|Understanding Open Source Software - by Red Hat's Mark Webbink, Esq.
Wednesday, December 31 2003 @ 09:20 AM EST
Here is a good article to share with your boss.
Mark Webbink, Senior Vice President and General Counsel of Red Hat, Inc., wrote this article for corporate attorneys, explaining free and open source software and comparing various open source licenses, detailing how the GPL really works, explaining US copyright law, and listing some corporate law office best practices for software, from the standpoint of what policies are prudent for the corporate environment.
He also explains how derivative works are defined, touches on the indemnification issue and the difference between open source and "shared source", and highlights some of the main myths and misconceptions about the GPL and open source.
I get email about this subject, so I know some of you are very interested in this subject, so I hope you enjoy finding answers in this thorough and accessible information.
This article was originally published in the March 2003 Journal of the New South Wales Society for Computers and the Law, and we republish with Mr. Webbink's kind permission.
What is Open Source Software?
The Open Source Initiative ("OSI") defines Open Source as software providing the following rights and obligations:
- No royalty or other fee imposed upon redistribution.
- Availability of the source code.
- Right to create modifications and derivative works.
- May require modified versions to be distributed as the original version plus patches.
- No discrimination against persons or groups.
- No discrimination against fields of endeavour.
- All rights granted must flow through to/with redistributed versions.
- The license applies to the program as a whole and each of its components.
- The license must not restrict other software, thus permitting the distribution of open source and closed source software together.
This definition clearly leaves room for a wide variety of licenses, and we will examine a number of those license types shortly. Although it is this OSI definition of Open Source to which the remainder of this paper relates, it is worthwhile to also examine the definition of Free Software, for often times the terms Free Software and Open Source are used interchangeably. While they are similar, there are differences worth appreciating.
When we speak of Free Software, we are not talking about freeware, i.e., software that is essentially in the public domain. Rather, we are talking about software that is licensed under the precepts of the Free Software Foundation ("FSF") and its flagship GNU General Public License.
According to the FSF definition:
"Free software is a matter of the users' freedom to run, copy, distribute, study, change and improve the software. More precisely, it refers to four kinds of freedom, for the users of the software:
- The freedom to run the program, for any purpose (freedom 0).
- The freedom to study how the program works, and adapt it to your needs (freedom 1). Access to the source code is a precondition for this.
- The freedom to redistribute copies so you can help your neighbour (freedom 2).
- The freedom to improve the program, and release your improvements to the public, so that the whole community benefits (freedom 3). Access to the source code is a precondition for this.
A program is free software if users have all of these freedoms."
Contrasting the Open Source and Free Software definitions, one finds that all Free Software is Open Source, but as administered by the Free Software Foundation, not all Open Source is Free Software. The difference principally arises from so-called license compatibility, but in large measure the differences are principally philosophical and not substantial.
Fundamentals of Copyright Law
To better appreciate Open Source software, we need a basic understanding of copyright law. Open source software is fundamentally grounded in copyright law . In order to appreciate the rights granted under Open Source licenses, one must first be familiar with the basic bundle of rights granted to the holder of a copyright. Under U.S. copyright law, those rights are:
- The exclusive right to copy the work;
- The exclusive right to make derivative works;
- The exclusive right to distribute the work;
- The exclusive right to perform the work; and
- The exclusive right to display the work.
These rights, in turn, are subject to certain limitations, such as rights of "fair use." Fair use includes the use of a work for purposes of criticism, comment, news reporting, teaching, scholarship or research and does not constitute infringement of the work. Whether a specific use is fair use is determined by a number of factors, including:
(1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
(2) the nature of the copyrighted work;
(3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
(4) the effect of the use upon the potential market for or value of the copyrighted work.
Works, such as software, may be placed in the public domain and exist outside of the scope of copyright law. However, with changes in the copyright law in the 1970's and 1980's, including the automatic application of copyright under the Berne Convention, it is no longer an easy task to contribute software to the public domain. Software (or any other body of work) that is in the public domain cannot, by definition, assert any restrictions on who or how it can be used, modified or distributed (though other laws, such as export controls, may still restrict some software's use or distribution). If Open Source software were in the public domain (that is, not subject to copyright because the author has disclaimed copyright in the work), any business or individual could use the software for any purpose without any copyright restriction, and there would be no requirements for legal review above and beyond ensuring compliance with other statutes (which apply equally to all other software, public domain, or not). Because Open Source software is not in the public domain, but instead protected by copyright law and licensed for use under certain, perhaps unconventional, terms, those terms must be understood.
A valid copyright license applies to a body of work and must assert at least one restriction. A copyright license that states no restrictions implicitly grants all rights, including rights to use, modify, distribute, etc. Most proprietary software copyright licenses assert restrictions on use (including definitions of "fair use", which, according to such licenses, usually does not include decompiling, reverse engineering, or other such uses), copying (usually only for the purposes of backup), and redistribution (usually only when acting as an authorized agent for the copyright owner).
Types of Open Source Licenses
Open source licenses may be broadly categorized into the following types: (1) those that apply no restrictions on the distribution of derivative works (we will call these Non-Protective Licenses because they do not protect the code from being used in non-Open Source applications); and (2) those that do apply such restrictions (we will call these Protective Licenses because they ensure that the code will always remain open/free).
To better appreciate the nature of these licenses, it is helpful to picture software licenses on a continuum based on the rights in copyright extended to the licensee. See Diagram 1 at the conclusion of this article.
Software that has been placed in the public domain is free of all restrictions, all rights under copyright having been granted to the public at large. Licensors of Non-Protective Open Source licenses retain their copyright, but they grant all rights under copyright to the licensee. Licensors of Protective Open Source licenses retain their copyright, grant all rights under copyright to the licensee, but apply at least one restriction, typically that the redistribution of the software, whether modified or unmodified, must be under the same license. Licensors of propriety licenses retain their copyright and only grant a few rights under copyright, typically only the rights to perform and display. The following table, where the BSD license is used as an example of a Non-Protective Open Source license and the GNU General Public License as an example of a Protective Open Source license, displays these contrasts - see Diagram 2 at the conclusion of this article.
