s14}).  However, those
 
who wish to license works under GPL are not required to automatically accept
 
changes made by the FSF for their own copyrighted works.
 

	
 
Each licensor may chose three different methods of licensing, as follows:
 

	
 
\begin{itemize}
 

	
 
\item explicitly name a single version of GPL for their work (usually
 
  indicated in shorthand by saying the license is ``GPLv$X$-only''), or
 

	
 
\item name no version of the GPL, thus they allow their downstream recipients
 
  to select any version of the GPL they choose (usually indicated in shorthand
 
  by saying the license is simply ``GPL''), or
 

	
 
\item name a specific version of GPL and give downstream recipients the
 
  option to choose that version ``or any later version as published by the
 
  FSF'' (usually indicated by saying the license is
 
  ``GPLv$X$-or-later'')\footnote{The shorthand of ``GPL$X+$'' is also popular
 
    for this situation.  The authors of this tutorial prefer ``-or-later''
 
    syntax, because it (a) mirrors the words ``or'' and ``later from the
 
    licensing statement, (b) the $X+$ doesn't make it abundantly clear that
 
    $X$ is clearly included as a license option and (c) the $+$ symbol has
 
    other uses in computing (such as with regular expressions) that mean
 
    something different.}
 
\end{itemize}
 

	
 
\label{license-compatibility-first-mentioned}
 

	
 
Oddly, this flexibility has received (in the opinion of the authors, undue)
 
criticism, primarily because of the complex and oft-debated notion of
 
``license compatibility'' (which is explained in detail in
 
\S~\ref{license-compatibility}).  Copyleft licenses are generally
 
incompatible with each other, because the details of how they implement
 
copyleft differs.  Specifically, copyleft works only because of its
 
requirement that downstream licensors use the \textit{same} license for
 
combined and modified works.  As such, software licensed under the terms of
 
``GPLv2-only'' cannot be combined with works licensed ``GPLv3-or-later''.
 
This is admittedly a frustrating outcome.
 

	
 
Other copyleft licenses that appeared after GPL, such as the Creative Commons
 
``Attribution-Share Alike'' licenses, the Eclipse Public License and the
 
Mozilla Public License \textbf{require} all copyright holders choosing
 
to use any version of those licenses to automatically allow use of their
 
copyrighted works under new versions.\footnote{CC-BY-SA-2.0 and greater only
 
permit licensing of adaptations under future versions; 1.0 did not have
 
any accomodation for future version compatibility.}  Of course, Creative
 
Commons, the Eclipse Foundation, and the Mozilla Foundation (like the FSF)
 
have generally served as excellent stewards of their licenses.  Copyright
 
holders using those licenses seems to find it acceptable to fully delegate
 
all future licensing decisions for their copyrights to these organizations
 
without a second thought.
 

	
 
However, note that FSF gives herein the control of copyright holders to
 
decide whether or not to implicitly trust the FSF in its work of drafting
 
future GPL versions.  The FSF, for its part, does encourage copyright holders
 
to chose by default ``GPLv$X$-or-later'' (where $X$ is the most recent
 
version of the GPL published by the FSF).  However, the FSF \textbf{does not
 
  mandate} that a choice to use any GPL requires a copyright holder ceding
 
its authority for future licensing decisions to the FSF.  In fact, the FSF
 
considered this possibility for GPLv3 and chose not to do so, instead opting
 
for the third-party steward designation clause discussed in
 
Section~\ref{GPLv3s14}.
 

	
 
\section{Complexities of Two Simultaneously Popular Copylefts}
 

	
 
Obviously most GPL advocates would prefer widespread migration to GPLv3, and
 
many newly formed projects who seek a copyleft license tend to choose a
 
GPLv3-based license.  However, many existing copylefted projects continue
 
with GPLv2-only or GPLv2-or-later as their default license.
 

