Changeset - e893d9d78296
[Not reviewed]
0 1 0
Bradley Kuhn (bkuhn) - 10 years ago 2014-03-18 22:33:18
bkuhn@ebb.org
Wordsmith and typo fixes.
1 file changed with 3 insertions and 3 deletions:
0 comments (0 inline, 0 general)
gpl-lgpl.tex
Show inline comments
...
 
@@ -1202,1540 +1202,1540 @@ 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,
 
constant s, 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 expressly rejected the AFC test and, instead, takes a
 
much narrower view of the meaning of derivative work for software. 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 IntÂ’l., Inc., 49 F.3d 807
 
(1st Cir. 1995). More specifically, the court held that a menu command
 
hierarchy for a computer program was uncopyrightable because it did not
 
merely explain and present the programÂ’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 CircuitÂ’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. It is also reasonable to expect that the First Circuit will read
 
the unprotectable elements set forth in \S~102(b) broadly, and, as such,
 
promulgate a definition of derivative work that is much narrower than that
 
which exists under the AFC test.
 

	
 
\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 amongst 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
 
have 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.
 

	
 
Perhaps not surprisingly, there have been few 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 overaggressiveness on the part of the plaintiff.
 
However, no cases involving Free Software licensing have ever gone to
 
court. As Free Software becomes an ever-increasingly important part of
 
the economy, it remains to be seen if battle lines will be
 
drawn over whether particular programs infringe the rights of Free
 
Software developers or whether the entire community, including industry,
 
adopts norms avoiding such risk.
 

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

	
 
\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}
 

	
 
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
 
modifies a GPLv2'd program, they must follow the terms of GPLv2~\S2 in making
 
those changes.  Thus, this sections ensures that the body of GPL'd software, as it
 
continues and develops, remains Free as in freedom.
 

	
 
To achieve that goal, GPLv2~\S2 first sets forth that the rights of
 
redistribution of modified versions are the same as those for verbatim
 
copying, as presented in GPLv2~\S1.  Therefore, the details of charging money,
 
keeping copyright notices intact, and other GPLv2~\S1 provisions are in tact
 
here as well.  However, there are three additional requirements.
 

	
 
The first (GPLv2~\S2(a)) requires that modified files carry ``prominent
 
notices'' explaining what changes were made and the date of such
 
changes. This section does not prescribe some specific way of
 
marking changes nor does it control the process of how changes are made.
 
Primarily, GPLv2~\S2(a) seeks to ensure that those receiving modified
 
versions know the history of changes to the software.  For some users,
 
it is important to know that they are using the standard version of
 
program, because while there are many advantages to using a fork,
 
there are a few disadvantages.  Users should be informed about the
 
historical context of the software version they use, so that they can
 
make proper support choices.  Finally, GPLv2~\S2(a) serves an academic
 
purpose --- ensuring that future developers can use a diachronic
 
approach to understand the software.
 

	
 
\medskip
 

	
 
The second requirement (GPLv2~\S2(b)) contains the four short lines that embody
 
the legal details of ``share and share alike''.  These 46 words are
 
considered by some to be the most worthy of careful scrutiny because
 
GPLv2~\S2(b), and they
 
can be a source of great confusion when not properly understood.
 

	
 
In considering GPLv2~\S2(b), first note the qualifier: it \textit{only} applies to
 
derivative works that ``you distribute or publish''.  Despite years of
 
education efforts on this matter, many still believe that modifiers
 
of GPL'd software \textit{must} to publish or otherwise
 
share their changes.  On the contrary, GPLv2~\S2(b) {\bf does not apply if} the
 
changes are never distributed.  Indeed, the freedom to make private,
 
personal, unshared changes to software for personal use only should be
 
protected and defended.\footnote{Most Free Software enthusiasts believe there is an {\bf
 
    moral} obligation to redistribute changes that are generally useful,
 
  and they often encourage companies and individuals to do so.  However, there
 
  is a clear distinction between what one {\bf ought} to do and what one
 
  {\bf must} do.}
 

	
 
Next, we again encounter the same matter that appears in GPLv2~\S0, in the
 
following text:
 
\begin{quote}
 
``...that in whole or part contains or is derived from the Program or any part thereof.''
 
\end{quote}
 
Again, the GPL relies here on what the copyright law says is a derivative
 
work.  If, under copyright law, the modified version ``contains or is
 
derived from'' the GPL'd software, then the requirements of GPLv2~\S2(b)
 
apply.  The GPL invokes its control as a copyright license over the
 
modification of the work in combination with its control over distribution
 
of the work.
 

	
 
The final clause of GPLv2~\S2(b) describes what the licensee must do if she is
 
distributing or publishing a work that is deemed a derivative work under
 
copyright law --- namely, the following:
 
\begin{quote}
 
[The work must] be licensed as a whole at no charge to all third parties
 
under the terms of this License.
 
\end{quote}
 
That is probably the most tightly-packed phrase in all of the GPL\@.
 
Consider each subpart carefully.
 

	
 
The work ``as a whole'' is what is to be licensed. This is an important
 
point that GPLv2~\S2 spends an entire paragraph explaining; thus this phrase is
 
worthy of a lengthy discussion here.  As a programmer modifies a software
 
program, she generates new copyrighted material --- fixing expressions of
 
ideas into the tangible medium of electronic file storage.  That
 
programmer is indeed the copyright holder of those new changes.  However,
 
those changes are part and parcel to the original work distributed to
 
the programmer under GPL\@. Thus, the license of the original work
 
affects the license of the new whole derivative work.
 

	
 
% {\cal I}
 
\newcommand{\gplusi}{$\mathcal{G\!\!+\!\!I}$}
 
\newcommand{\worki}{$\mathcal{I}$}
 
\newcommand{\workg}{$\mathcal{G}$}
 

	
 
\label{separate-and-independent}
 

	
 
It is certainly possible to take an existing independent work (called
 
\worki{}) and combine it with a GPL'd program (called \workg{}).  The
 
license of \worki{}, when it is distributed as a separate and independent
 
work, remains the prerogative of the copyright holder of \worki{}.
 
However, when \worki{} is combined with \workg{}, it produces a new work
 
that is the combination of the two (called \gplusi{}). The copyright of
 
this combined work, \gplusi{}, is held by the original copyright
 
holder of each of the two works.
 

	
 
In this case, GPLv2~\S2 lays out the terms by which \gplusi{} may be
 
distributed and copied.  By default, under copyright law, the copyright
 
holder of \worki{} would not have been permitted to distribute \gplusi{};
 
copyright law forbids it without the expressed permission of the copyright
 
holder of \workg{}. (Imagine, for a moment, if \workg{} were a proprietary
 
product --- would its copyright holders  give you permission to create and distribute
 
\gplusi{} without paying them a hefty sum?)  The license of \workg{}, the
 
GPL, states the  options for the copyright holder of \worki{}
 
who may want to create and distribute \gplusi{}.  GPL's pregranted
 
permission to create and distribute derivative works, provided the terms
 
of GPL are upheld, goes far above and beyond the permissions that one
 
would get with a typical work not covered by a copyleft license.  (Thus, to
 
say that this restriction is any way unreasonable is simply ludicrous.)
 

	
 
\medskip
 

	
 
The next phrase of note in GPLv2~\S2(b) is ``licensed \ldots at no charge.''
 
This phrase  confuses many.  The sloppy reader points out this as ``a
 
contradiction in GPL'' because (in their confused view) that clause of GPLv2~\S2 says that redistributors cannot
 
charge for modified versions of GPL'd software, but GPLv2~\S1 says that
 
they can.  Avoid this confusion: the ``at no charge'' \textbf{does not} prohibit redistributors from
 
charging when performing the acts governed by copyright
 
law,\footnote{Recall that you could by default charge for any acts not
 
governed by copyright law, because the license controls are confined
 
by copyright.} but rather that they cannot charge a fee for the
 
\emph{license itself}.  In other words, redistributors of (modified
 
and unmodified) GPL'd works may charge any amount they choose for
 
performing the modifications on contract or the act of transferring
 
the copy to the customer, but they may not charge a separate licensing
 
fee for the software.
 

	
 
GPLv2~\S2(b) further states that the software must ``be licensed \ldots to all
 
third parties.''  This too yields some confusion, and feeds the
 
misconception mentioned earlier --- that all modified versions must made
 
available to the public at large.  However, the text here does not say
 
that.  Instead, it says that the licensing under terms of the GPL must
 
extend to anyone who might, through the distribution chain, receive a copy
 
of the software.  Distribution to all third parties is not mandated here,
 
but GPLv2~\S2(b) does require redistributors to license the derivative works in
 
a way that extends to all third parties who may ultimately receive a
 
copy of the software.
 