Non-Protective Open Source licenses include: Academic Free License v.1.2; Apache Software License v.1.1; Artistic; Attribution Assurance license; BSD License; Eiffel Forum License; Intel Open Source License for CDSA/CSSM Implementation; MIT License; Open Group Test Suite License; Q Public License v.1.0; Sleepycat License; Sun Industry Standards Source License; University of Illinois/NCSA Open Source License; Vovida Software License v.1.0; W3C Software Notice and License; X.Net, Inc. License; zlib/libpng License; and Zope Public License v.2.0.
Protective Open Source licenses include: Apple Public Source License v.1.2; Artistic License; Common Public License v.1.0; GNU General Public License v.2.0; GNU Lesser General Public License v.2.1; IBM Public License v.1.0; Jabber Open Source License v.1.0; MITRE Collaborative Virtual Workspace License; Motosoto Open Source License v.0.9.1; Mozilla Public License v.1.0 and v.1.1; Nethack General Public License; Noika Open Source License v.1.0a; OCLC Research Public License v.1.0; Open Software License v.1.1; Python License; Python Software Foundation License v.2.1.1; Ricoh Source Code Public License v.1.0; and Sun Public License v.1.0.
All of these, and additional new licenses, can be found on the Open Source Initiative website.
Some Open Source licenses of both types include other provisions, such as restrictions on the use of trademarks, express grants of license with respect to applicable patents, disclaimers of warranties, indemnification of copyright holders in commercial distributions, and disclaimers of liability. However, none of these provisions are as fundamentally important as the obligations/restrictions that are imposed on redistribution rights under the Protective Open Source licenses, and it is with those restrictions on redistribution that we next focus.
The GNU General Public License
As of this writing, the GNU General Public License ("GPL") is the most pervasive license of Open Source software. Of all the software to which it has been applied, none is better known than the Linux® kernel. In fact, the GPL has been applied to a majority of those software modules that are included in the best known of the Linux® distributions, such as Red Hat® Linux®. Its wide appeal among the Open Source community stems from the fact that it falls into that category of Open Source licenses which obligate parties who wish to redistribute such software, either in original or modified (derivative) form, to do so under the terms of the license agreement under which such software was received (all of which we refer to as Protective licenses). That is, having been granted the right to use, modify and redistribute the software under the GPL, the GPL requires you to extend those same privileges under the same terms to others who receive the software from you. This is the common thread that governs Protective licenses, and for that reason, we will focus on the GPL as the standard for Protective licenses.
The GPL provides certain rights to anyone receiving a license to software governed by the GPL. At the same time, it imposes very few obligations except on those who wish to redistribute the software: Those rights and obligations are:
- The right to copy and redistribute so long as you include a copyright notice and a disclaimer of warranties. You may charge for the cost of distribution and you may offer warranty protection for a fee.
- The right to make derivative works for your own use.
- The right to distribute derivative works so long as you:
This section, and the obligation to license under the GPL, does not apply to works which are independent works distributed with the GPL'd work and which run on the GPL'd works.
- Identify the work as modified;
- License it under the GPL; and
- Provide the license information interactively if the program normally runs interactively.
- You may distribute the work only in executable form so long as the source code is:
- distributed with the object code;
- offered by a written offer, valid for a period of at least
three years, to make the source code available for no more than the cost of distribution; and
- for non-commercial distributions, accompanied with the offer the redistributing party received as to the availability of the source code.
- You may not impose restrictions on any of these rights.
This is a simple, yet elegant approach. Basically, the licensor is permitting any licensee to exercise virtually all of the rights available under copyright, i.e., the right to copy, the right to make derivative works, the right to distribute, the right to perform, the right to display. The only obligation imposed is, if the licensee, in turn, wishes to distribute the software to other parties, they agree to do so only under the GPL. The sole purpose of these restrictions is to preserve the integrity of the original grant of freedom through any path of redistribution and to make it impossible for anybody to create a version of the software that offers less freedom to any recipient than the original version would have granted. To paraphrase, the GPL states "once free, always free."
Note that the GPL has no relevance to the case where a party licenses the software and chooses not to redistribute it. This is true whether the party is an individual, a corporation, a corporate conglomerate, or the government. As noted by the FSF, when the GPL refers to "You" in the context of a corporation, it means the parent company and all of the controlled subsidiaries of that parent. Similarly, when "You" is addressing a unit of government, it means that unit of government and all of the subdivisions of that government that are under the direct control of that government. In that context, "You" can readily mean the entire federal government of the U.S. or it could mean any state or commonwealth government, including the agencies of that state or commonwealth government.
The GPL does not require that a licensee, who has not made a distribution of the software to another, provide copies of that software to any party who so demands it. The restrictions of the GPL apply only in the case of where GPL'd software is being provided to another party, and the GPL pertains only to the preservation of its original purpose-nothing more.
Based on the foregoing, we can divide the types of Open Source usage into categories, and analyze the legal implications of the GPL for each category. The interesting categories are:
- Users who use only GPL binaries as they would any other similar program;
- Users who modify GPL sources to handle local configuration issues or to address internal requirements and not for distribution to others; and
- Users who modify GPL sources and redistribute them for fun and/or profit.
In case (1), the GPL affects these users not at all; use of the Open Source GNU Emacs TM text editor does not imply that the act of saving a file changes the ownership of the file to the FSF, nor does compilation of a file by Open Source GNU C Compiler cause the resulting object code to belong to the FSF, nor does setting a breakpoint in an executable cause the executable to suddenly become the property of the FSF. Thus, the normal use of GPL software (i.e., use like one would use any other commercial software) in a commercial environment poses no extraordinary legal problems. The wide distribution of Linux operating system software in the last several years for use on commercial web and enterprise servers is ample evidence that there is no legal reason to not use Open Source software if you happen to think it is better than the proprietary alternatives.
In case (2), the locally modified software by definition confers to its users access to the locally modified sources. There is no requirement within the GPL that such local modifications be disclosed to any other party.