	
 
While GPLv3 introduces many improvements --- many of which were designed to
 
increase adoption by for-profit companies --- GPLv2 remains a widely used and
 
extremely popular license.  The GPLv2 is, no doubt, a good and useful
 
license.
 

	
 
However, unlike GPLv1 before it,
 
GPLv2 remains an integral part of the copyleft licensing infrastructure.  As such, those who seek to have expertise in current
 
topics of copyleft licensing need to study both the GPLv2 and GPLv3 family of
 
licenses.
 

	
 
Furthermore, GPLv3 is more easily understood by first studying GPLv2.
 
This is not only because of their chronological order, but also because much
 
of the discussion material available for GPLv3 tends to talk about GPLv3 in
 
contrast to GPLv2.  As such, a strong understanding of GPLv2 helps in
 
understanding most of the third-party material found regarding GPLv3.  Thus,
 
the following chapter begins a deep discussion of GPLv2.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Running Software and Verbatim Copying}
 
\label{run-and-verbatim}
 

	
 

	
 
This chapter begins the deep discussion of the details of the terms of
 
GPLv2\@. In this chapter, we consider the first two sections: GPLv2 \S\S
 
0--2. These are the straightforward sections of the GPL that define the
 
simplest rights that the user receives.
 

	
 
\section{GPLv2~\S0: Freedom to Run}
 
\label{GPLv2s0}
 

	
 
GPLv2~\S0, the opening section of GPLv2, sets forth that copyright law governs
 
the work.  It specifically points out that it is the ``copyright
 
holder'' who decides if a work is licensed under its terms and explains
 
how the copyright holder might indicate this fact.
 

	
 
A bit more subtly, GPLv2~\S0 makes an inference that copyright law is the only
 
system that can restrict the software.  Specifically, it states:
 
\begin{quote}
 
Activities other than copying, distribution and modification are not
 
covered by this License; they are outside its scope.
 
\end{quote}
 
In essence, the license governs \emph{only} those activities, and all other
 
activities are unrestricted, provided that no other agreements trump GPLv2
 
(which they cannot; see Sections~\ref{GPLv2s6} and~\ref{GPLv2s7}).  This is
 
very important, because the Free Software community heavily supports
 
users' rights to ``fair use'' and ``unregulated use'' of copyrighted
 
material.  GPLv2 asserts through this clause that it supports users' rights
 
to fair and unregulated uses.
 

	
 
Fair use (called ``fair dealing'' in some jurisdictions) of copyrighted
 
material is an established legal doctrine that permits certain activities
 
regardless of whether copyright law would otherwise restrict those activities.
 
Discussion of the various types of fair use activity are beyond the scope of
 
this tutorial.  However, one important example of fair use is the right to
 
quote portions of the text in a larger work so as to criticize or suggest
 
changes.  This fair use right is commonly used on mailing lists when
 
discussing potential improvements or changes to Free Software.
 

	
 
Fair use is a doctrine established by the courts or by statute.  By
 
contrast, unregulated uses are those that are not covered by the statue
 
nor determined by a court to be covered, but are common and enjoyed by
 
many users.  An example of unregulated use is reading a printout of the
 
program's source code like an instruction book for the purpose of learning
 
how to be a better programmer.  The right to read something that you have
 
access to is and should remain unregulated and unrestricted.
 

	
 
\medskip
 

	
 
Thus, the GPLv2 protects users' fair and unregulated use rights precisely by
 
not attempting to cover them.  Furthermore, the GPLv2 ensures the freedom
 
to run specifically by stating the following:
 
\begin{quote}
 
''The act of running the Program is not restricted.''
 
\end{quote}
 
Thus, users are explicitly given the freedom to run by GPLv2~\S0.
 

	
 
\medskip
 

	
 
The bulk of GPLv2~\S0 not yet discussed gives definitions for other terms used
 
throughout.  The only one worth discussing in detail is ``work based on
 
the Program''.  The reason this definition is particularly interesting is
 
not for the definition itself, which is rather straightforward, but
 
because it clears up a common misconception about the GPL\@.
 