	
 
In summary, GPLv2\ 2(b) says what terms under which the third parties must
 
receive this no-charge license.  Namely, they receive it ``under the terms
 
of this License'', the GPLv2.  When an entity \emph{chooses} to redistribute
 
a derivative work of GPL'd software, the license of that whole 
 
work must be GPL and only GPL\@.  In this manner, GPLv2~\S2(b) dovetails nicely
 
with GPLv2~\S6 (as discussed in Section~\ref{GPLv2s6} of this tutorial).
 

	
 
\medskip
 

	
 
The final paragraph of GPLv2~\S2 is worth special mention.  It is possible and
 
quite common to aggregate various software programs together on one
 
distribution medium.  Computer manufacturers do this when they ship a
 
pre-installed hard drive, and GNU/Linux distribution vendors do this to
 
give a one-stop CD or URL for a complete operating system with necessary
 
applications.  The GPL very clearly permits such ``mere aggregation'' with
 
programs under any license.  Despite what you hear from its critics, the
 
GPL is nothing like a virus, not only because the GPL is good for you and
 
a virus is bad for you, but also because simple contact with a GPL'd
 
code-base does not impact the license of other programs.  A programmer must
 
expended actual effort  to cause a work to fall under the terms
 
of the GPL.  Redistributors are always welcome to simply ship GPL'd
 
software alongside proprietary software or other unrelated Free Software,
 
as long as the terms of GPL are adhered to for those packages that are
 
truly GPL'd.
 

	
 
\section{GPLv2~\S3: Producing Binaries}
 
\label{GPL-Section-3}
 

	
 
Software is a strange beast when compared to other copyrightable works.
 
It is currently impossible to make a film or a book that can be truly
 
obscured.  Ultimately, the full text of a novel, even one written by
 
William Faulkner, must presented to the reader as words in some
 
human-readable language so that they can enjoy the work.  A film, even one
 
directed by David Lynch, must be perceptible by human eyes and ears to
 
have any value.
 

	
 
Software is not so.  While the source code --- the human-readable
 
representation of software is of keen interest to programmers -- users and
 
programmers alike cannot make the proper use of software in that
 
human-readable form.  Binary code --- the ones and zeros that the computer
 
can understand --- must be predicable and attainable for the software to
 
be fully useful.  Without the binaries, be they in object or executable
 
form, the software serves only the didactic purposes of computer science.
 

	
 
Under copyright law, binary representations of the software are simply
 
derivative works of the source code.  Applying a systematic process (i.e.,
 
``compilation''\footnote{``Compilation'' in this context refers to the
 
  automated computing process of converting source code into binaries.  It
 
  has absolutely nothing to do with the term ``compilation'' in copyright statues.}) to a work of source code yields binary code. The binary
 
code is now a new work of expression fixed in the tangible medium of
 
electronic file storage.
 

	
 
Therefore, for GPL'd software to be useful, the GPL, since it governs the
 
rules for creation of derivative works, must grant permission for the
 
generation of binaries.  Furthermore, notwithstanding the relative
 
popularity of source-based GNU/Linux distributions like Gentoo, users find
 
it extremely convenient to receive distribution of binary software.  Such
 
distribution is the redistribution of derivative works of the software's
 
source code.  GPLv2~\S3 addresses the matter of creation and distribution of
 
binary versions.
 

	
 
Under GPLv2~\S3, binary versions may be created and distributed under the
 
terms of GPLv2~\S1--2, so all the material previously discussed applies
 
here.  However, GPLv2~\S3 must go a bit further.  Access to the software's
 
source code is an incontestable prerequisite for the exercise of the
 
fundamental freedoms to modify and improve the software.  Making even
 
the most trivial changes to a software program at the binary level is
 
effectively impossible.  GPLv2~\S3 must ensure that the binaries are never
 
distributed without the source code, so that these freedoms are passed
 
through the distribution chain.
 

	
 
GPLv2~\S3 permits distribution of binaries, and then offers three options for
 
distribution of source code along with binaries. The most common and the
 
least complicated is the option given under GPLv2~\S3(a).
 

	
 
GPLv2~\S3(a) offers the option to directly accompany the source code alongside
 
the distribution of the binaries.  This is by far the most convenient
 
option for most distributors, because it means that the source-code
 
provision obligations are fully completed at the time of binary
 
distribution (more on that later).
 

	
 
Under GPLv2~\S3(a), the source code provided must be the ``corresponding source
 
code.''  Here ``corresponding'' primarily means that the source code
 
provided must be that code used to produce the binaries being distributed.
 
That source code must also be ``complete''.   GPLv2~\S3's penultimate paragraph
 
explains in detail what is meant by ``complete''.  In essence, it is all
 
the material that a programmer of average skill would need to actually use
 
the source code to produce the binaries she has received.  Complete source
 
is required so that, if the licensee chooses, she should be able to
 
exercise her freedoms to modify and redistribute changes.  Without the
 
complete source, it would not be possible to make changes that were
 
actually directly derived from the version received.
 

	
 
Furthermore, GPLv2~\S3 is defending against a tactic that has in fact been
 
seen in GPL enforcement.  Under GPL, if you pay a high price for
 
a copy of GPL'd binaries (which comes with corresponding source, of
 
course), you have the freedom to redistribute that work at any fee you
 
choose, or not at all.  Sometimes, companies attempt a GPL-violating
 
cozenage whereby they produce very specialized binaries (perhaps for
 
an obscure architecture).  They then give source code that does
 
correspond, but withhold the ``incantations'' and build plans they
 
used to make that source compile into the specialized binaries.
 
Therefore, GPLv2~\S3 requires that the source code include ``meta-material'' like
 
scripts, interface definitions, and other material that is used to
 
``control compilation and installation'' of the binaries.  In this
 
manner, those further down the distribution chain are assured that
 
they have the unabated freedom to build their own derivative works
 
from the sources provided.
 

	
 
Software distribution comes in many
 
forms.  Embedded manufacturers, for example, have the freedom to put
 
GPL'd software into mobile devices with very tight memory and space
 
constraints.  In such cases, putting the source right alongside the
 
binaries on the machine itself might not be an option.  While it is
 
recommended that this be the default way that people comply with GPL, the
 
GPL does provide options when such distribution is infeasible.
 

	
 
GPLv2~\S3, therefore, allows source code to be provided on any physical
 
``medium customarily used for software interchange.''  By design, this
 
phrase covers a broad spectrum --- the phrase seeks to pre-adapt to
 
changes in  technology.  When GPLv22 was first published in June
 
1991, distribution on magnetic tape was still common, and CD was
 
relatively new.  By 2002, CD is the default.  By 2007, DVD's were the
 
default.  Now, it's common to give software on USB drives and SD card.  This
 
language in the license must adapt with changing technology.
 

	
 
Meanwhile, the binding created by the word ``customarily'' is key.  Many
 
incorrectly believe that distributing binary on CD and source on the
 
Internet is acceptable.  In the corporate world in industrialized countries, it is indeed customary to
 
simply download a CDs' worth of data quickly.  However, even today in the USA, many computer users are not connected to the Internet, and most people connected
 
to the Internet still have limited download speeds.  Downloading
 
CDs full of data is not customary for them in the least.  In some cities
 
in Africa, computers are becoming more common, but Internet connectivity
 
is still available only at a few centralized locations.  Thus, the
 
``customs'' here are normalized for a worldwide userbase.  Simply
 
providing source on the Internet --- while it is a kind, friendly and
 
useful thing to do --- is not usually sufficient.
 

	
 
Note, however, a major exception to this rule, given by the last paragraph
 
of GPLv2~\S3. \emph{If} distribution of the binary files is made only on the
 
Internet (i.e., ``from a designated place''), \emph{then} simply providing
 
the source code right alongside the binaries in the same place is
 
sufficient to comply with GPLv2~\S3.
 

	
 
\medskip
 

	
 
As is shown above, Under GPLv2~\S3(a), embedded manufacturers can put the
 
binaries on the device and ship the source code along on a CD\@.  However,
 
sometimes this turns out to be too costly.  Including a CD with every
 
device could prove too costly, and may practically (although not legally)
 
prohibit using GPL'd software. For this situation and others like it, GPlv2\S~3(b) is available.
 