In case (3), we get to the group of users for whom the GPL was really written. Users redistributing modified or unmodified versions of Open Source software must obey the GPL's "Golden Rule" of licensing the distributed software under the GPL and not adding any downstream restrictions. To the extent that somebody wants to profit from GPL'd software by using traditional proprietary license restrictions, those restrictions will prove difficult if not impossible to apply. Note, however, earning profit because of the GPL is both legal and encouraged.
From this analysis we are left needing a definition of what constitutes a derivative work in software.
What is a Derivative Work?
The U.S. Copyright Act defines a derivative work as:
"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"."
Thus, a work that is based on one or more preexisting works constitutes a derivative work to the extent that the new material added constitutes an original work of authorship. Such new material may include editorial revisions, annotations, elaborations or other modifications. Derivative works may transform the original work, such as in a translation, including translating software from one computer language to another, or they may combine the original work with other works, such as in a compilation like Red Hat® Linux®. Copyright protection in a derivative work or compilation extends only to the material contributed by the author of such work, and does not grant rights in preexisting material included in the new work.
Where does the law stand on derivative works in software?
The law on derivative works in software is not well established. The U.S. Copyright Act does not specifically address derivative works in software, and there are no U.S. Supreme Court cases immediately on point. Most of the case law has developed among the various U.S. Courts of Appeals, but even there the law varies from one circuit to the next.
The Copyright Act provides an important definition in addition to that of "derivative works", that of "computer programs", which are defined as:
"a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result."
In addition, the Copyright Act limits the scope of what is covered by copyright by excluding certain subject matter. §102(b) of the Act provides:
"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."
Perhaps the most established of the tests for derivative works in software is the "abstraction, filtration, and comparison" ("AFC") test established by the Second Circuit. Under the threepart AFC test, a court first determines (abstracts) the constituent structural parts of the original program. From these structural parts, the court then filters all unprotectable portions, including those unprotectable matters defined in §102(b) of the Copyright Act and elements that are in the public domain. In the final step, the Court compares any remaining code containing creative expression to the structure of the second program to determine whether the software program in question is sufficiently similar to the pre-existing work to justify a finding that the second program is a derivative work of the first. This AFC approach has been adopted by three other circuits: the Fifth, Tenth and Eleventh.
Of the remaining nine U.S. Courts of Appeal, only one has adopted a clear test for derivative works in software. The Ninth Circuit's test is based on analytical dissection, which first considers whether there are substantial similarities in both the ideas and expressions of the two works and then utilizes analytic dissection to determine whether any similar features are protected by copyright. The similar elements are categorized by the degree of protection they are to be afforded. "Thin" protection is afforded to non-copyrightable facts or ideas that derive copyright protection only from the manner in which those facts or ideas are aligned and presented. "Broad" protection is afforded to copyrightable expression. The court uses these standards to make a subjective comparison of the works to determine whether, as a whole, they are sufficiently similar to justify a finding that one is a derivative work of the other.
How do these tests apply to derivative works in Open Source software?
In addressing derivative works, Open Source software requires special consideration. This is due principally to the fact that Open Source software, by definition, permits the making of derivative works. Under a Non-Protective license, the new portions of such a derivative work may be licensed under the license of choice of the author, and there is little likelihood of an infringement dispute.
The case is much different with a Protective license because it requires derivative works to be licensed under the same license as the original work. Here the question largely becomes one degree of copying versus adequate avoidance of derivation. Where Open Source software licensed under a Protective license appears to have been copied, in whole or in part, into a larger work, which is then licensed under a different license than the original work, the question of derivative work and infringement would be determined by the courts using the tests outlined above. However, this is not the case where the subsequent author maintains the original Protective license with respect to the original work but licenses the new work under a different license, for here the subsequent author has not infringed the rights of the original author except to the extent that the new work can be determined to be a derivative work of the original. This latter instance requires an entirely different approach to determining derivation.
Where the original work continues to be licensed under a Protective license and the new work is licensed under an alternative license, the following factors are to be considered when determining whether the new work is a derivative of the original:
- The substantiality of the new work;
- Whether any part of the original work has been modified; and
- How such modification has been accomplished.
This analysis is consistent with the distinction drawn by the GPL itself. Clause 2 of the GPL states:
"Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather the intent is to exercise the right to control the distribution of derivative or collective works based on the Program. In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of storage or distribution medium does not bring the other work under the scope of this license."
For example, if the work in question is a database written entirely by you, and the Program in question is a GPL'd operating system (one of many to which the database may have been ported), the distribution of the database with the operating system on a volume of storage (such as the system hard disk) would not confer the GPL of the operating system to the database software. On the other hand, if modifications are made to the Program (the operating system) in order to accommodate the work (the database), then those modifications, which are a derivative work of the Program, would need to be made available under the GPL. No modifications to the work (the database) need be redistributed in this case.
In summary, the legal requirements of the GPL are quite straightforward for commercial software providers: if you want to use a proprietary revenue capture model, keep your works (i.e., the code) separate from GPL'd works, keep the modifications made to each fully independent, and there will be no problems protecting your primary works. At the same time, any modifications you make to software that are already covered by the GPL will be subject to the GPL.
Myths About Open Source
Before leaving this discussion of Open Source licensing it is worthwhile to address some of the myths or misconceptions that have arisen around Open Source.
Myth 1 - Open Source software is "viral" and undermines intellectual property rights.
This myth is particularly rich. First, as already noted, Open Source software is fundamentally grounded in copyright law. As with the holder of any copyright, the copyright holder for a piece of Open Source software gets to elect which rights he/she will grant to others. Open Source authors simply choose to grant more rights than proprietary vendors. The mere fact that an Open Source author using a Protective license insists that derivative works that are distributed to others be licensed under the same license should be contrasted with proprietary software licenses that simply deny the licensee the right to create derivative works or to redistribute them. Each is an exercise in intellectual property rights, and neither is wrong.
Myth 2 - Open Source software is more prone to claims of intellectual property infringement.
