|
Open Source Software: What Is It and How Does It Work?" - By Dr. Ben Kremer |
|
Sunday, April 04 2004 @ 09:04 PM EDT
|
Solicitor Dr. Ben Kremer and his firm, Freehills in Australia, have graciously given Groklaw permission to reproduce his recent article, "Open Source Software: What Is It and How Does It Work?" originally published in their newsletter and available on the Internet here. I asked for permission because I find it to be one of the clearest, most understandable articles I've come across yet on the GPL, what it is legally (contract or license), and how it works, and I thought it was important to present it to as wide an audience as possible and to have it in Groklaw's searchable collection. Thank you, Dr. Kremer and Freehills.
************************************************************
Open Source Software: What Is It and How Does It Work?
~ by Dr. Ben Kremer
Open source software is playing an increasingly important role in the marketplace. It is, however, still subject to misunderstanding at both the lay and legal level. This article will attempt to clarify some of the uncertainty by addressing two fundamental issues—what open source software is, and how it works.
What is open source software?
The term 'software' refers to two different, but related, things. One is 'source code', which is a set of human-readable and understandable instructions that comprise the 'recipe' from which an executable program can be made. The other is 'object code', which is the actual executable program—the bundle of 1s and 0s which are fed to a computer's microprocessor to make that computer perform various operations.
For various reasons, software is invariably written as source code. That code cannot itself be executed by a computer, but is 'compiled' to form object code by a specialised program. 1 The process is hard to reverse (a technique called 'decompilation'), as the process of compilation generally strips out things such as comments inserted by the programmers to describe how the program works, and names of variables or subroutines. (This information generally cannot be restored during decompilation, so decompiled source code is often difficult to understand and modify.) Commonly, various functions can be incorporated in source code so that different versions of object code—such as 'demo' programs; program versions with greater or restricted functionality; or programs optimised for different computers or operating systems—can be produced with minimal effort.
The 'closed source' model
Historically, software vendors kept source code to themselves (indeed, invariably protecting it as valuable trade secrets) and sold, or licensed, only the object code versions of programs to their customers. Customers can run those pieces of object code, but cannot view or modify the source code (and hence modify the behaviour of the program) without dealing with the software vendor. Licence terms commonly prohibit customers from decompiling the object code. As a result, they can only use whatever is provided to them, along with any modifications they could prevail upon the software supplier to incorporate. This model is called 'closed source' software, and continues to be used by the majority of commercial software companies.
The 'open source' model
'Open source' software, simply put, is software whose source code is released openly, alongside the object code. Each user thus has access to the original source code (with no need to produce an inferior version by decompilation). The intent is that any user of the software is 'free' to modify, enhance or customise the program's source code, and use it to produce modified object code—ie to produce a customised, updated or improved version of the program. Hence the common description of open source software as 'free software': 'free' is used in the sense of 'free to modify' rather than 'free as in beer', although much open source software is both.
The hardest conceptual problem about open source software is how to ensure people play by the rules. There are many models, but the most common is to require any person who redistributes an open source program (whether in its original form, or with any changes they have made) to also redistribute the accompanying source code. The requirement is necessary to overcome human nature: after all, taking a program, making some changes, and redistributing the modified version while keeping the modified source code to yourself is an intuitively valuable way of creating and preserving a competitive edge over others. Anyone else who wants to replicate your modifications to the code must reduplicate your work themselves. Such a user could thus springboard off the earlier open source program to produce a closed source model for themselves.
The answer is easy, but apparently counterintuitive to many: it's done by copyright law.
Copyright and the GPL
The myth of contract
Most people seem to assume that, if one person (the original producer of some open source software) wants to control or limit another's actions (the subsequent modifier of that software), then there must be some kind of contract between the two. While it might be possible to achieve such protection by contract, it would not work in practice. As a general matter, to be enforceable, a contract must be entered into voluntarily by each party. There is usually no way to force another person into a contract. Thus, if the original producer of open source code made it available to others to access without restriction, there would usually not be—as a practical matter—any way to force that other person into the relevant contract. Even if it were, the transaction costs involved would be significant.