	
 
GPLv2~\S3(b) allows a distributor of binaries to instead provide a written
 
offer for source code alongside those binaries.  This is useful in two
 
specific ways.  First, it may turn out that most users do not request the
 
source, and thus the cost of producing the CDs is saved --- a financial
 
and environmental windfall.  In addition, along with a GPLv2~\S3(b) compliant
 
offer for source, a binary distributor might choose to \emph{also} give a
 
URL for source code.  Many who would otherwise need a CD with source might
 
turn out to have those coveted high bandwidth connections, and are able to
 
download the source instead --- again yielding environmental and financial
 
windfalls.
 

	
 
However, note that regardless of how many users prefer to get the
 
source online, GPLv2~\S3(b) does place lasting long-term obligations on the
 
binary distributor.  The binary distributor must be prepared to honor
 
that offer for source for three years and ship it out (just as they
 
would have had to do under GPLv2~\S3(a)) at a moment's notice when they
 
receive such a request.  There is real organizational cost here:
 
support engineers must be trained how to route source requests, and
 
source CD images for every release version for the last three years
 
must be kept on hand to burn such CDs quickly. The requests might not
 
even come from actual customers; the offer for source must be valid
 
for ``any third party''.
 

	
 
That phrase is another place where some get confused --- thinking again
 
that full public distribution of source is required.  The offer for source
 
must be valid for ``any third party'' because of the freedoms of
 
redistribution granted by GPLv2~\S\S1--2.  A company may ship a binary image
 
and an offer for source to only one customer.  However, under GPL, that
 
customer has the right to redistribute that software to the world if she
 
likes.  When she does, that customer has an obligation to make sure that
 
those who receive the software from her can exercise their freedoms under
 
GPL --- including the freedom to modify, rebuild, and redistribute the
 
source code.
 

	
 
GPLv2~\S3(c) is created to save her some trouble, because by itself GPLv2~\S3(b)
 
would unfairly favor large companies.  GPLv2~\S3(b) allows the
 
separation of the binary software from the key tool that people can use
 
to exercise their freedom. The GPL permits this separation because it is
 
good for redistributors, and those users who turn out not to need the
 
source.  However, to ensure equal rights for all software users, anyone
 
along the distribution chain must have the right to get the source and
 
exercise those freedoms that require it.
 

	
 
Meanwhile, GPLv2~\S3(b)'s compromise primarily benefits companies who
 
distribute binary software commercially.  Without GPLv2~\S3(c), that benefit
 
would be at the detriment of the companies' customers; the burden of
 
source code provision would be unfairly shifted to the companies'
 
customers.  A customer, who had received binaries with a GPLv2~\S3(b)-compliant
 
offer, would be required under GPLv2 (sans GPLv2~\S3(c)) to acquire the source,
 
merely to give a copy of the software to a friend who needed it.  GPLv2~\S3(c)
 
reshifts this burden to entity who benefits from GPLv2~\S3(b).
 

	
 
GPLv2~\S3(c) allows those who undertake \emph{noncommercial} distribution to
 
simply pass along a GPLv2~\S3(b)-compliant source code offer.  The customer who
 
wishes to give a copy to her friend can now do so without provisioning the
 
source, as long as she gives that offer to her friend.  By contrast, if
 
she wanted to go into business for herself selling CDs of that software,
 
she would have to acquire the source and either comply via GPLv2~\S3(a), or
 
write her own GPLv2~\S3(b)-compliant source offer.
 

	
 
This process is precisely the reason why a GPLv2~\S3(b) source offer must be
 
valid for all third parties.  At the time the offer is made, there is no
 
way of knowing who might end up noncommercially receiving a copy of the
 
software.  Companies who choose to comply via GPLv2~\S3(b) must thus be
 
prepared to honor all incoming source code requests.  For this and the
 
many other additional necessary complications under GPLv2~\S\S3(b--c), it is
 
only rarely a better option than complying via GPLv2~\S3(a).
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{GPL's Implied Patent Grant}
 
\label{gpl-implied-patent-grant}
 

	
 
We digress again briefly from our section-by-section consideration of GPLv2
 
to consider the interaction between the terms of GPL and patent law. The
 
GPLv2, despite being silent with respect to patents, actually confers on its
 
licensees more rights to a licensor's patents than those licenses that
 
purport to address the issue. This is the case because patent law, under
 
the doctrine of implied license, gives to each distributee of a patented
 
article a license from the distributor to practice any patent claims owned
 
or held by the distributor that cover the distributed article. The
 
implied license also extends to any patent claims owned or held by the
 
distributor that cover ``reasonably contemplated uses'' of the patented
 
article. To quote the Federal Circuit Court of Appeals, the highest court
 
for patent cases other than the Supreme Court:
 

	
 
\begin{quotation}
 
Generally, when a seller sells a product without restriction, it in
 
effect promises the purchaser that in exchange for the price paid, it will
 
not interfere with the purchaser's full enjoyment of the product
 
purchased. The buyer has an implied license under any patents of the
 
seller that dominate the product or any uses of the product to which the
 
parties might reasonably contemplate the product will be put.
 
\end{quotation}
 
Hewlett-Packard Co. v. Repeat-O-Type Stencil Mfg. Corp., Inc., 123 F.3d
 
1445 (Fed. Cir. 1997).
 

	
 
Of course, Free Software is licensed, not sold, and there are indeed
 
restrictions placed on the licensee, but those differences are not likely
 
to prevent the application of the implied license doctrine to Free
 
Software, because software licensed under the GPL grants the licensee the
 
right to make, use, and sell the software, each of which are exclusive
 
rights of a patent holder. Therefore, although the GPLv2 does not expressly
 
grant the licensee the right to do those things under any patents the
 
licensor may have that cover the software or its reasonably contemplated
 
uses, by licensing the software under the GPLv2, the distributor impliedly
 
licenses those patents to the GPLv2 licensee with respect to the GPLv2'd
 
software.
 

	
 
An interesting issue regarding this implied patent license of GPLv2'd
 
software is what would be considered ``uses of the [software] to which
 
the parties might reasonably contemplate the product will be put.'' A
 
clever advocate may argue that the implied license granted by GPLv2 is
 
larger in scope than the express license in other Free Software
 
licenses with express patent grants, in that, the patent license
 
clause of many of those licenses are specifically limited to the
 
patent claims covered by the code as licensed by the patentee.
 

	
 
To the contrary, GPLv2's implied patent license grants the GPLv2 licensee a
 
patent license to do much more than just that because the GPLv2 licensee,
 
under the doctrine of implied patent license, is free to practice any
 
patent claims held by the licensor that cover ``reasonably contemplated
 
uses'' of the GPL'd code, which may very well include creation and
 
distribution of derivative works since the GPL's terms, under which the
 
patented code is distributed, expressly permits such activity.
 

	
 
Further supporting this result is the Federal Circuit's pronouncement that
 
the recipient of a patented article has, not only an implied license to
 
make, use, and sell the article, but also an implied patent license to
 
repair the article to enable it to function properly, Bottom Line Mgmt.,
 
Inc. v. Pan Man, Inc., 228 F.3d 1352 (Fed. Cir. 2000). Additionally, the
 
Federal Circuit extended that rule to include any future recipients of the
 
patented article, not just the direct recipient from the distributor.
 
This theory comports well with the idea of Free Software, whereby software
 
is distributed amongst many entities within the community for the purpose
 
of constant evolution and improvement. In this way, the law of implied
 
patent license used by the GPLv2 ensures that the community mutually
 
benefits from the licensing of patents to any single community member.
 

	
 
Note that simply because GPLv2'd software has an implied patent license does
 
not mean that any patents held by a distributor of GPLv2'd code become
 
worthless. To the contrary, the patents are still valid and enforceable
 
against either:
 

	
 
\begin{enumerate}
 
 \renewcommand{\theenumi}{\alph{enumi}}
 
 \renewcommand{\labelenumi}{\textup{(\theenumi)}}
 

	
 
\item any software other than that licensed under the GPLv2 by the patent
 
  holder, and
 

	
 
\item any party that does not comply with the GPLv2
 
with respect to the licensed software.
 
\end{enumerate}
 

	
 
\newcommand{\compB}{$\mathcal{B}$}
 
\newcommand{\compA}{$\mathcal{A}$}
 

	
 
For example, if Company \compA{} has a patent on advanced Web browsing, but
 
also licenses a Web browsing software program under the GPLv2, then it
 
cannot assert the patent against any party that takes a license to its
 
program under the GPLv2. However, if a party uses that program without
 
complying with the GPLv2, then Company \compA{} can assert, not just copyright
 
infringement claims against the non-GPLv2-compliant party, but also
 
infringement of the patent, because the implied patent license only
 
extends to use of the software in accordance with the GPLv2. Further, if
 
Company \compB{} distributes a competitive advanced Web browsing program,
 
Company \compA{} is free to assert its patent against any user or
 
distributor of that product. It is irrelevant whether Company \compB's
 
program is distributed under the GPLv2, as Company \compB{} can not grant
 
implied licenses to Company \compA's patent.
 