The suggestion of the proprietary vendor is that, because the Open Source development model relies on a vast network of Open Source developers who are not necessarily under the control of the distributor, the code produced is far more likely to be exposed to intellectual property infringement claims. The facts simply do not bear this out. While there undeniably have been such claims against some Open Source development projects and/or distributors, the claims have been few and far between.
Myth 3 - Unlike proprietary vendors, Open Source software vendors do not provide warranties or indemnity against intellectual property infringement.
That is true, but no more true for Open Source vendors than for proprietary vendors. For example, the Windows 98 license expressly disclaims any warranty of non-infringement.
Myth 4 - The GNU General Public License is risky because it has never been tested in court.
True again. But which is riskier, licensing practices that are constantly being challenged or those that, in their simplicity and effectiveness, have avoided challenge.
Myth 5 - Making your source code viewable to some users is the equivalent of Open Source.
Open Source provides value to its customers and users by giving them total control over their computing environments. The customer gets to choose whether to run the standard version or whether modifications are desirable. The customer can not only see the bugs, he/she can fix the bugs. Making source code merely viewable to a few users does not help them understand the code, does not let them modify the code, and most importantly, does not let them fix the code when it breaks. This approach to source code "sharing" equates to entering a public library only to find there is no card catalogue and all of the books are in locked glass cases. Yes, you can root around and find the titles of the books, but you have no ability to gain knowledge from them. Proprietary software seeks to maximize its value solely in monetary terms by achieving a monopoly. Open Source software maximizes its value by assuring that a monopoly cannot be achieved.
Myth 6 - Open Source methods do not produce innovation.
This is a myth. The Open Source community: (a) developed the Apache webserver which is used to run the majority of webservers in the world today; (b) developed Sendmail, the most popular e-mail management software; and (c) developed BIND, the basis for using domain names instead of IP addresses to locate websites. Clearly, Open Source is capable of advancing the art of software.
Without belaboring this point, let us turn to best practices that a corporate law office should maintain with respect to software, whether Open Source or proprietary.
Corporate Law Office Best Practices for Software
As with any form of intellectual property, there are risks associated with licensing the use of software. Some of those risks may relate specifically to Open Source software, but most often they relate to all software, regardless of the form of license. Following are a series of best practices that every corporate law office should implement across their company:
1. Do not permit the uncontrolled importation of software onto company computers.
Do not permit employees to download freeware, shareware, or Open Source software onto company computers without first clearing the license terms with the legal department. At the same time, bar the use of proprietary software except to the extent that the company can account for the permitted licenses. In other words, know what you are putting on your machines--to do otherwise exposes your company to risk.
2. Deal with reputable software vendors with financial staying power.
One of the biggest risks a company takes is adopting software that has no future. Equally true is licensing software from a company without the financial wherewithal to maintain and protect that software. Know your vendors. Know their financial strength, know their policies on licensing, know their responsiveness, and know that their software is reliable.
3. Know how the software will be used.
It's one thing if Open Source is to be used as an operating system on a backoffice server, it is something altogether different if that same Open Source software is to be modified and embedded in a product. The former is not problematic; the latter may be. At the same time, make sure your IT folks are well aware of the typical proprietary restrictions which prohibit reverse engineering or modification. While some proprietary vendors may permit such activities under a special development license or a community source code license, they do not generally permit the activities under their general commercial licenses. It may be worthwhile to categorize each item of software and its permitted uses, e.g., approved for general use in executable form only, approved for use at the source code level in specialized applications or modified applications, and not approved for any use. Finally, nature of use is important in knowing whether the software will be distributed outside the company, potentially triggering Open Source licensing restrictions.
4. Have a means for documenting what software, and what version of that software, is in use.
Knowing this information and having ready access to it will help assure licensing compliance and at the same time permit IT managers the ability to manage the IT architecture and its advancement.
5. Require documentation of all internal software development projects.
This includes modification of Open Source software. Such documentation should indicate the source of any base software that is modified, all of the authors of the developed software, prior projects (both internal and with prior employers) on which such developers worked, and the identification of any known related intellectual property, particularly patents.
These are but a few suggestions. They are meant to address those issues most commonly found in software, including Open Source software.
For those interested in learning more about Open Source, the following websites are suggested reading:
- When I talk about copyright law in this paper, I am discussing U.S. copyright law as embodied in Title 17 of the United States Code. The United States is a signatory to the Berne Convention covering copyright, and much of U.S. copyright law is very similar to that of other Berne signatory countries. However, there are provisions in copyright law in the U.S. that are unique to the U.S., such as copyright registration. Persons in countries other than the U.S. should consult local legal counsel specializing in copyright law.
- §1-106, Title 17, U.S. Code.
- §1-107, Title 17, U.S. Code.
- 37 CFR 201.26 defines public domain computer software as software which has been publicly distributed with an explicit disclaimer of copyright protection by the copyright owner. As the Free Software Foundation has stated, public domain software means software that is not copyrighted.
- Under the Judicial Improvements Act of 1990, which authorized the creation of a national shareware registry, software copyright owners may donate their software to the public domain by assigning it to the Machine-Readable Collections Reading Room of the Library of Congress. 37 Code of Federal Regulations Part 201.26 (1991).
- 17 U.S. Code §101.
- 17 U.S. Code §103.
- For an in depth discussion of the state of the law, see "Software Derivative Work: A Circuit Dependent Determination", Dan Ravicher, October 31, 2002, http://www.pbwt.com/Attorney/files/ravicher 1.pdf.
- 17 U.S. Code §101.
- Computer Associates Intl. VC. 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); Mitek Holdings, Inc. v. Arce Engineering Co., Inc., 89 F.3d 1548.
- Apple Computer, Inc. v. Microsoft Corp., 35 F.3d 1435 (9th Cir. 1994).
|Authored by: freeio on Wednesday, December 31 2003 @ 06:49 PM EST|
|This is a most useful thing. I would suppose that those who spend their time in|
the depths of legalese for a living do actually appreciate a well-written
explanation of the relationships between the various parts which make up any
complex concept. And make no mistake, software copyright and licensing are
The concepts and rationale behind the free software licenses are not hard to
grasp. but they do run sufficiently counter to the prevailing monetization of
everything that they can be difficult to believe. Why would someone do that?