Enter copyright
However, software—in both its source and object forms—is a copyright work if certain conditions (such as originality and substantiality) are met. The original producer of open source software will thus own copyright in it. 2 Copyright is a bundle of exclusive rights, including the rights to control the copying and adaptation of the copyright work. The default position is that only the copyright owner may reproduce the work (ie make a more or less exact copy of it, or a substantial part of it), or make an adaptation (ie change it into another equivalent but different form of expression, such as translating it. 3 In the context of computer programs, this includes the sole right to copy the source code, make modifications to it, or compile it to form object code, and the sole right to copy the object code thereby produced.
Licences
There is nothing, however, to stop the copyright owner from permitting others to do one or more of those acts. There is also nothing to stop that permission being granted on terms. A grant of permission to a person to do something that that person is otherwise forbidden to do is quite common. In legal terms, that is the precise definition of a licence. 4 Licences can come about by contract, or by oral permission (subject to any applicable statute), and can be simple or convoluted, and narrow or broad, but all share the same basic form: a person who otherwise has the power to exclude another from an act, and who has the power to authorise that other person to do the act, does so.
The power to exclude can come from a number of sources, but the most common is property or statute: the occupier of land can allow another person to come onto that land, but in the absence of such permission could generally sue them for trespass. The owner of a car can authorise another person to use it, when without that permission such use would be a trespass and conversion of the fuel. And a person who holds copyright—which is itself a form of statutory property—can authorise another person to do some, any or all of the various rights which that copyright initially vests exclusively in the owner.
The way open source is usually made legally effective is to make the requirement to disclose the source code of any modification of the original source code a term of the permission to deal with the original—copyrighted—source code. Provided subsequent users adhere to the conditions upon which permission is granted, they are within the scope of the permission, and they cannot be sued for infringing the original user's copyright when they make any copies, or adaptations, of the original source code. Such qualified consent is common: the permission given to enter a person's house to attend a party does not give you the right to live there indefinitely; the permission given to enter a shop does not allow you to remain and start busking; and the permission given to affix a person's trade mark to specified goods does not allow you to affix that mark to any other goods. Stray outside the scope of consent, and you are again subject to the legal regime which gives the owner the right to exclude you.
The GNU public licence
The way this operates for open source software can be clearly seen in the 'GPL'—the 'GNU public licence'—one of the most common licences under which open source software is distributed. 5 In addition to various other requirements, section 3 of the GPL provides:
'You may copy and distribute the Program (or a work based on it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you also do one of the following:
a) Accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange.' (italics added)
Although paragraphs 3(b) and (c) allow two other means of compliance with the section, the method in paragraph (a) is probably the most common. Its effect is simple: if the original program is released under the GPL, then you may copy and distribute it, or a modified version, without infringing the copyright in that program provided you also distribute the relevant source code. The provision of your source code is integral to your ability to deal with the original source code. If you fail to distribute your source code, you have gone outside the terms of your original permission to deal with the copyright work, and (in the absence of any applicable defence, or other permission to deal with the work) you become liable for infringing the copyright in it.
To some people, this seems to be a novel use of copyright law. It isn't really: replace 'accompany the work with source code' with 'pay the copyright owner $x per copy', and you have something approaching a standard royalty contract. However, the elegant simplicity hides much power: remedies for breach of copyright are often more powerful than remedies for simple breach of contract. There are many subtle areas involved in open source law and the GPL, but the basic point is quite simple: 'it's the copyright, stupid'.
1 Called, strangely enough, a compiler.
2 This article does not discuss the various legal provisions affecting ownership of copyright, such as where an employee creates a copyright work in the course of his employment, or where a work is made 'for hire', or created jointly by several authors.