	
 
This result also reassures companies that they need not fear losing their
 
proprietary value in patents to competitors through the GPLv2 implied patent
 
license, as only those competitors who adopt and comply with the GPLv2's
 
terms can benefit from the implied patent license. To continue the
 
example above, Company \compB{} does not receive a free ride on Company
 
\compA's patent, as Company \compB{} has not licensed-in and then
 
redistributed Company A's advanced Web browser under the GPLv2. If Company
 
\compB{} does do that, however, Company \compA{} still has not lost
 
competitive advantage against Company \compB{}, as Company \compB{} must then,
 
when it re-distributes Company \compA's program, grant an implied license
 
to any of its patents that cover the program. Further, if Company \compB{}
 
relicenses an improved version of Company A's program, it must do so under
 
the GPLv2, meaning that any patents it holds that cover the improved version
 
are impliedly licensed to any licensee. As such, the only way Company
 
\compB{} can benefit from Company \compA's implied patent license, is if it,
 
itself, distributes Company \compA's software program and grants an
 
implied patent license to any of its patents that cover that program.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Defending Freedom on Many Fronts}
 

	
 
Chapters~\ref{run-and-verbatim} and~\ref{source-and-binary} presented the
 
core freedom-defending provisions of GPLv2\@, which are in GPLv2~\S\S0--3.
 
GPLv2\S\S~4--7 of the GPLv2 are designed to ensure that GPLv2~\S\S0--3 are
 
not infringed, are enforceable, are kept to the confines of copyright law but
 
also  not trumped by other copyright agreements or components of other
 
entirely separate legal systems.  In short, while GPLv2~\S\S0--3 are the parts
 
of the license that defend the freedoms of users and programmers,
 
GPLv2~\S\S4--7 are the parts of the license that keep the playing field clear
 
so that \S\S 0--3 can do their jobs.
 

	
 
\section{GPLv2~\S4: Termination on Violation}
 
\label{GPLv2s4}
 

	
 
GPLv2~\S4 is GPLv2's termination clause.  Upon first examination, it seems
 
strange that a license with the goal of defending users' and programmers'
 
freedoms for perpetuity in an irrevocable way would have such a clause.
 
However, upon further examination, the difference between irrevocability
 
and this termination clause becomes clear.
 

	
 
The GPL is irrevocable in the sense that once a copyright holder grants
 
rights for someone to copy, modify and redistribute the software under terms
 
of the GPL, they cannot later revoke that grant.  Since the GPL has no
 
provision allowing the copyright holder to take such a prerogative, the
 
license is granted as long as the copyright remains in effect.\footnote{In
 
  the USA, due to unfortunate legislation, the length of copyright is nearly
 
  perpetual, even though the Constitution forbids perpetual copyright.} The
 
copyright holders have the right to relicense the same work under different
 
licenses (see Section~\ref{Proprietary Relicensing} of this tutorial), or to
 
stop distributing the GPLv2'd version (assuming GPLv2~\S3(b) was never used),
 
but they may not revoke the rights under GPLv2 already granted.
 

	
 
In fact, when an entity looses their right to copy, modify and distribute
 
GPL'd software, it is because of their \emph{own actions}, not that of the
 
copyright holder.  The copyright holder does not decided when GPLv2~\S4
 
termination occurs (if ever); rather, the actions of the licensee determine
 
that.
 

	
 
Under copyright law, the GPL has granted various rights and freedoms to
 
the licensee to perform specific types of copying, modification, and
 
redistribution.  By default, all other types of copying, modification, and
 
redistribution are prohibited.  GPLv2~\S4 says that if you undertake any of
 
those other types (e.g., redistributing binary-only in violation of GPLv2~\S3),
 
then all rights under the license --- even those otherwise permitted for
 
those who have not violated --- terminate automatically.
 

	
 
GPLv2~\S4 makes GPLv2 enforceable.  If licensees fail to adhere to the
 
license, then they are stuck without any permission under to engage in
 
activities covered by copyright law.  They must completely cease and desist
 
from all copying, modification and distribution of the GPL'd software.
 

	
 
At that point, violating licensees must gain the forgiveness of the copyright
 
holders to have their rights restored.  Alternatively, the violators could
 
negotiate another agreement, separate from GPL, with the copyright
 
holder.  Both are common practice, although
 
\tutorialpartsplit{as discussed in \textit{A Practical Guide to GPL
 
    Compliance}, there are }{Chapter~\ref{compliance-understanding-whos-enforcing}
 
  explains further } key differences between these two very different uses of GPL.
 

	
 
\section{GPLv2~\S5: Acceptance, Copyright Style}
 
\label{GPLv2s5}
 

	
 
GPLv2~\S5 brings us to perhaps the most fundamental misconception and common
 
confusion about GPLv2\@. Because of the prevalence of proprietary software,
 
most users, programmers, and lawyers alike tend to be more familiar with
 
EULAs. EULAs are believed by their authors to be contracts, requiring
 
formal agreement between the licensee and the software distributor to be
 
valid. This has led to mechanisms like ``shrink-wrap'' and ``click-wrap''
 
as mechanisms to perform acceptance ceremonies with EULAs.
 

	
 
The GPL does not need contract law to ``transfer rights.''  No rights
 
are transfered between parties. By contrast, the GPL is a permission
 
The GPL does not need contract law to ``transfer rights.''  Usually, no rights
 
are transfered between parties.  By contrast, the GPL is primarily a permission
 
slip to undertake activities that would otherwise have been prohibited
 
by copyright law. As such, it needs no acceptance ceremony; the
 
by copyright law.  As such, GPL needs no acceptance ceremony; the
 
licensee is not even required to accept the license.
 

	
 
However, without the GPL, the activities of copying, modifying and
 
distributing the software would have otherwise been prohibited.  So, the
 
GPL says that you only accepted the license by undertaking activities that
 
you would have otherwise been prohibited without your license under GPL\@.
 
This is a certainly subtle point, and requires a mindset quite different
 
from the contractual approach taken by EULA authors.
 

	
 
An interesting side benefit to GPLv2~\S5 is that the bulk of users of Free
 
Software are not required to accept the license.  Undertaking fair and
 
unregulated use of the work, for example, does not bind you to the GPL,
 
since you are not engaging in activity that is otherwise controlled by
 
copyright law.  Only when you engage in those activities that might have an
 
impact on the freedom of others does license acceptance occur, and the
 
terms begin to bind you to fair and equitable sharing of the software.  In
 
other words, the GPL only kicks in when it needs to for the sake of
 
freedom.
 

	
 
\section{Using GPL Both as a Contract and Copyright License}
 

	
 
\section{GPLv2~\S6: GPL, My One and Only}
 
\label{GPLv2s6}
 

	
 
A point that was glossed over in Section~\ref{GPLv2s4}'s discussion of GPLv2~\S4
 
was the irrevocable nature of the GPL\@. The GPLv2 is indeed irrevocable,
 
and it is made so formally by GPLv2~\S6.
 

	
 
The first sentence in GPLv2~\S6 ensures that as software propagates down the
 
distribution chain, that each licensor can pass along the license to each
 
new licensee. Under GPLv2~\S6, the act of distributing automatically grants a
 
license from the original licensor to the next recipient. This creates a
 
chain of grants that ensure that everyone in the distribution has rights
 
under the GPLv2\@. In a mathematical sense, this bounds the bottom ---
 
making sure that future licensees get no fewer rights than the licensee before.
 

	
 
The second sentence of GPLv2~\S6 does the opposite; it bounds from the top. It
 
prohibits any licensor along the distribution chain from placing
 
additional restrictions on the user. In other words, no additional
 
requirements may trump the rights and freedoms given by GPLv2\@.
 

	
 
The final sentence of GPLv2~\S6 makes it abundantly clear that no individual
 
entity in the distribution chain is responsible for the compliance of any
 
other. This is particularly important for noncommercial users who have
 
passed along a source offer under GPLv2~\S3(c), as they cannot be assured that
 
the issuer of the offer will honor their GPLv2~\S3 obligations.
 

	
 
In short, GPLv2~\S6 says that your license for the software is your one and
 
only copyright license allowing you to copy, modify and distribute the
 
software.
 