	
 
The GPL is often mistakenly criticized because it fails to give a
 
definition of ``derivative work'' or ``combined work''.  In fact, it would be incorrect and
 
problematic if the GPL attempted to define these terms.  A copyright license, in
 
fact, has no control over the rules of copyright themselves.  Such rules are
 
the domain of copyright law and the courts --- not the licenses that utilize
 
those systems.
 

	
 
Copyright law as a whole does not propose clear and straightforward guidelines
 
for identifying the derivative and/or combined works of software.  However,
 
no copyright license --- not even the GNU GPL --- can be blamed for this.
 
Legislators and court opinions must give us guidance in borderline cases.
 
Meanwhile, lawyers will likely based their conclusions on the application of rules
 
made in the context of literary or artistic copyright to the different
 
context of computer programming and by analyzing the (somewhat limited) case
 
law and guidance available from various sources.
 
(Chapter~\ref{derivative-works} discusses this issue in depth.)
 

	
 

	
 
\section{GPLv2~\S1: Verbatim Copying}
 
\label{GPLv2s1}
 

	
 
GPLv2~\S1 covers the matter of redistributing the source code of a program
 
exactly as it was received. This section is quite straightforward.
 
However, there are a few details worth noting here.
 

	
 
The phrase ``in any medium'' is important.  This, for example, gives the
 
freedom to publish a book that is the printed copy of the program's source
 
code.  It also allows for changes in the medium of distribution.  Some
 
vendors may ship Free Software on a CD, but others may place it right on
 
the hard drive of a pre-installed computer.  Any such redistribution media
 
is allowed.
 

	
 
Preservation of copyright notice and license notifications are mentioned
 
specifically in GPLv2~\S1.  These are in some ways the most important part of
 
the redistribution, which is why they are mentioned by name.  GPL
 
always strives to make it abundantly clear to anyone who receives the
 
software what its license is.  The goal is to make sure users know their
 
rights and freedoms under GPL, and to leave no reason that users might be
 
surprised the software is GPL'd. Thus
 
throughout the GPL, there are specific references to the importance of
 
notifying others down the distribution chain that they have rights under
 
GPL.
 

	
 
Also mentioned by name is the warranty disclaimer. Most people today do
 
not believe that software comes with any warranty.  Notwithstanding
 
\href{http://mlis.state.md.us/2000rs/billfile/hb0019.htm}{Maryland's} and \href{http://leg1.state.va.us/cgi-bin/legp504.exe?001+ful+SB372ER}{Virginia's} UCITA bills, there are few or no implied warranties with software.
 
However, just to be on the safe side, GPL clearly disclaims them, and the
 
GPL requires re-distributors to keep the disclaimer very visible. (See
 
Sections~\ref{GPLv2s11} and~\ref{GPLv2s12} of this tutorial for more on GPL's
 
warranty disclaimers.)
 

	
 
Note finally that GPLv2~\S1 creates groundwork for the important defense of
 
commercial freedom.  GPLv2~\S1 clearly states that in the case of verbatim
 
copies, one may make money.  Re-distributors are fully permitted to charge
 
for the re-distribution of copies of Free Software. In addition, they may
 
provide the warranty protection that the GPL disclaims as an additional
 
service for a fee. (See Section~\ref{Business Models} for more discussion
 
on making a profit from Free Software redistribution.)
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 

	
 
\chapter{Derivative Works: Statute and Case Law}
 
\label{derivative-works}
 

	
 
As described in the \hyperref[copyleft-definition]{earlier general discussion
 
  of copyleft}, strong copyleft licenses such as the GPL seek to uphold
 
software freedom via the copyright system.  This principle often causes
 
theoretical or speculative dispute among lawyers, because ``the work'' ---
 
the primary unit of consideration under most copyright rules -- is not a unit
 
of computer programming. In order to determine whether a ``routine'' an
 
``object'', a ``function'', a ``library'' or any other unit of software is
 
part of one ``work'' when combined with other GPL'd code, we must ask a
 
question that copyright law will not directly answer in the same technical
 
terms.
 