3 In the context of computer programs, adaptation will likely include re-expressing the work in any other type of code. Compiling source code probably produces an adaptation of that code, rather than reproducing it: see 'Copyright and Computer Programs: Data Access v Powerflex Before the High Court' (1997) 20 Sydney Law Review 296, available online at http://www.law.usyd.edu.au/~slr/v20/n2/bhc_kremer.html.
4 See Federal Commissioner of Taxation v United Aircraft Corporation (1943) 68 CLR 525 at 533: 'A licence provides an excuse for an act which would otherwise be unlawful as, for example, an entry upon a person's land, or the infringement of a patent or copyright. It is an authority to do something which would otherwise be wrongful or illegal or inoperative.'
5 Accessible at www.gnu.org/licenses/gpl.html. In a somewhat geeky in-joke, GNU itself stands for 'GNU's Not Unix', which is a recursively defined acronym where G actually stands for 'GNU'. Recursion (where a subroutine invokes itself) is a fairly commonplace feature of computer code, and non-geeks are advised simply to smile, back away, and leave such jokes alone.
This article is not intended to be nor should it be relied upon as a substitute for legal or other professional advice.
Copyright © 2004 Dr. Ben Kremer
|
|
Authored by: Anonymous on Sunday, April 04 2004 @ 09:57 PM EDT |
Wrong link:
http://www.groklaw.net/www.law.usyd.edu.au/~slr/v20/n2/bhc_kremer.html[ Reply to This | # ]
|
|
Authored by: jdg on Sunday, April 04 2004 @ 09:57 PM EDT |
Mistakes and Corrections Here
---
SCO is trying to appropriate the "commons"; don't let them[ Reply to This | # ]
|
|
Authored by: Taran on Sunday, April 04 2004 @ 10:01 PM EDT |
Actually, it's the General Public License - which, to the chagrin of whoever
responds to this post - I say is actually a Free Software license which has been
adopted by 'Open Source'.
Of course, whoever responds may say I'm a zealot or something, but hey... it's
accurate if you know the history, so I'll let that stand on it's own. Why not
just call it FOSS? Seems like everyone could work together that way.
The GPL itself allows the four freedoms which allow intellectual usability which
is important for a business which defines it's business logic, rather than
having the business logic defined by other companies.[ Reply to This | # ]
|
|
Authored by: tknarr on Sunday, April 04 2004 @ 10:28 PM EDT |
I especially like his analogy to a standard royalty contract. I've always
considered the GPL to be very similar to most other software agreements, the
only difference being the coin you pay in. Other licenses merely require you to
pay in dollars/euros/(substitute-your-currency-here) while the GPL requires you
to pay in kind: you received code, you pay in code.[ Reply to This | # ]
|
|
Authored by: surak on Sunday, April 04 2004 @ 10:28 PM EDT |
I>Historically, software vendors kept source code to themselves (indeed,
invariably protecting it as valuable trade secrets) and sold, or licensed, only
the object code versions of programs to their customers.
This is part
of the problem. This is just flat out wrong. The reason many people tend to
think this is true is because when *they* got into computers, closed source was
actually the norm. But it is more accurate to say that historically the source
has been open, rather than closed.
When Richard Stallman got the idea to
start the Free Software Foundation, he encountered something he'd never found
before: a closed source printer driver. He had become accustomed to being able
to change the printer driver's source code under previous revisions of the
system he was running on. But when the new version came out -- poof, not only
was the source code not included, but they refused to give him a copy unless he
signed a draconian NDA.
You see, before this time, source code either
generally accompanied the program, the program was distributed only as source
(like Unix), or was at the very least generally available. There were no
licenses, no GPL, everything was done on the honour system -- there was a
unwritten rule that source code was part of the system.
The inventor of
closed source, btw, is probably Microsoft. They were the first company in
existence to sell only software. At the time, it was considered a pretty risky
proposition. It's only today when closed source is the rule and not the
exception that we think that "historically" software is provided as closed
source. End users today put up with closed-source proprietarily-locked-in
software only because they don't know any better...as far as most of them know
it's always been this way. But, as I said, that's just blatantly false
misinformation.