	
 
\section{GPLv2~\S7: ``Give Software Liberty or Give It Death!''}
 
\label{GPLv2s7}
 

	
 
In essence, GPLv2~\S7 is a verbosely worded way of saying for non-copyright
 
systems what GPLv2~\S6 says for copyright. If there exists any reason that a
 
distributor knows of that would prohibit later licensees from exercising
 
their full rights under GPL, then distribution is prohibited.
 

	
 
Originally, this was designed as the title of this section suggests --- as
 
a last ditch effort to make sure that freedom was upheld. However, in
 
modern times, it has come to give much more. Now that the body of GPL'd
 
software is so large, patent holders who would want to be distributors of
 
GPL'd software have a tough choice. They must choose between avoiding
 
distribution of GPL'd software that exercises the teachings of their
 
patents, or grant a royalty-free, irrevocable, non-exclusive license to
 
those patents. Many companies, including IBM, the largest patent holder
 
in the world, have chosen the latter.
 

	
 
Thus, GPLv2~\S7 rarely gives software death by stopping its distribution.
 
Instead, it is inspiring patent holders to share their patents in the same
 
freedom-defending way that they share their copyrighted works.
 

	
 
\section{GPLv2~\S8: Excluding Problematic Jurisdictions}
 
\label{GPLv2s8}
 

	
 
GPLv2~\S8 is rarely used by copyright holders. Its intention is that if a
 
particular country, say Unfreedonia, grants particular patents or allows
 
copyrighted interfaces (no country to our knowledge even permits those
 
yet), that the GPLv2'd software can continue in free and unabated
 
distribution in the countries where such controls do not exist.
 

	
 
It is a partial ``out'' from GPLv2~\S7. Without GPLv2~\S8, if a copyright holder
 
knew of a patent in a particular country licensed in a GPL-incompatible
 
way, then she could not distribute under GPL, because the work could
 
legitimately end up in the hands of citizens of Unfreedonia.
 

	
 
It is an inevitable but sad reality that some countries are freer than
 
others. GPLv2~\S8 exists to permit distribution in those countries that are
 
free without otherwise negating parts of the license.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Odds, Ends, and Absolutely No Warranty}
 

	
 
GPLv2~\S\S0--7 constitute the freedom-defending terms of the GPLv2. The remainder
 
of the GPLv2 handles administrivia and issues concerning warranties and
 
liability.
 

	
 
\section{GPLv2~\S9: FSF as Stewards of GPL}
 
\label{GPLv2s9}
 

	
 
FSF reserves the exclusive right to publish future versions of the GPL\@;
 
GPLv2~\S9 expresses this. While the stewardship of the copyrights on the body
 
of GPL'd software around the world is shared among thousands of
 
individuals and organizations, the license itself needs a single steward.
 
Forking of the code is often regrettable but basically innocuous. Forking
 
of licensing is disastrous.
 

	
 
FSF has only released two versions of GPL --- in 1989 and 1991. GPL
 
version 3 is under current internal drafting. FSF's plan is to have a
 
long and engaging comment period. The goal of GPL is to defend freedom, and
 
a gigantic community depends on that freedom now. FSF hopes to take all
 
stakeholders' opinions under advisement.
 

	
 
\section{GPLv2~\S10: Relicensing Permitted}
 
\label{GPLv2s10}
 

	
 
GPLv2~\S10 reminds the licensee of what is already implied by the nature of
 
copyright law. Namely, the copyright holder of a particular software
 
program has the prerogative to grant alternative agreements under separate
 
copyright licenses.
 

	
 
\section{GPLv2~\S11: No Warranty}
 
\label{GPLv2s11}
 

	
 
All warranty disclaimer language tends to be shouted in all capital
 
letters. Apparently, there was once a case where the disclaimer language
 
of an agreement was negated because it was not ``conspicuous'' to one of
 
the parties. Therefore, to make such language ``conspicuous,'' people
 
started placing it in bold or capitalizing the entire text. It now seems
 
to be voodoo tradition of warranty disclaimer writing.
 

	
 
Some have argued the GPL is unenforceable in some jurisdictions because
 
its disclaimer of warranties is impermissibly broad. However, GPLv2~\S11
 
contains a jurisdictional savings provision, which states that it is to be
 
interpreted only as broadly as allowed by applicable law. Such a
 
provision ensures that both it, and the entire GPL, is enforceable in any
 
jurisdiction, regardless of any particular law regarding the
 
permissibility of certain warranty disclaimers.
 

	
 
Finally, one important point to remember when reading GPLv2~\S11 is that GPLv2~\S1
 
permits the sale of warranty as an additional service, which GPLv2~\S11 affirms.
 

	
 
\section{GPLv2~\S12: Limitation of Liability}
 
\label{GPLv2s12}
 

	
 
There are many types of warranties, and in some jurisdictions some of them
 
cannot be disclaimed. Therefore, usually agreements will have both a
 
warranty disclaimer and a limitation of liability, as we have in GPLv2~\S12. \S
 
11 thus gets rid of all implied warranties that can legally be
 
disavowed. GPLv2~\S12, in turn, limits the liability of the actor for any
 
warranties that cannot legally be disclaimed in a particular jurisdiction.
 

	
 
Again, some have argued the GPL is unenforceable in some jurisdictions
 
because its limitation of liability is impermissibly broad. However, \S
 
12, just like its sister, GPLv2~\S11, contains a jurisdictional savings
 
provision, which states that it is to be interpreted only as broadly as
 
allowed by applicable law. As stated above, such a provision ensures that
 
both GPLv2~\S12, and the entire GPL, is enforceable in any jurisdiction,
 
regardless of any particular law regarding the permissibility of limiting
 
liability.
 

	
 
So end the terms and conditions of the GNU General Public License.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{GPLv3}
 
\label{GPLv3}
 

	
 
\section{Understanding GPLv3 As An Upgraded GPLv2}
 

	
 
\section{GPLv3~\S0: Giving In On ``Defined Terms''}
 

	
 
\section{GPLv3~\S1: Understanding CCS}
 

	
 
\section{GPLv3~\S2: Basic Permissions}
 

	
 
\section{GPLv3~\S3: What Hath DMCA Wrought}
 
\label{GPLv3s3}
 
\section{GPLv3~\S4: Verbatim Copying}
 

	
 
\section{GPLv3~\S5: Modified Source}
 

	
 
\section{GPLv3~\S6: Non-Source and Corresponding Source}
 

	
 
\section{Understanding License Compatibility}
 
\label{license-compatibility}
 

	
 
\section{GPLv3~\S7: Explicit Compatibility}
 

	
 
\section{GPLv3~\S8: A Lighter Termination}
 

	
 
\section{GPLv3~\S9: Acceptance}
 

	
 
\section{GPLv3~\S10: Explicit Downstream License}
 

	
 
\section{GPLv3~\S11: Explicit Patent Licensing}
 
\label{GPLv3s11}
 
\section{GPLv3~\S12: Familiar as GPLv2 \S 7}
 

	
 
\section{GPLv3~\S13: The Great Affero Compromise}
 

	
 
\section{GPLv3~\S14: So, When's GPLv4?}
 
\label{GPlv2s14}
 
\section{GPLv3~\S15--17: Warranty Disclaimers and Liability Limitation}
 

	
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{The Lesser GPL}
 

	
 
As we have seen in our consideration of the GPL, its text is specifically
 
designed to cover all possible derivative works under copyright law. Our
 
goal in designing GPL was to make sure that any derivative work of GPL'd
 
software was itself released under GPL when distributed. Reaching as far
 
as copyright law will allow is the most direct way to reach that goal.
 

	
 
However, while the strategic goal is to bring as much Free Software
 
into the world as possible, particular tactical considerations
 
regarding software freedom dictate different means. Extending the
 
copyleft effect as far as copyright law allows is not always the most
 
prudent course in reaching the goal. In particular situations, even
 
those of us with the goal of building a world where all published
 
software is Free Software realize that full copyleft does not best
 
serve us. The GNU Lesser General Public License (``GNU LGPL'') was
 
designed as a solution for such situations.
 

	
 
\section{The First LGPL'd Program}
 

	
 
The first example that FSF encountered where such altered tactics were
 
needed was when work began on the GNU C Library. The GNU C Library would
 
become (and today, now is) a drop-in replacement for existing C libraries.
 
On a Unix-like operating system, C is the lingua franca and the C library
 
is an essential component for all programs. It is extremely difficult to
 
construct a program that will run with ease on a Unix-like operating
 
system without making use of services provided by the C library --- even
 
if the program is written in a language other than C\@. Effectively, all
 
user application programs that run on any modern Unix-like system must
 
make use of the C library.
 