Well, this explanation will help with the rationale and the mechanics, for those
who have not been in this territory before.
(I am reminded of my first time living in South Texas, where my employer (and
soon best friend - it does actually happen!) had to explain to me that the local
response to offers of hospitality was expected to be yes, of course I will
accept it. To turn down an offer of hospitality was considered an insult (i.e.
I'd rather die in this wilderness than accept anything from the likes of you!).
This turned my world a bit off-axis, as I had been raised in the
Minnesota/Nebraska version where much is offered, but you must always turn it
down, because that is the expected behavior. Two world-views, two cultures,
many worlds apart...)
73 de w4ti
[ Reply to This | # ]
|Authored by: tazer on Wednesday, December 31 2003 @ 07:06 PM EST|
|Already sent the link to my boss, but I'm preaching to the choir there. We're|
done migrating our servers to GNU/Linux and are working on workstations this
SCO -->Santa Cruz Operation ->Tarantella
SCOG ->The SCO Group -------->Caldera
[ Reply to This | # ]
|Authored by: jrw on Wednesday, December 31 2003 @ 07:26 PM EST|
|Happy New Year Everyone.|
And may this bright new year bring all of us, both collectively and
individually, all the things we both desire and deserve.
Good Luck and God Speed.
[ Reply to This | # ]
|Authored by: Anonymous on Wednesday, December 31 2003 @ 07:41 PM EST|
|What got changed with this page? It runs off my screen to the |
right. Most (all previous) Groklaw pages size to my browser.
This (dare I say it) , seems made to fit a WINDOWS screen.
Makes it hard to read. Although I will try. Afterall , Groklaw
is an interesting place. I have learned alot here.
A retired CPA, who has come to appreciate what a GOOD lawyer
[ Reply to This | # ]
|Authored by: rsmith on Wednesday, December 31 2003 @ 08:37 PM EST|
The following "best practice" misses the point:
2. Deal with
reputable software vendors with financial staying power.
the key benefits of Free and/or Open Source Software is that you're not
tied to a particular vendor. Has the company doing your software maintenance
folded? No problem, select another to take over.
A free/open source
program doesn't die as long as there is a community keeping it
Never ascribe to malice that which is adequately explained
[ Reply to This | # ]
|Authored by: Ted Powell on Wednesday, December 31 2003 @ 09:06 PM EST|
|Simplification can sometimes result in unfortunate changes to perceived
The Open Source Initiative ("OSI") defines Open Source
as software providing the following rights and
This appears to say that an
Open Source licence "May require modified versions to be distributed",
specifically in the form of "the original version plus patches."
4. May require modified versions to be
distributed as the original version plus
writer had a mental "if any" in there somewhere, but it didn't make it into the
document. As the item stands, it confirms some people's worst fears. The rest of
us generally won't interpret it that way, of course, but we aren't the ones it
was written for.
A quick check at http://www.opensource.org/d
ocs/definition.php reveals that the intended meaning is more
4. May limit distribution of modified versions to the
original version plus patches.
Even that doesn't tell the full
story of criterion (4), but at least it's less of a breeding ground for
BTW, don't try to find the OSI site
www.osi.org; it won't work.
not determined by majority vote.
[ Reply to This | # ]
|Authored by: blacklight on Wednesday, December 31 2003 @ 11:39 PM EST|
|groklaw's publication of Mark Webbink's article underlines groklaw's unique|
value-add as a communications medium for critical information to get to OSS.
Given the comprehensiveness and length of Mark Webbink's contribution, I just
don't see any of the rag trade press printing it in its entirety on their own
website let alone allowing a flow of comments on it.
And speaking of comments, I will start by nitpicking Mark Webbink's statement
that the GPL offers virtually all the privileges allowed by the copyrights: the
GPL is intended by the copyrights owners (or their proxies) to give expanded
privileges beyond the default privileges authorized by copyrights law, provided
that certain conditions are met.
[ Reply to This | # ]
|Authored by: bobn on Wednesday, December 31 2003 @ 11:44 PM EST|
|The definition of derivative works given here implies to me (IANAL) that SCOX is|
in for a very rough ride on trying to claim NUMA, RCU, etc.
Thanks PJ. Happy New Year.
[ Reply to This | # ]
|Authored by: belzecue on Wednesday, December 31 2003 @ 11:56 PM EST|
|This idea has been touched on before, but it bears repeating:|
By throwing everything and the kitchen sink at Linux in an attempt to stop it,
SCO have helped to show, in a very public way, that Linux is unstoppable.
Almost. The SCO controversy has demonstrated to those who would prefer to see
Linux put on a short leash (M$) that patent violation is probably the only thing
capable of halting Linux in its tracks.
But even if Microsoft chose to fire their big patent guns at Linux, would they?
Have they missed their chance to steer corporate customers around the Linux pot
of gold? I say yes, they have. CTOs and CIOs would not be happy if M$ forced
them to rip out their existing Linux servers because of patent issues.
"Take our road or you ain't going nowhere" is the SCO line, and
that's not winning any hearts and minds in the backoffice trenches. Besides,
M$ needs the illusion of competition and the illusion of free choice. They
wanted the illusion but they ended up facing the real thing.
They tried to lock in users even more, but (apologies here) the more they
tightened their grip, the more star systems slipped through their fingers.
They tried heavy discounting, and to their surprise customers cosied up to Linux
anyway. As a Microsoft bean counter, that's got to screw with your head
So what to do? Microsoft will 'innovate'. Okay, so their meaning is not the
one you and I take for granted. Innovation, in the M$ world, means consume and
control, buy and bury, catch up and cry 'bleeding edge'. So, to fight against
Linux, Microsoft will 'innovate'.