[ Reply to This | # ]
|
- Misinformation - Authored by: jkondis on Sunday, April 04 2004 @ 10:31 PM EDT
- Misinformation - Authored by: Anonymous on Monday, April 05 2004 @ 01:43 AM EDT
- Misinformation - Authored by: vonbrand on Sunday, April 04 2004 @ 11:02 PM EDT
- Misinformation - Authored by: rkrishnam_can01 on Monday, April 05 2004 @ 12:11 AM EDT
- MS isn't the first - Authored by: Anonymous on Monday, April 05 2004 @ 12:23 AM EDT
- Early closed source. - Authored by: davcefai on Monday, April 05 2004 @ 01:20 AM EDT
- Remember BBS's? - Authored by: Anonymous on Monday, April 05 2004 @ 01:33 AM EDT
- Yup - Authored by: Anonymous on Monday, April 05 2004 @ 10:48 AM EDT
- Re: Yup - Authored by: Anonymous on Tuesday, April 06 2004 @ 12:45 AM EDT
- Re: Yup - Authored by: surak on Tuesday, April 06 2004 @ 07:47 AM EDT
- Misinformation - Authored by: Anonymous on Monday, April 05 2004 @ 01:53 AM EDT
- Misinformation - Authored by: Anonymous on Monday, April 05 2004 @ 04:38 AM EDT
- Misinformation - Authored by: soronlin on Monday, April 05 2004 @ 07:20 AM EDT
- Misinformation - Authored by: blacklight on Monday, April 05 2004 @ 07:31 AM EDT
- Misinformation - Authored by: Anonymous on Monday, April 05 2004 @ 07:43 AM EDT
- Misinformation - Authored by: codswallop on Monday, April 05 2004 @ 08:02 AM EDT
- Misinformation - Authored by: cyclists2 on Monday, April 05 2004 @ 08:45 AM EDT
- Misinformation - Authored by: Peter H. Salus on Monday, April 05 2004 @ 09:27 AM EDT
- Misinformation - Authored by: decsnake on Monday, April 05 2004 @ 01:31 PM EDT
|
Authored by: Anonymous on Sunday, April 04 2004 @ 10:30 PM EDT |
here [ Reply to This | # ]
|
|
Authored by: pesc on Sunday, April 04 2004 @ 10:42 PM EDT |
Historically, software vendors kept source code to themselves...
Well,
historically, most important software such as the operating system, compilers
and software libraries did not come from "software vendors" but from the
hardware vendor. The software could not easily run on the competitors machines
anyway, so it was fairly common that you could get the source code for these
components from the hardware vendor. This allowed you to learn more about how to
use the hardware you have bought, and fix minor problems yourself. (The same
benefits that are touted for open source software.) The hardware vendor retained
the copyright to its code.
So in the early 80's I had the opportunity to
examine the source code for the SINTRAN-III OS for the NORD-100 computers (a
Norwegian brand) and for VAX/VMS (which is often regarded as a very proprietary
operating system).
So I would argue that the notion of "Open Source" is at
least as old as "Closed Source" (although it wasn't called "Open Source" at that
time). The rise of "Software vendors" and secretly guarded "Closed source" are
much newer. [ Reply to This | # ]
|
|
Authored by: rsteinmetz70112 on Sunday, April 04 2004 @ 11:08 PM EDT |
Clear concise and coherant.
cc: IBM legal team.[ Reply to This | # ]
|
|
Authored by: mobrien_12 on Sunday, April 04 2004 @ 11:44 PM EDT |
Thanks again for finding and publishing all these good anti-fud stories PJ. [ Reply to This | # ]
|
|
Authored by: Thomas Frayne on Sunday, April 04 2004 @ 11:46 PM EDT |
In
license plus offer , I stated that I agree with Kremer that the GPL can be
interpreted as a license with conditions. However, interpreting the GPL as a
license plus a public offer makes all the legal implications clearer.