	
 
By the time work began on the GNU implementation of the C libraries, there
 
were already many C libraries in existence from a variety of vendors.
 
Every proprietary Unix vendor had one, and many third parties produced
 
smaller versions for special purpose use. However, our goal was to create
 
a C library that would provide equivalent functionality to these other C
 
libraries on a Free Software operating system (which in fact happens today
 
on modern GNU/Linux systems, which all use the GNU C Library).
 

	
 
Unlike existing GNU application software, however, the licensing
 
implications of releasing the GNU C Library (``glibc'') under GPL were
 
somewhat different. Applications released under GPL would never
 
themselves become part of proprietary software. However, if glibc were
 
released under GPL, it would require that any application distributed for
 
the GNU/Linux platform be released under GPL\@.
 

	
 
Since all applications on a Unix-like system depend on the C library, it
 
means that they must link with that library to function on the system. In
 
other words, all applications running on a Unix-like system must be
 
combined with the C library to form a new whole derivative work that is
 
composed of the original application and the C library. Thus, if glibc
 
were GPL'd, each and every application distributed for use on GNU/Linux
 
would also need to be GPL'd, since to even function, such applications
 
would need to be combined into larger derivative works by linking with
 
glibc.
 

	
 
At first glance, such an outcome seems like a windfall for Free Software
 
advocates, since it stops all proprietary software development on
 
GNU/Linux systems. However, the outcome is a bit more subtle. In a world
 
where many C libraries already exist, many of which could easily be ported
 
to GNU/Linux, a GPL'd glibc would be unlikely to succeed. Proprietary
 
vendors would see the excellent opportunity to license their C libraries
 
to anyone who wished to write proprietary software for GNU/Linux systems.
 
The de-facto standard for the C library on GNU/Linux would likely be not
 
glibc, but the most popular proprietary one.
 

	
 
Meanwhile, the actual goal of releasing glibc under GPL --- to ensure no
 
proprietary applications on GNU/Linux --- would be unattainable in this
 
scenario. Furthermore, users of those proprietary applications would also
 
be users of a proprietary C library, not the Free glibc.
 

	
 
The Lesser GPL was initially conceived to handle this scenario. It was
 
clear that the existence of proprietary applications for GNU/Linux was
 
inevitable. Since there were so many C libraries already in existence, a
 
new one under GPL would not stop that tide. However, if the new C library
 
were released under a license that permitted proprietary applications
 
to link with it, but made sure that the library itself remained Free,
 
an ancillary goal could be met. Users of proprietary applications, while
 
they would not have the freedom to copy, share, modify and redistribute
 
the application itself, would have the freedom to do so with respect to
 
the C library.
 

	
 
There was no way the license of glibc could stop or even slow the creation
 
of proprietary applications on GNU/Linux. However, loosening the
 
restrictions on the licensing of glibc ensured that nearly all proprietary
 
applications at least used a Free C library rather than a proprietary one.
 
This trade-off is central to the reasoning behind the LGPL\@.
 

	
 
Of course, many people who use the LGPL today are not thinking in these
 
terms. In fact, they are often choosing the LGPL because they are looking
 
for a ``compromise'' between the GPL and the X11-style liberal licensing.
 
However, understanding FSF's reasoning behind the creation of the LGPL is
 
helpful when studying the license.
 

	
 

	
 
\section{What's the Same?}
 

	
 
Much of the text of the LGPL is identical to the GPL\@. As we begin our
 
discussion of the LGPL, we will first eliminate the sections that are
 
identical, or that have the minor modification changing the word
 
``Program'' to ``Library.''
 

	
 
First, LGPLv2.1~\S1, the rules for verbatim copying of source, are
 
equivalent to those in GPLv2~\S1.
 

	
 
Second, LGPLv2.1~\S8 is equivalent GPLv2~\S4\@. In both licenses, this
 
section handles termination in precisely the same manner.
 

	
 
LGPLv2.1~\S9 is equivalent to GPLv2~\S5\@. Both sections assert that
 
the license is a copyright license, and handle the acceptance of those
 
copyright terms.
 

	
 
LGPLv2.1~\S10 is equivalent to GPLv2~\S6. They both protect the
 
distribution system of Free Software under these licenses, to ensure that
 
up, down, and throughout the distribution chain, each recipient of the
 
software receives identical rights under the license and no other
 
restrictions are imposed.
 

	
 
LGPLv2.1~\S11 is GPLv2~\S7. As discussed, it is used to ensure that
 
other claims and legal realities, such as patent licenses and court
 
judgments, do not trump the rights and permissions granted by these
 
licenses, and requires that distribution be halted if such a trump is
 
known to exist.
 

	
 
LGPLv2.1~\S12 adds the same features as GPLv2~\S8. These sections are
 
used to allow original copyright holders to forbid distribution in
 
countries with draconian laws that would otherwise contradict these
 
licenses.
 

	
 
LGPLv2.1~\S13 sets up FSF as the steward of the LGPL, just as GPLv2~\S9
 
does for GPL. Meanwhile, LGPLv2.1~\S14 reminds licensees that copyright
 
holders can grant exceptions to the terms of LGPL, just as GPLv2~\S10
 
reminds licensees of the same thing.
 

	
 
Finally, the assertions of no warranty and limitations of liability are
 
identical; thus LGPLv2.1~\S15 and LGPLv2.1~\S16 are the same as GPLv2~\S11 and \S
 
12.
 

	
 
As we see, the entire latter half of the license is identical.
 
The parts which set up the legal boundaries and meta-rules for the license
 
are the same. It is our intent that the two licenses operate under the
 
same legal mechanisms and are enforced precisely the same way.
 

	
 
We strike a difference only in the early portions of the license.
 
Namely, in the LGPL we go into deeper detail of granting various permissions to
 
create derivative works, so the redistributors can make
 
some proprietary derivatives. Since we simply do not allow the
 
license to stretch as far as copyright law does regarding what
 
derivative works must be relicensed under the same terms, we must go
 
further to explain which derivative works we will allow to be
 
proprietary. Thus, we'll see that the front matter of the LGPL is a
 
bit more wordy and detailed with regards to the permissions granted to
 
those who modify or redistribute the software.
 

	
 
\section{Additions to the Preamble}
 

	
 
Most of LGPL's Preamble is identical, but the last seven paragraphs
 
introduce the concepts and reasoning behind creation of the license,
 
presenting a more generalized and briefer version of the story with which
 
we began our consideration of LGPL\@.
 

	
 
In short, FSF designed LGPL for those edge cases where the freedom of the
 
public can better be served by a more lax licensing system. FSF doesn't
 
encourage use of LGPL automatically for any software that happens to be a
 
library; rather, FSF suggests that it only be used in specific cases, such
 
as the following:
 

	
 
\begin{itemize}
 

	
 
\item To encourage the widest possible use of a Free Software library, so
 
  it becomes a de-facto standard over similar, although not
 
  interface-identical, proprietary alternatives
 

	
 
\item To encourage use of a Free Software library that already has
 
  interface-identical proprietary competitors that are more developed
 

	
 
\item To allow a greater number of users to get freedom, by encouraging
 
  proprietary companies to pick a Free alternative for its otherwise
 
  proprietary products
 

	
 
\end{itemize}
 

	
 
LGPL's preamble sets forth the limits to which the license seeks to go in
 
chasing these goals. LGPL is designed to ensure that users who happen to
 
acquire software linked with such libraries have full freedoms with
 
respect to that library. They should have the ability to upgrade to a newer
 
or modified Free version or to make their own modifications, even if they
 
cannot modify the primary software program that links to that library.
 

	
 
Finally, the preamble introduces two terms used throughout the license to
 
clarify between the different types of derivative works: ``works that use
 
the library,'' and ``works based on the library.''  Unlike GPL, LGPL must
 
draw some lines regarding derivative works. We do this here in this
 
license because we specifically seek to liberalize the rights afforded to
 
those who make derivative works. In GPL, we reach as far as copyright law
 
allows. In LGPL, we want to draw a line that allows some derivative works
 
copyright law would otherwise prohibit if the copyright holder exercised
 
his full permitted controls over the work.
 

	
 
\section{An Application: A Work that Uses the Library}
 

	
 
In the effort to allow certain proprietary derivative works and prohibit
 
others, LGPL distinguishes between two classes of derivative works:
 
``works based on the library,'' and ``works that use the library.''  The
 
distinction is drawn on the bright line of binary (or runtime) derivative
 
works and source code derivatives. We will first consider the definition
 
of a ``work that uses the library,'' which is set forth in LGPLv2.1~\S5.
 