M$ Problem 1: anything you can do, Linux can do better... You have finite
resources; Linux has global resources. Cry copycat if you like, but programmers
love playing with cool ideas. Linux coders will take your ideas (which cannot
be copyrighted), apply that global brainpower, and build something wonderful --
from the ground up (why would they want to imitate Microsofts flawed and
insecure methods and 'ways of doing things'?). If you want to stay in the
game you will need to copy Linux's improved implementation of the idea back
into Windows (don't forget the press release about your latest
M$ Problem 2: Open is a dirty word. Open goes against every twisted fiber in
Microsoft's body. Open means less lock-in. Open means ugliness exposed. Open
means questions asked. Linux means openness, and once customers have got a
taste of that, they never want to go back. Microsoft becoming 'open' is the
same as the magician telling the audience exactly how the smoke and mirrors
fooled them. With a collective sigh and a vague sense of being cheated, the
audience trudge off in search of something genuinely exciting, something that
holds their amazement precisely *because* they can see how it works without the
cloak of deception and obfuscation.
M$ Problem 3: This whole Linux thing would go away if only we (MS) could
befriend it then smother it. Oh, that's right... the GPL might actually hold
up in court, which means we can't swallow Linux whole. Tell that Heise guy
we're cancelling his BMW order.
Hmm, so back to innovation. Customers will flock to Windows for the innovation.
Customers will not flock to Windows if Linux offers the same innovations.
Ergo, Linux must not be allowed to adopt 'our' (M$) innovations. Thus, we
must patent each innovation so that the threat of a lawsuit stifles innovation
in Linux. We can't kill Linux outright, but we can squeeze and squeeze and
squeeze, and slowly but surely we will choke the life out of that sucker. Hey,
it worked all those other times, right?
To recap: M$ cannot kill Linux or even try. The IT industry will not tolerate
that. M$ needs to get Windows ahead in the 'must have' feature race so that
Linux becomes 'totally five minutes ago, dude'. To get ahead, M$ must cripple
Linux with threats of patent litigation. I just hope, for their sake, it
doesn't backfire like their other attempts to kill Linux. Would IBM, in the
spirit of mutually assured destruction, throw a few patent nukes back at M$?
[ Reply to This | # ]
|Authored by: cr on Thursday, January 01 2004 @ 12:21 AM EST|
Freeware, at least in the corner of the WinDOS world occupied by
shareware, typically does NOT mean public-domain. In comparison with
shareware, where a payment is either expected or functionally required (by
feature-crippling or timebomb) for continued use of the binary program, freeware
is software which the author allows you to freely distribute and use in its
released binary form only. Sometimes the freeware version is a "Lite" version of
their shareware offering (the NoteTab editor has such a freeware junior
edition), but the source-code for the program is always withheld.
high-quality example of this kind of program distribution is David Harris's
POP/SMTP/IMAP client program Pegasus Mail, available at http://www.pmail.com/.
I usually recommend
Pegasus to people who are still stuck using Windows. The program has been
well-maintained since the Win3.1 days. To my knowledge, the program has had one
major hole, quickly fixed.
Other freeware programs have had less
dependable histories; Netlab, for instance, a handy Windows bundle of
networking utilities with a simple GUI by Alexander Danileiko, is now
all-but-unavailable on the Internet. Depending on freeware is a bit of a gamble,
since the author might decide at any time to:
...and, without the source code, the end-user is unable
to work around these limitations and vulnerabilities.
- take the program off the
Net, making it unavailable to new users
- go out of business or lose
interest, orphaning the program
- turn the program into shareware (NoteWorthy Composer, a
Windows-based music composition editor and MIDI sequencer, went this route) for
any updates or bugfixes
- insert treacherous code into the program (I
recall reading about an executable-compressor that had backdoor code in it)
- decide not to update the program to deal with changed circumstances
(Pegasus Mail is unlikely to be ported to Linux, per author's
[ Reply to This | # ]
|Authored by: DaveAtFraud on Thursday, January 01 2004 @ 01:30 AM EST|
|The article nicely sidesteps the difference between the GPL and the LGPL when|
discussing derivative works. There wouldn't be two different forms of the GPL
if there wasn't a difference in how the GPL authors wished for the two forms to
My understanding: libraries licensed under the LGPL can be called by a
proprietary program with no requirement that the proprietary program itself be
GPLed. Librariies licensed under the GPL are "viral" and their use
by a proprietary program requires that the proprietary program itself be GPLed.
It remains to be seen whether the GPL will hold up as applied to the GPL
definition of derivative works for this case in particular.
Side note: I somehow doubt if nuances such as dynamic vs. static linking of
software libraries will ever get encoded into copyright law even as a result of
some particular case law. Any explanation of GPL vs. LGPL applicability should
not not delve into technical nuances.
BTW, Happy New Year everyone!
Quietly implementing RFC 1925 wherever I go.
[ Reply to This | # ]
|Authored by: Anonymous on Thursday, January 01 2004 @ 04:46 AM EST|
|"Derivative works may transform the original work, such as in a|
translation, including translating software from one computer language to
another, or they may combine the original work with other works, such as in a
compilation like Red Hat® Linux®."
While otherwise a very useful summary of a highly confusing field, I must take
strong exception to this statement, as it is both inaccurate and dangerously
A compilation of other works is not, as I understand it, a derivative work: It
is a copy of those works included, no more and no less. Outside of software, a
fiction anthology or a nonfiction collection of articles will contain separate
copyright statements for each item, with a phrasing like 'reproduced with
permission of ...' for each item. Furthermore, the GPL -specifically- mentions
that a simple collection of programs in the same media is not considered - by
the FSF and other users of the GPL - to be a derivative work. Similarly, Redhat
is -not- a derivative work of all the programs included in it.
Each RPM file is a reproduction of the program it contains. The Redhat Linux
-Kernel- is a derivative work, because Redhat modifies the kernel before
release, and other modified programs are, individually, derivative works, but
the collection of software that makes up a RedHat Linux distribution is not, in
these terms, a derivative work.