	
 
Therefore, this chapter digresses from  discussion of GPL's exact text to
 
consider the matter of combined and/or derivative works --- a concept that we must
 
understand fully before considering GPLv2~\S\S2--3\@.  At least under USA
 
copyright law, The GPL, and Free
 
Software licensing in general, relies critically on the concept of
 
``derivative work'' since software that is ``independent,'' (i.e., not
 
``derivative'') of Free Software need not abide by the terms of the
 
applicable Free Software license. As much is required by \S~106 of the
 
Copyright Act, 17 U.S.C. \S~106 (2002), and admitted by Free Software
 
licenses, such as the GPL, which (as we have seen) states in GPLv2~\S0 that ``a
 
`work based on the Program' means either the Program or any derivative
 
work under copyright law.'' It is being a derivative work of Free Software
 
that triggers the necessity to comply with the terms of the Free Software
 
license under which the original work is distributed. Therefore, one is
 
left to ask, just what is a ``derivative work''? The answer to that
 
question differs depending on which court is being asked.
 

	
 
The analysis in this chapter sets forth the differing definitions of
 
derivative work by the circuit courts. The broadest and most
 
established definition of derivative work for software is the
 
abstraction, filtration, and comparison test (``the AFC test'') as
 
created and developed by the Second Circuit. Some circuits, including
 
the Ninth Circuit and the First Circuit, have either adopted narrower
 
versions of the AFC test or have expressly rejected the AFC test in
 
favor of a narrower standard. Further, several other circuits have yet
 
to adopt any definition of derivative work for software.
 

	
 
As an introductory matter, it is important to note that literal copying of
 
a significant portion of source code is not always sufficient to establish
 
that a second work is a derivative work of an original
 
program. Conversely, a second work can be a derivative work of an original
 
program even though absolutely no copying of the literal source code of
 
the original program has been made. This is the case because copyright
 
protection does not always extend to all portions of a program's code,
 
while, at the same time, it can extend beyond the literal code of a
 
program to its non-literal aspects, such as its architecture, structure,
 
sequence, organization, operational modules, and computer-user interface.
 

	
 
\section{The Copyright Act}
 

	
 
The copyright act is of little, if any, help in determining the definition
 
of a derivative work of software. However, the applicable provisions do
 
provide some, albeit quite cursory, guidance. Section 101 of the Copyright
 
Act sets forth the following definitions:
 

	
 
\begin{quotation}
 
A ``computer program'' is a set of statements or instructions to be used
 
directly or indirectly in a computer in order to bring about a certain
 
result.
 

	
 
A ``derivative work'' is a work based upon one or more preexisting works,
 
such as a translation, musical arrangement, dramatization,
 
fictionalization, motion picture version, sound recording, art
 
reproduction, abridgment, condensation, or any other form in which a work
 
may be recast, transformed, or adapted. A work consisting of editorial
 
revisions, annotations, elaborations, or other modifications which, as a
 
whole, represent an original work of authorship, is a ``derivative work.''
 
\end{quotation}
 

	
 
These are the only provisions in the Copyright Act relevant to the
 
determination of what constitutes a derivative work of a computer
 
program. Another provision of the Copyright Act that is also relevant to
 
the definition of derivative work is \S~102(b), which reads as follows:
 

	
 
\begin{quotation}
 
In no case does copyright protection for an original work of authorship
 
extend to any idea, procedure, process, system, method of operation,
 
concept, principle, or discovery, regardless of the form in which it is
 
described, explained, illustrated, or embodied in such work.
 