	
 
We noted in our discussion of GPLv2~\S3 (discussed in
 
Section~\ref{GPL-Section-3} of this document) that binary programs when
 
compiled and linked with GPL'd software are derivative works of that GPL'd
 
software. This includes both linking that happens at compile-time (when
 
the binary is created) or at runtime (when the binary -- including library
 
and main program both -- is loaded into memory by the user). In GPL,
 
binary derivative works are controlled by the terms of the license (in GPLv2~\S3),
 
and distributors of such binary derivatives must release full
 
corresponding source\@.
 

	
 
In the case of LGPL, these are precisely the types of derivative works
 
we wish to permit. This scenario, defined in LGPL as ``a work that uses
 
the library,'' works as follows:
 

	
 
\newcommand{\workl}{$\mathcal{L}$}
 
\newcommand{\lplusi}{$\mathcal{L\!\!+\!\!I}$}
 

	
 
\begin{itemize}
 

	
 
\item A new copyright holder creates a separate and independent work,
 
  \worki{}, that makes interface calls (e.g., function calls) to the
 
  LGPL'd work, called \workl{}, whose copyright is held by some other
 
  party. Note that since \worki{} and \workl{} are separate and
 
  independent works, there is no copyright obligation on this new copyright
 
  holder with regard to the licensing of \worki{}, at least with regard to
 
  the source code.
 

	
 
\item The new copyright holder, for her software to be useful, realizes
 
  that it cannot run without combining \worki{} and \workl{}.
 
  Specifically, when she creates a running binary program, that running
 
  binary must be a derivative work, called \lplusi{}, that the user can
 
  run.
 

	
 
\item Since \lplusi{} is a derivative work of both \worki{} and \workl{},
 
  the license of \workl{} (the LGPL) can put restrictions on the license
 
  of \lplusi{}. In fact, this is what LGPL does.
 

	
 
\end{itemize}
 

	
 
We will talk about the specific restrictions LGPLv2.1 places on ``works
 
that use the library'' in detail in Section~\ref{lgpl-section-6}. For
 
now, focus on the logic related to how the LGPLv2.1 places requirements on
 
the license of \lplusi{}. Note, first of all, the similarity between
 
this explanation and that in Section~\ref{separate-and-independent},
 
which discussed the combination of otherwise separate and independent
 
works with GPL'd code. Effectively, what LGPLv2.1 does is say that when a
 
new work is otherwise separate and independent, but has interface
 
calls out to an LGPL'd library, then it is considered a ``work that
 
uses the library.''
 

	
 
In addition, the only reason that LGPLv2.1 has any control over the licensing
 
of a ``work that uses the library'' is for the same reason that GPL has
 
some say over separate and independent works. Namely, such controls exist
 
because the {\em binary combination\/} (\lplusi{}) that must be created to
 
make the separate work (\worki{}) at all useful is a derivative work of
 
the LGPLv2.1'd software (\workl{}).
 

	
 
Thus, a two-question test that will help indicate if a particular work is
 
a ``work that uses the library'' under LGPLv2.1 is as follows:
 

	
 
\begin{enumerate}
 

	
 
\item Is the source code of the new copyrighted work, \worki{}, a
 
  completely independent work that stands by itself, and includes no
 
  source code from \workl{}?
 

	
 
\item When the source code is compiled, does it create a derivative work
 
  by combining with \workl{}, either by static (compile-time) or dynamic
 
  (runtime) linking, to create a new binary work, \lplusi{}?
 
\end{enumerate}
 

	
 
If the answers to both questions are ``yes,'' then \worki{} is most likely
 
a ``work that uses the library.''  If the answer to the first question
 
``yes,'' but the answer to the second question is ``no,'' then most likely
 
\worki{} is neither a ``work that uses the library'' nor a ``work based on
 
the library.''  If the answer to the first question is ``no,'' but the
 
answer to the second question is ``yes,'' then an investigation into
 
whether or not \worki{} is in fact a ``work based on the library'' is
 
warranted.
 

	
 
\section{The Library, and Works Based On It}
 

	
 
In short, a ``work based on the library'' could be defined as any
 
derivative work of LGPL'd software that cannot otherwise fit the
 
definition of a ``work that uses the library.''  A ``work based on the
 
library'' extends the full width and depth of copyright derivative works,
 
in the same sense that GPL does.
 

	
 
Most typically, one creates a ``work based on the library'' by directly
 
modifying the source of the library. Such a work could also be created by
 
tightly integrating new software with the library. The lines are no doubt
 
fuzzy, just as they are with GPL'd works, since copyright law gives us no
 
litmus test for derivative works of a software program.
 

	
 
Thus, the test to use when considering whether something is a ``work
 
based on the library'' is as follows:
 

	
 
\begin{enumerate}
 

	
 
\item Is the new work, when in source form, a derivative work under
 
  copyright law of the LGPL'd work?
 

	
 
\item Is there no way in which the new work fits the definition of a
 
  ``work that uses the library''?
 
\end{enumerate}
 

	
 

	
 
If the answer is ``yes'' to both these questions, then you most likely
 
have a ``work based on the library.''  If the answer is ``no'' to the
 
first but ``yes'' to the second, you are in a gray area between ``work
 
based on the library'' and a ``work that uses the library.''
 

	
 
In our years of work with the LGPLv2.1, however, we have never seen a work
 
of software that was not clearly one or the other; the line is quite
 
bright. At times, though, we have seen cases where a derivative work
 
appeared in some ways to be a work that used the library and in other
 
ways a work based on the library. We overcame this problem by
 
dividing the work into smaller subunits. It was soon discovered that
 
what we actually had were three distinct components: the original
 
LGPL'd work, a specific set of works that used that library, and a
 
specific set of works that were based on the library. Once such
 
distinctions are established, the licensing for each component can be
 
considered independently and the LGPLv2.1 applied to each work as
 
prescribed.
 

	
 

	
 
\section{Subtleties in Defining the Application}
 

	
 
In our discussion of the definition of ``works that use the library,'' we
 
left out a few more complex details that relate to lower-level programming
 
details. The fourth paragraph of LGPLv2.1~\S5 covers these complexities,
 
and it has been a source of great confusion. Part of the confusion comes
 
because a deep understanding of how compiler programs work is nearly
 
mandatory to grasp the subtle nature of what LGPLv2.1~\S5, \P 4 seeks to
 
cover. It helps some to note that this is a border case that we cover in
 
the license only so that when such a border case is hit, the implications
 
of using LGPL continue in the expected way.
 

	
 
To understand this subtle point, we must recall the way that a compiler
 
operates. The compiler first generates object code, which are the binary
 
representations of various programming modules. Each of those modules is
 
usually not useful by itself; it becomes useful to a user of a full program
 
when those modules are {\em linked\/} into a full binary executable.
 

	
 
As we have discussed, the assembly of modules can happen at compile-time
 
or at runtime. Legally, there is no distinction between the two --- both
 
create a derivative work by copying and combining portions of one work and
 
mixing them with another. However, under LGPL, there is a case in the
 
compilation process where the legal implications are different.
 
Specifically, while we know that a ``work that uses the library'' is one
 
whose final binary is a derivative work, but whose source is not, there
 
are cases where the object code --- that intermediate step between source
 
and final binary --- is a derivative work created by copying verbatim code
 
from the LGPL'd software.
 

	
 
For efficiency, when a compiler turns source code into object code, it
 
sometimes places literal portions of the copyrighted library code into the
 
object code for an otherwise separate independent work. In the normal
 
scenario, the derivative would not be created until final assembly and
 
linking of the executable occurred. However, when the compiler does this
 
efficiency optimization, at the intermediate object code step, a
 
derivative work is created.
 

	
 
LGPLv2.1~\S5\P4 is designed to handle this specific case. The intent of
 
the license is clearly that simply compiling software to ``make use'' of
 
the library does not in itself cause the compiled work to be a ``work
 
based on the library.''  However, since the compiler copies verbatim,
 
copyrighted portions of the library into the object code for the otherwise
 
separate and independent work, it would actually cause that object file to be a
 
``work based on the library.''  It is not FSF's intent that a mere
 
compilation idiosyncrasy would change the requirements on the users of the
 
LGPLv2.1'd software. This paragraph removes that restriction, allowing the
 
implications of the license to be the same regardless of the specific
 
mechanisms the compiler uses underneath to create the ``work that uses the
 
library.''
 