In fact, if it -were- considered a single derivative work, it almost certainly
could not be distributed legally by anybody, because the GPL is constantly
incompatible with some license or other (or vice versa, if you prefer), most
collections of open-source software contain works which cannot be merged into a
single 'work' and meet the criteria of both licenses. The FSF is constantly
advocating to get people to toe the line with GPL compatability, but there are
many exceedingly useful packages that don't have that.
(qmail and xv come immediately to mind - xv, of course, is not open source even
though it's source distributed, and qmail is just barely OSI-definition
open-source and very much not FSF-defined free software. I don't know if Redhat
actually contains either of these packages - but I thought Debian was the only
distribution picky enough to accept only 'truly Free' software and be even
close to able to consider itself entirely 'GPLable', to coin a term.)
I would be more inclined to choose an example like 'Galeon is a derivative work
of Mozilla,' personally.
Of course, I am not a lawyer and not a part of RedHat, so perhaps I'm missing
something, but I have been around free software and discussing the legal
ramifications of the GPL since before the term 'open source' was coined, so I
hope I have some clue what I'm talking about.
[ Reply to This | # ]
|Authored by: Anonymous on Thursday, January 01 2004 @ 11:32 AM EST|
| Imagine that for a GPL software, I create a function that return any nth prime|
number in a constant time.
Now I need a function like this in my company's software (a trade secret one,
with commercial distributions). The easy way is to copy the function code.
It seems obvious that somewhere in source and particularly in running software,
I specify that this GPL part is used.
But at this point, and after all that I've read, it seems that the whole
company's software must become a GPL software, except if I had place, at first
step, my function into a LGPL library instead of a GPL software ... Is that true
If I modified the function to return an array of prime number, it appears to me
like a derivative work, that MUST be given back to the open source community,
but only the function, not the whole software source code.
So, what's can be really considered as a derivative work ? Must I release all
my function as library under LGPL ?
In this example it's my own GPL function code, but in fact it could be any
function or part of a GPL software.
The §102(b) of the Copyright Act and the "AFC" test may applied to
this case and any similar one. But it does not look clear to me ... (perhaps
it's because I'm french and my english is not perfect :)
[ Reply to This | # ]
|Authored by: hooper on Thursday, January 01 2004 @ 05:11 PM EST|
|Just my opinion.|
I don't care to listen to a word RedHat has to say about open source. I see
RedHat as a traitor to the open source movement with the words they write and
language they use. More than a few feel as I do.
Using a name for what they call their new open source project that they had no
right to use. Telling the world that they would be better off using Windows.
That is RedHat.
I put little if any faith in any corporate Linux. I prefer noncorporate
sponsored open source works for both business and personal use.
I know many feel that Redhat is a leading factor in the fight against SCO, but I
don't see it that way at all. Nor do I see Novell/Suse as a good trend. Take it
for what you will, Novell is the king of proprietary networking and now being
they do not have a product worth anything, they purchase suse Linux. Greed is
everywhere. Not just SCO.
If your looking for a hero, look for those who contribute without backstabbing
the community when their model of a desktop or anything else doesn't pan out.
Sorry for the negativity, but I enjoy the open source movement and
participation. Not corporate greed and those who try to play the fence until
[ Reply to This | # ]
|Authored by: Anonymous on Thursday, January 01 2004 @ 05:56 PM EST|
|I am trying to reconcile the following two statments:|
The GPL provides certain rights to anyone
receiving a license to software governed by the GPL. At the same time, it
imposes very few obligations except on those who wish to redistribute the
software: Those rights and obligations are:
1. The right to copy and
redistribute so long as you include a copyright notice and a disclaimer of
warranties. You may charge for the cost of distribution and you may offer
warranty protection for a fee.
From Red Hat's Web
Red Hat Enterprise Linux is the premier operating system for
open source computing. It's sold by annual subscription, runs
on seven system architectures, is certified by top enterprise software and
hardware vendors, and backed by a Red Hat Network subscription and up to 24x7
support with one-hour response.
If you look a little further, you
will find that per unit prices for intel based systems range from $179
(basic workstation) to $2595 (premium AS).
Seems to me that
$179 more than covers distribution cost, also, warranty protection is not
being offered, it is being required.
If this is allowed under the GPL,
then great, but I don't expect Redhat to get very far with these prices on the
lower end systems. On larger systems, maybe - I don't have any recent prices
[ Reply to This | # ]
|Authored by: Anonymous on Friday, January 02 2004 @ 01:53 AM EST|
|For crying out loud, GPL is not an Open Source licence, it is a FREE SOURCE|
LICENCE. Learn the difference, otherwise don't mix GPL into your 'article'!
[ Reply to This | # ]
|Authored by: Anonymous on Friday, January 02 2004 @ 05:44 AM EST|
...but the article points out that a user of
is allowed to use it for whatever purposes they may choose
which includes running proprietary software on it; the
apply to someone who redistributes. So,
in the case of a dynamically linked
library the author of
a proprietary work is not redistributing, the user must
obtain the library from another source, in which case I
don't see how
copying code into RAM from disk is relevent.
[ Reply to This | # ]
|Authored by: Anonymous on Sunday, January 04 2004 @ 03:12 PM EST|
|If you look at the whole issue of license and copyright
rationally and make an
analogy to programming you get
What are the odds
that a person without a computer science
background could write code without
every executing it that
would contain no bugs and no security holes?
of the question is:
What are the odds that a person without a legal
could write a license and copyright without ever testing it
court that would contain no errors and would survive
attack by lawyers?
know the answer. So I believe it is pointless to
talk about license and
copyright issues. Since open source
programmers can't afford lawyers I believe
that every open
source license and copyright will fail as soon as it gets
court. GPL won't because it was written by a lawyer
trained in IP law. But
programs that have been GPLed can
still lose because they misappropriate
Programmers, being hyperrationals, believe that they can
legal issues and craft licenses (or copy
existing licenses) that will survive in
court. But the
"english" words in a license have special meanings which
correspond to the dictionary definitions. Just like
"registers" doesn't mean
"heating elements embedded in a
floor or wall". It takes years to understand the
meanings. Worse yet, every court (machine) has its own
which might not be agreed to by another
or higher court.