\end{quotation}
 

	
 
Therefore, before a court can ask whether one program is a derivative work
 
of another program, it must be careful not to extend copyright protection
 
to any ideas, procedures, processes, systems, methods of operation,
 
concepts, principles, or discoveries contained in the original program. It
 
is the implementation of this requirement to ``strip out'' unprotectable
 
elements that serves as the most frequent issue over which courts
 
disagree.
 

	
 
\section{Abstraction, Filtration, Comparison Test}
 

	
 
As mentioned above, the AFC test for determining whether a computer
 
program is a derivative work of an earlier program was created by the
 
Second Circuit and has since been adopted in the Fifth, Tenth, and
 
Eleventh Circuits. Computer Associates Intl., Inc. v. Altai, Inc., 982
 
F.2d 693 (2nd Cir. 1992); Engineering Dynamics, Inc. v. Structural
 
Software, Inc., 26 F.3d 1335 (5th Cir. 1994); Kepner-Tregoe,
 
Inc. v. Leadership Software, Inc., 12 F.3d 527 (5th Cir. 1994); Gates
 
Rubber Co. v. Bando Chem. Indust., Ltd., 9 F.3d 823 (10th Cir. 1993);
 
Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997); Bateman
 
v. Mnemonics, Inc., 79 F.3d 1532 (11th Cir. 1996); and, Mitek Holdings,
 
Inc. v. Arce Engineering Co., Inc., 89 F.3d 1548 (11th Cir. 1996).
 

	
 
Under the AFC test, a court first abstracts from the original program its
 
constituent structural parts. Then, the court filters from those
 
structural parts all unprotectable portions, including incorporated ideas,
 
expression that is necessarily incidental to those ideas, and elements
 
that are taken from the public domain. Finally, the court compares any and
 
all remaining kernels of creative expression to the structure of the
 
second program to determine whether the software programs at issue are
 
substantially similar so as to warrant a finding that one is the
 
derivative work of the other.
 

	
 
Often, the courts that apply the AFC test will perform a quick initial
 
comparison between the entirety of the two programs at issue in order to
 
help determine whether one is a derivative work of the other. Such a
 
holistic comparison, although not a substitute for the full application of
 
the AFC test, sometimes reveals a pattern of copying that is not otherwise
 
obvious from the application of the AFC test when, as discussed below,
 
only certain components of the original program are compared to the second
 
program. If such a pattern is revealed by the quick initial comparison,
 
the court is more likely to conclude that the second work is indeed a
 
derivative of the original.
 

	
 
\subsection{Abstraction}
 

	
 
The first step courts perform under the AFC test is separation of the
 
work's ideas from its expression. In a process akin to reverse
 
engineering, the courts dissect the original program to isolate each level
 
of abstraction contained within it. Courts have stated that the
 
abstractions step is particularly well suited for computer programs
 
because it breaks down software in a way that mirrors the way it is
 
typically created. However, the courts have also indicated that this step
 
of the AFC test requires substantial guidance from experts, because it is
 
extremely fact and situation specific.
 

	
 
By way of example, one set of abstraction levels is, in descending order
 
of generality, as follows: the main purpose, system architecture, abstract
 
data types, algorithms and data structures, source code, and object
 
code. As this set of abstraction levels shows, during the abstraction step
 
of the AFC test, the literal elements of the computer program, namely the
 
source and object code, are defined as particular levels of
 
abstraction. Further, the source and object code elements of a program are
 
not the only elements capable of forming the basis for a finding that a
 
second work is a derivative of the program. In some cases, in order to
 
avoid a lengthy factual inquiry by the court, the owner of the copyright in
 
the original work will submit its own list of what it believes to be the
 
protected elements of the original program. In those situations, the court
 
will forgo performing its own abstraction, and proceed to the second step of
 
the AFC test.
 

	
 
\subsection{Filtration}
 

	
 
The most difficult and controversial part of the AFC test is the second
 
step, which entails the filtration of protectable expression contained in
 
the original program from any unprotectable elements nestled therein. In
 
determining which elements of a program are unprotectable, courts employ a
 
myriad of rules and procedures to sift from a program all the portions
 
that are not eligible for copyright protection.
 