	
 
As it turns out, we have only once had anyone worry about this specific
 
idiosyncrasy, because that particular vendor wanted to ship object code
 
(rather than final binaries) to their customers and was worried about
 
this edge condition. The intent of clarifying this edge condition is
 
primarily to quell the worries of software engineers who understand the
 
level of verbatim code copying that a compiler often does, and to help
 
them understand that the full implications of LGPLv2.1 are the same regardless
 
of the details of the compilation progress.
 

	
 
\section{LGPLv2.1~\S6 \& LGPLv2.1~\S5: Combining the Works}
 
\label{lgpl-section-6}
 
Now that we have established a good working definition of works that
 
``use'' and works that ``are based on'' the library, we will consider the
 
rules for distributing these two different works.
 

	
 
The rules for distributing ``works that use the library'' are covered in
 
LGPLv2.1~\S6\@. LGPLv2.1~\S6 is much like GPLv2~\S3, as it requires the release
 
of source when a binary version of the LGPL'd software is released. Of
 
course, it only requires that source code for the library itself be made
 
available. The work that ``uses'' the library need not be provided in
 
source form. However, there are also conditions in LGPLv2.1~\S6 to make sure
 
that a user who wishes to modify or update the library can do so.
 

	
 
LGPLv2.1~\S6 lists five choices with regard to supplying library source
 
and granting the freedom to modify that library source to users. We
 
will first consider the option given by \S 6(b), which describes the
 
most common way currently used for LGPLv2.1 compliance on a ``work that
 
uses the library.''
 

	
 
LGPLv2.1~\S6(b) allows the distributor of a ``work that uses the library'' to
 
simply use a dynamically linked, shared library mechanism to link with the
 
library. This is by far the easiest and most straightforward option for
 
distribution. In this case, the executable of the work that uses the
 
library will contain only the ``stub code'' that is put in place by the
 
shared library mechanism, and at runtime the executable will combine with
 
the shared version of the library already resident on the user's computer.
 
If such a mechanism is used, it must allow the user to upgrade and
 
replace the library with interface-compatible versions and still be able
 
to use the ``work that uses the library.''  However, all modern shared
 
library mechanisms function as such, and thus LGPLv2.1~\S6(b) is the simplest
 
option, since it does not even require that the distributor of the ``work
 
2based on the library'' ship copies of the library itself.
 

	
 
LGPLv2.1~\S6(a) is the option to use when, for some reason, a shared library
 
mechanism cannot be used. It requires that the source for the library be
 
included, in the typical GPL fashion, but it also has a requirement beyond
 
that. The user must be able to exercise her freedom to modify the library
 
to its fullest extent, and that means recombining it with the ``work based
 
on the library.''  If the full binary is linked without a shared library
 
mechanism, the user must have available the object code for the ``work
 
based on the library,'' so that the user can relink the application and
 
build a new binary.
 

	
 
The remaining options in LGPLv2.1~\S6 are very similar to the other choices
 
provided by GPLv2~\S3. There are some additional options, but time does
 
not permit us in this course to go into those additional options. In
 
almost all cases of distribution under LGPL, either LGPLv2.1~\S6(a) or LGPLv2.1~\S6(b) are
 
exercised.
 

	
 
\section{Distribution of the Combined Works}
 

	
 
Essentially, ``works based on the library'' must be distributed under the
 
same conditions as works under full GPL\@. In fact, we note that 
 
LGPLv2.1~\S2 is nearly identical in its terms and requirements to GPLv2~\S2.
 
There are again subtle differences and additions, which time does not
 
permit us to cover in this course.
 

	
 
\section{And the Rest}
 

	
 
The remaining variations between LGPL and GPL cover the following
 
conditions:
 

	
 
\begin{itemize}
 

	
 
\item Allowing a licensing ``upgrade'' from LGPL to GPL\@ (in LGPLv2.1~\S3)
 

	
 
\item Binary distribution of the library only, covered in LGPLv2.1~\S4,
 
  which is effectively equivalent to LGPLv2.1~\S3
 

	
 
\item Creating aggregates of libraries that are not derivative works of
 
  each other, and distributing them as a unit (in LGPLv2.1~\S7)
 

	
 
\end{itemize}
 

	
 

	
 
Due to time constraints, we cannot cover these additional terms in detail,
 
but they are mostly straightforward. The key to understanding LGPLv2.1 is
 
understanding the difference between a ``work based on the library'' and a
 
``work that uses the library.''  Once that distinction is clear, the
 
remainder of LGPLv2.1 is close enough to GPL that the concepts discussed in
 
our more extensive GPL unit can be directly applied.
 

	
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
\chapter{Integrating the GPL into Business Practices}
 

	
 
Since GPL'd software is now extremely prevalent through the industry, it
 
is useful to have some basic knowledge about using GPL'd software in
 
business and how to build business models around GPL'd software.
 

	
 
\section{Using GPL'd Software In-House}
 

	
 
As discussed in Sections~\ref{GPLv2s0} and~\ref{GPLv2s5} of this tutorial,
 
the GPL only governs the activities of copying, modifying and
 
distributing software programs that are not governed by the license.
 
Thus, in FSF's view, simply installing the software on a machine and
 
using it is not controlled or limited in any way by GPL\@. Using Free
 
Software in general requires substantially fewer agreements and less
 
license compliance activity than any known proprietary software.
 

	
 
Even if a company engages heavily in copying the software throughout the
 
enterprise, such copying is not only permitted by GPLv2~\S\S1 and 3, but it is
 
encouraged!  If the company simply deploys unmodified (or even modified)
 
Free Software throughout the organization for its employees to use, the
 
obligations under the license are very minimal. Using Free Software has a
 
substantially lower cost of ownership --- both in licensing fees and in
 
licensing checking and handling -- than the proprietary software
 
equivalents.
 

	
 
\section{Business Models}
 
\label{Business Models}
 

	
 
Using Free Software in house is certainly helpful, but a thriving
 
market for Free Software-oriented business models also exists. There is the
 
traditional model of selling copies of Free Software distributions.
 
Many companies, including IBM and Red Hat, make substantial revenue
 
from this model. IBM primarily chooses this model because they have
 
found that for higher-end hardware, the cost of the profit made from
 
proprietary software licensing fees is negligible. The real profit is
 
in the hardware, but it is essential that software be stable, reliable
 
and dependable, and the users be allowed to have unfettered access to
 
it. Free Software, and GPL'd software in particular (because IBM can
 
be assured that proprietary versions of the same software will not
 
exists to compete on their hardware) is the right choice.
 

	
 
Red Hat has actually found that a ``convenience fee'' for Free Software,
 
when set at a reasonable price (around \$60 or so), can produce some
 
profit. Even though Red Hat's system is fully downloadable on their
 
Web site, people still go to local computer stores and buy copies of their
 
box set, which is simply a printed version of the manual (available under
 
a Free license as well) and the Free Software system it documents.
 

	
 
\medskip
 

	
 
However, custom support, service, and software improvement contracts
 
are the most widely used models for GPL'd software. The GPL is
 
central to their success, because it ensures that the code base
 
remains common, and that large and small companies are on equal
 
footing for access to the technology. Consider, for example, the GNU
 
Compiler Collection (GCC). Cygnus Solutions, a company started in the
 
early 1990s, was able to grow steadily simply by providing services
 
for GCC --- mostly consisting of new ports of GCC to different or new,
 
embedded targets. Eventually, Cygnus was so successful that
 
it was purchased by Red Hat where it remains a profitable division.
 

	
 
However, there are very small companies like CodeSourcery, as well as
 
other medium-sized companies like MontaVista and OpenTV that compete in
 
this space. Because the code-base is protect by GPL, it creates and
 
demands industry trust. Companies can cooperate on the software and
 
improve it for everyone. Meanwhile, companies who rely on GCC for their
 
work are happy to pay for improvements, and for ports to new target
 
platforms. Nearly all the changes fold back into the standard
 
versions, and those forks that exist remain freely available.
 

	
 
\medskip
 

	
 
\label{Proprietary Relicensing}
 

	
 
A final common business model that is perhaps the most controversial is
 
proprietary relicensing of a GPL'd code base. This is only an option for
 
software in which a particular entity is the sole copyright holder. As
 
discussed earlier in this tutorial, a copyright holder is permitted under
 
copyright law to license a software system under her copyright as many
 
different ways as she likes to as many different parties as she wishes.
 

	
 
Some companies, such as MySQL AB and TrollTech, use this to their
 
financial advantage with regard to a GPL'd code base. The standard
 
version is available from the company under the terms of the GPL\@.
0 comments (0 inline, 0 general)