Even Darl McBride's
brother, who IS a lawyer, doesn't seem
to understand IP law well enough to argue
it in court.
Professional lawyers get slaughered in court. We
As the lead programmer of a free software project I've
to endless flame-wars by programmers who
think they understand IP law. While I
applaud the effort
to educate evidenced by this article I think it will
lead to yet more flame wars. Write code. Give it away.
Never ask what
becomes of it. That way lies madness.
That said, I must confess to reading
Groklaw 3 times a day.
I only wish PJ could program as we could use her
on our project.
[ Reply to This | # ]
|Authored by: Anonymous on Sunday, January 04 2004 @ 04:03 PM EST|
|<quote>Now you may think the GPL (and/or other GNU licensed works like the|
LGPL, FDL, etc...) protects your work. What you may not realize is that
Copyright Law is the _ultimate_ law. The software that cracked the encryption
and revealed the follies of popular "Internet Filtering" software
were perfect examples. The software was released GPL, but then revoked later.
How? Because the creators ultimately have "all rights reserved" to
their copyright, and can revoke any license at any time (like they did when the
popular filtering software vendors bought the rights).
Now one way you can "protect" your free software/works from being
submissive to this "hole" in Copyright Law is to assign all rights
to the Free Software Foundation. In fact, this is exactly what the FSF
recommends you do with any GNU licensed work. If you have not done so already,
consider doing this with any GPL, LGPL, FDL or other GNU-licensed work that you
do not plan to "dual-license"</quote>
Is this correct?
[ Reply to This | # ]
|Authored by: Anonymous on Sunday, January 04 2004 @ 05:24 PM EST|
|I understand that if I develop software using GPL tools, my customer has the|
right to the source code. My question is, what if my customer has a concern
about me distributing the source code, that I developed for them, to someone
In other words, is there any way that I as the developer, using GPL software,
can design a contract that would allow my customer to do what they want with the
source code, but limit my rights to distribute?
Unless my customers can be assured that there specific business processes are
protected, by more than my honesty, I don't see how I can sell my programming
services, using GPL tools.
[ Reply to This | # ]
|Authored by: Anonymous on Sunday, January 04 2004 @ 06:10 PM EST|
|And yet, the reasons to how GPL would actually be better than BSD/MIT license is|
simply not there. It only states that GPL is a good license.
Yes the viral effect is there, not worse than proprietary it states, but very
much worse than Non protective licenses which is actually the same as the word
FREE. The word restrict is not same as free.....
Worst people are know to be those who say "It's free, but..." or
"I'm not rasist, but..." or "I don't hate you, but..."
and so forth... same is it with the GPL, here and it's the lesson.
Please consider license choice and consider freedom as a viable aspect of it,
that would mean BSD/MIT
[ Reply to This | # ]
|Authored by: drobson on Sunday, January 04 2004 @ 07:28 PM EST|
|This article is very helpful. I have a couple of related questions and hope|
someone can point me to information about dual licensing. Specifically...
Several vendors seem to offer two versions of their software: a GPL version and
a proprietary-licensed version which is available for a fee; prohibits revision
and redistribution; and includes support.
 Is this legitimate? Is it possible for the author to distribute the same
code by two seemingly conflicting licenses at the same time?
 I am writing a program and would like to distribute version 1.0 via GPL and
simultaneously sell 2.0 (which contains much 1.0 code) under a
commercial-restricted license. Later, when 3.0 is ready, I may elect to make 2.0
GPL and 3.0 commercial-restricted. I understand that it is not possible for a
secondary user to do this. Is it possible for the author to do it?
[ Reply to This | # ]
|Authored by: Anonymous on Sunday, January 04 2004 @ 09:01 PM EST|
|One thing that I have never gotten a simple answer on is this one:|
1 - Assume X is a GPL application that supports plugin modules.
2 - I write mod_mine for X.
Now, must I license the module as GPL/LGPL? If so, that's the only way the GPL
to me looks viral. Otherwise, it can be a platform on which I can build other
things. And if I find something in the core code that's wrong, I can still
contribute. Of course, there'd be labeling in the module--header files etc.
But beyond, I _should_ be able to just write my module.
[ Reply to This | # ]
|Authored by: Anonymous on Tuesday, January 06 2004 @ 09:49 AM EST|
Who owns the copyright to a derivative work?
This may be a problem unique
to open source software, but it's come up in my own work, and this seems a
compelling place to address it. I'll describe a generalization of my own work
to illustrate the problem.
I downloaded an open source, GPL'd piece of
software from sourceforge, consisting of (we'll say) 10 packages, where each
package is a separate file. Two of the 10 packages remain unaltered, 2 have
been nominally altered, the remaining six have been gutted, and there are now 10
additional modules or packages. By gutted, I mean that perhaps 20% of the
original code is recognizably intact, the rest has been altered to the point of
unrecognizability or deleted entirely, and significant additions have been
This raises a number of questions. First, does "copyright on the
software" apply to all the packages as a whole, or to each individual file?
Assuming that it in fact applies to the packages or individual files, it's easy
to see that unlatltered packages are copyrighted to the original author, and
entirely new packages are copyright to me. What of the remaining packages? If
I make significant changes, the work or package is now "derivative," but to whom
should the copyright be attributed? Does the original author still hold
copyright? Do I, having modified the work? Do we share copyright?
be important to note that my goal is to release a new, derivative work. That
is, my intentions exceed the goals of the original software and its author. In
short, I'm forking development.
I also plan to release the resulting
software under GPL, so from another developer's standpoint, the question may be
moot. That is, I will extend the same rights extended to me, and thus the code
will be "out there," irrespective of whose name is on it. Nonetheless, I'm
having difficulty deciding how to appropriately indicate copyright
In the spirit of open source development, I plan on including
everyone's name as an AUTHOR of the software, but there's a significant
difference between including names in an AUTHOR file, and claiming
Any thoughts anyone has on this would be most welcome, not least
being where else I might pose this question.
[ Reply to This | # ]