I continue to think that the license plus offer interpretation is the
easiest one for both lawyers and lay persons to understand. However, either
interpretation makes it obvious that the GPL validly places binding legal
obligations on those who distribute GPL'd programs, and that those who violate
those obligations and continue to distribute violate copyright law.
[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 12:23 AM EDT |
I'm glad Dr Kremer noted "'free to modify' rather than 'free as in
beer'". However, as so many people do, he's confusing Open Source and Free
Software. The article continually confuses the two. These are terms with very
specific meanings, used by the Open Source Initiative and the Free Software
Foundation. Richard Stallman created the FSF, and the GPL.
Open Source software is NOT Free Software. There are many different Open Source
licenses, and some of them are not compatible with the GPL.
See osi.org and fsf.org for lists of licenses, licensing terms, and which
licenses are GPL-compatible.
I think the GPL is a work of genius. How about we give credit where credit is
due, and use the right terminology. :)
Here are Richard Stallman's Four Freedoms, which are fundamental to
understanding the GPL and Free 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 neighbor (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.
[ Reply to This | # ]
|
|
Authored by: xyzzy on Monday, April 05 2004 @ 01:20 AM EDT |
It would not serve this articles purpose (an excellent PHM article) but it would
be good for there to be a counter point specifically about the shared source and
licensed proprietary source that you can use under very strict rules and only
when you pay extra contrasting to Open Source.[ Reply to This | # ]
|
|
Authored by: ExcludedMiddle on Monday, April 05 2004 @ 01:36 AM EDT |
This article is quite good! Now I wish that it just made ONE point more
clearly:
If you just download a FOSS program and run it on your own system, you may
modify the code all you like without releasing it to the public. It's
distributing the program that triggers the release of source code. We've brought
this up again and again in these discussions.
That point is the one that seems to be missing from most of these articles about
FOSS software, and arguably needs clarification the most, from a legal
standpoint. It's also the most powerful concept behind FOSS. Critics often harp
on the idea that those "nameless" creators of the software could
disappear tomorrow, and then you'd be stuck. They seem to fail to understand
that you still have the source, and could continue just fine from there, no
matter what happens to the creators.
I just can't emphasize how powerful the concept of having access to the source
in the first place is. And if you consider the changes you make secret, have no
fears that you'll need to show the world your code as long as you don't
distribute. Most companies out there are not computer vendors, and just want to
modify the programs for their own business purposes. There's no issues with
releasing source code in these situations.[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 01:38 AM EDT |
For all of the people that are still confused by Daniel Wallace's ramblings
how "IBM knows GPL is a contract", here is that text from IBM's
countersuit:
28. The Linux kernel is subject to the GPL as it is
comprised of programs and other works that contain notices placed by
contributing copyright holders permitting distribution under the terms of the
GPL. The Linux developers' public agreement to apply GPL terms expresses
in a binding legal form the conscious public covenant that defines the
open-source community -- a covenant that SCO itself supported as a Linux
company for many years.
The above text does not say that the
public covenant (contract) between Linux developers is the text of the GPL, as
Wallace would have you believe. The above says that all those developers,
including SCO, agreed (i.e. formed a contract, covenant, agreement,
understanding) to license their works under the GPL. They did that through
discussion and their actions. That is completely different
than saying that GPL is the text of the contract they have formed.
The
"apply GPL terms" part relates to copyrighted works, not
people.
It would be like saying that developers of Apache formed a
contract that is the Apache Software Licence. Complete nonsense. One needs to
sign a piece of paper to become an Apache committer and sign off copyrights to
Apache Foundation. In the same sense, Linux developers agreed (without actually
having a predefined piece of paper) that they will continue releasing their work
under the GPL, while each preserving copyright to the respective
part.
Daniel Wallace is the master of manipulation and word twisting.
Beware of his writings, they are highly misleading. [ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 01:46 AM EDT |
Yesterday I saw an article with the word "fiasco" in the headline. I
immidiately assumed that this was another story about SCO.