	
 
First, as set forth in \S~102(b) of the Copyright Act, any and all ideas
 
embodied in the program are to be denied copyright protection. However,
 
implementing this rule is not as easy as it first appears. The courts
 
readily recognize the intrinsic difficulty in distinguishing between ideas
 
and expression and that, given the varying nature of computer programs,
 
doing so will be done on an ad hoc basis. The first step of the AFC test,
 
the abstraction, exists precisely to assist in this endeavor by helping
 
the court separate out all the individual elements of the program so that
 
they can be independently analyzed for their expressive nature.
 

	
 
A second rule applied by the courts in performing the filtration step of
 
the AFC test is the doctrine of merger, which denies copyright protection
 
to expression necessarily incidental to the idea being expressed. The
 
reasoning behind this doctrine is that when there is only one way to
 
express an idea, the idea and the expression merge, meaning that the
 
expression cannot receive copyright protection due to the bar on copyright
 
protection extending to ideas. In applying this doctrine, a court will ask
 
whether the program's use of particular code or structure is necessary for
 
the efficient implementation of a certain function or process. If so, then
 
that particular code or structure is not protected by copyright and, as a
 
result, it is filtered away from the remaining protectable expression.
 

	
 
A third rule applied by the courts in performing the filtration step of
 
the AFC test is the doctrine of scenes a faire, which denies copyright
 
protection to elements of a computer program that are dictated by external
 
factors. Such external factors can include:
 

	
 
\begin{itemize}
 

	
 
  \item The mechanical
 
specifications of the computer on which a particular program is intended
 
to operate
 

	
 
  \item Compatibility requirements of other programs with which a
 
program is designed to operate in conjunction
 

	
 
  \item Computer manufacturers'
 
design standards
 

	
 
  \item Demands of the industry being serviced, and widely accepted programming practices within the computer industry
 

	
 
\end{itemize}
 

	
 
Any code or structure of a program that was shaped predominantly in
 
response to these factors is filtered out and not protected by
 
copyright. Lastly, elements of a computer program are also to be filtered
 
out if they were taken from the public domain or fail to have sufficient
 
originality to merit copyright protection.
 

	
 
Portions of the source or object code of a computer program are rarely
 
filtered out as unprotectable elements. However, some distinct parts of
 
source and object code have been found unprotectable. For example,
 
constants, the invariable integers comprising part of formulas used to
 
perform calculations in a program, are unprotectable. Further, although
 
common errors found in two programs can provide strong evidence of
 
copying, they are not afforded any copyright protection over and above the
 
protection given to the expression containing them.
 

	
 
\subsection{Comparison}
 

	
 
The third and final step of the AFC test entails a comparison of the
 
original program's remaining protectable expression to a second
 
program. The issue will be whether any of the protected expression is
 
copied in the second program and, if so, what relative importance the
 
copied portion has with respect to the original program overall. The
 
ultimate inquiry is whether there is ``substantial'' similarity between
 
the protected elements of the original program and the potentially
 
derivative work. The courts admit that this process is primarily
 
qualitative rather than quantitative and is performed on a case-by-case
 
basis. In essence, the comparison is an ad hoc determination of whether
 
the protectable elements of the original program that are contained in the
 
second work are significant or important parts of the original program. If
 
so, then the second work is a derivative work of the first. If, however,
 
the amount of protectable elements copied in the second work are so small
 
as to be de minimis, then the second work is not a derivative work of the
 
original.
 

	
 
\section{Analytic Dissection Test}
 

	
 
The Ninth Circuit has adopted the analytic dissection test to determine
 
whether one program is a derivative work of another. Apple Computer,
 
Inc. v. Microsoft Corp., 35 F.3d 1435 (9th Cir. 1994). The analytic
 
dissection test first considers whether there are substantial similarities
 
in both the ideas and expressions of the two works at issue. Once the
 
similar features are identified, analytic dissection is used to determine
 
whether any of those similar features are protected by copyright. This
 
step is the same as the filtration step in the AFC test. After identifying
 
the copyrightable similar features of the works, the court then decides
 
whether those features are entitled to ``broad'' or ``thin''
 
protection. ``Thin'' protection is given to non-copyrightable facts or
 
ideas that are combined in a way that affords copyright protection only
 
from their alignment and presentation, while ``broad'' protection is given
 
to copyrightable expression itself. Depending on the degree of protection
 
afforded, the court then sets the appropriate standard for a subjective
 
comparison of the works to determine whether, as a whole, they are
 
sufficiently similar to support a finding that one is a derivative work of
 
the other. ``Thin'' protection requires the second work be virtually
 
identical in order to be held a derivative work of an original, while
 
``broad'' protection requires only a ``substantial similarity.''
 

	
 
\section{No Protection for ``Methods of Operation''}
 

	
 
The First Circuit has taken the position that the AFC test is inapplicable 
 
when the works in question relate to unprotectable elements set forth in 
 
\S~102(b).  Their approach results in a much narrower definition
 
of derivative work for software in comparison to other circuits. Specifically, 
 
the
 
First Circuit holds that ``method of operation,'' as used in \S~102(b) of
 
the Copyright Act, refers to the means by which users operate
 
computers. Lotus Development Corp. v. Borland Int'l., Inc., 49 F.3d 807
 
(1st Cir. 1995).  In Lotus, the court held that a menu command
 
hierarchy for a computer program was uncopyrightable because it did not
 
merely explain and present the program's functional capabilities to the
 
user, but also served as a method by which the program was operated and
 
controlled. As a result, under the First Circuit's test, literal copying
 
of a menu command hierarchy, or any other ``method of operation,'' cannot
 
form the basis for a determination that one work is a derivative of
 
another.  As a result, courts in the First Circuit that apply the AFC test
 
do so only after applying a broad interpretation of \S~102(b) to filter out
 
unprotected elements. E.g., Real View, LLC v. 20-20 Technologies, Inc., 
 
683 F. Supp.2d 147, 154 (D. Mass. 2010).
 

	
 

	
 
\section{No Test Yet Adopted}
 

	
 
Several circuits, most notably the Fourth and Seventh, have yet to
 
declare their definition of derivative work and whether or not the
 
AFC, Analytic Dissection, or some other test best fits their
 
interpretation of copyright law. Therefore, uncertainty exists with
 
respect to determining the extent to which a software program is a
 
derivative work of another in those circuits. However, one may presume
 
that they would give deference to the AFC test since it is by far the
 
majority rule among those circuits that have a standard for defining
 
a software derivative work.
 

	
 
\section{Cases Applying Software Derivative Work Analysis}
 

	
 
In the preeminent case regarding the definition of a derivative work for
 
software, Computer Associates v. Altai, the plaintiff alleged that its
 
program, Adapter, which was used to handle the differences in operating
 
system calls and services, was infringed by the defendant's competitive
 
program, Oscar. About 30\% of Oscar was literally the same code as
 
that in Adapter. After the suit began, the defendant rewrote those
 
portions of Oscar that contained Adapter code in order to produce a new
 
version of Oscar that was functionally competitive with Adapter, without
 
having any literal copies of its code. Feeling slighted still, the
 
plaintiff alleged that even the second version of Oscar, despite having no
 
literally copied code, also infringed its copyrights. In addressing that
 
question, the Second Circuit promulgated the AFC test.
 

	
 
In abstracting the various levels of the program, the court noted a
 
similarity between the two programs' parameter lists and macros. However,
 
following the filtration step of the AFC test, only a handful of the lists
 
and macros were protectable under copyright law because they were either
 
in the public domain or required by functional demands on the
 
program. With respect to the handful of parameter lists and macros that
 
did qualify for copyright protection, after performing the comparison step
 
of the AFC test, it was reasonable for the district court to conclude that
 
they did not warrant a finding of infringement given their relatively minor
 
contribution to the program as a whole. Likewise, the similarity between
 
the organizational charts of the two programs was not substantial enough
 
to support a finding of infringement because they were too simple and
 
obvious to contain any original expression.
 

	
 
In the case of Oracle America v. Google, 872 F. Supp.2d 974 (N.D. Cal. 2012),
 
the Northern District of California District Court examined the question of 
 
whether the application program interfaces (APIs) associated with the Java
 
programming language are entitled to copyright protection.  While the 
 
court expressly declined to rule whether all APIs are free to use without 
 
license (872 F. Supp.2d 974 at 1002), the court held that the command 
 
structure and taxonomy of the APIs were not protectable under copyright law.
 
Specifically, the court characterized the command structure and taxonomy as
 
both a ``method of operation'' (using an approach not dissimilar to the 
 
First Circuit's analysis in Lotus) and a ``functional requirement for 
 
compatibility'' (using Sega v. Accolade, 977 F.2d 1510 (9th Cir. 1992) and
 
Sony Computer Ent. v. Connectix, 203 F.3d 596 (9th Cir. 2000) as analogies),
 
and thus unprotectable subject matter under \S~102(b). 
 

	
 
Perhaps not surprisingly, there have been few other cases involving a highly
 
detailed software derivative work analysis. Most often, cases involve
 
clearer basis for decision, including frequent bad faith on the part of
 
the defendant or over-aggressiveness on the part of the plaintiff.  
 

	
 
\section{How Much Do Derivative Works Matter?}
 

	
 
It is certainly true that GPL intends for any work that is determined a
 
``derivative work'' under copyright law must be licensed as a whole under
 
GPL\@, as will be discussed in the following chapter.  However, as we finish
 
up our discussion derivative works, we must note that preparation of a
 
derivative work is by far not the only way to create a new work covered by
 
GPL\@.
 

	
 
In fact, while derivative work preparation is perhaps the most exciting area
 
of legal issues to consider, the more mundane ways to create a new work
 
covered by GPL are much more common.  For example, copyright statutes
 
generally require permission from the copyright holder to grant explicit
 
permission to modify a work in any manner.  As discussed in the next chapter,
 
the GPL {\em does} grant such permission, but requires the modified work must
 
also be licensed under the terms of the GPL (and only GPL:
 
see\S~\label{GPLv2s6} in this tutorial).  Determining whether software was
 
modified is a substantially easier analysis than the derivative work
 
discussions and considerations in this chapter.
 

	
 
The question of derivative works, when and how they are made, is undoubtedly
 
an essential discussion in the interpretation and consideration of copyleft.
 
That is why this chapter was included in this tutorial.  However, as we
 
return from this digression and resume discussion of the detailed text of the
 
GPLv2, we must gain a sense of perspective: most GPL questions center around
 
questions of modification and distribution, not preparation of derivative
 
works.  Derivative work preparation is ultimately a small subset of the types
 
of modified versions of the software a developer might create, thus, while an
 
excessive focus on derivative works indulges us in the more exciting areas of
 
copyleft, we must keep a sense of perspective regarding their relative
 
importance.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 

	
 
\chapter{Modified Source and Binary Distribution}
 
\label{source-and-binary}
 

	
 
In this chapter, we discuss the two core sections that define the rights
 
and obligations for those who modify, improve, and/or redistribute GPL'd
 
software. These sections, GPLv2~\S\S2--3, define the central core rights and
 
requirements of GPLv2\@.
 

	
 
\section{GPLv2~\S2: Share and Share Alike}
 
\label{GPLv2s2}
 

	
 
For many, this is where the ``magic'' happens that defends software
 
freedom upon redistribution.  GPLv2~\S2 is the only place in GPLv2
 
that governs the modification controls of copyright law.  If users
0 comments (0 inline, 0 general)