This must be the first sign that I´ve read way to much about this SCO nonsense.
Good thing that I am going to vacation so that I can 'detox'.
Lars
[ Reply to This | # ]
|
|
Authored by: stevenj on Monday, April 05 2004 @ 02:58 AM EDT |
But then, it doesn't seem like he's visited GNU's web site, considering the
other terminology he's slightly garbled.[ Reply to This | # ]
|
|
Authored by: JOff on Monday, April 05 2004 @ 03:23 AM EDT |
" [..] you have something approaching a standard royalty contract."
The very idea of a roalty is that something (money) flows back from
the user to the author. This is not the case with the GPL: nothing
flows directly backwards. Any user may modify the source, and if
he passes the result to someone else, he has to give away the
sources with the program. Only as a side effect this usually makes
some kind of a publication: it could as well stay with the new user
without ever becoming public. And only if the modification is
published it will become usefull to others. The important thing is:
the publication of any piece of the code is not forbidden, so if
something is usefull, it is very likely to become public sooner or
later.
Now everybody who modifies a piece of GPLed software and
gives it away is in a specific position: having inherited the work of
others he has to allow further inheritance.
This is more like the use of scientific publications: If someone
publishes the results of his scientific research, generally he will
cite other publications correctly and name them (to form a
continous trail of argumentation in written papers open to follow for
everyone). If one fails to cite others that contributed to his work, he
will be discredited in his scientific community. Anybody relies on
the past work of others, and you have to give credit where credit is
due: you have to name those publications that your own work is
based on.
You are allowed to use other's work for your own publication, as
long as you play to the rules: you will cite correctly, you will also
mention opinions that do not agree with you, and you will allways
tell who did that past work. By doing so you put your successor in
research, that may be anybody who reads your work, into the very
same position you have been when you started your own work: all
the past publications are available to you as the base for your
research.
(All that about the scientific publication process describes some
ideal, unrealistc, theoretical view, given all those licenses, fees,
patents and the like).
JOff [ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 07:42 AM EDT |
As many people have pointed out, there is no such license as the "GNU
public license" and there are plenty of "General Public
Licenses." But there is only one "GNU General Public License
(GPL)" which is the license for most of the GNU System including the Linux
kernel.
It is a good sign that lawyers in Australia seem to start discovering the topic
and this provides a solid first step. I think it contains too many weaknesses to
be called a "first-tier" reference, though.
P.S. Referring to the most wide-spread Free Software license by the co-opting
terminology "Open Source" is definitely going to turn some stomachs
wich the people who wrote (and still maintain) the GNU GPL.[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 08:19 AM EDT |
I've just found an interesting site on the BBC webpage that the readers of
Groklaw might appreciate.
http://news.bbc.co.uk/1/hi/in_depth/business/2000/microsoft/default.stm
It concerns the history and implications of the Microsoft antitrust case. I
recommened the video archive, on the top right of the webpage, as it shows
statements by Bill Gates and Janet Reno after the break-up ruling. It's
interesting to see how much of a influence the government can have over the
outcome of Anti-trust rulings in the United States.
Further, the shear lies spouted by Gates during his statements go far to
illustrate to mindset of the person Free Software has acting against it. Gates
is so ingrained in his rather unique perception of innovation and competition
that by the end of his six minute rant, one almost expects to see him carted off
to some asylum.
Stephen[ Reply to This | # ]
|
|
Authored by: meat straw on Monday, April 05 2004 @ 09:27 AM EDT |
Wired has an article on
problems with patent law andtech. [ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 09:42 AM EDT |
It seems reasonable that some things encourage innovation and others discourage
it. I suspect that the GPL does not discourage innovation and probably
encourages it. I also suspect that GPL may be better at encouraging innovation
than a BSD type license (but I can't actually prove that).
Other things ,like having to fill in a five page document to contribute to an
open source project, may discourage innovation. I suspect that a closed source
monoculture discourages innovation.
I have a couple of questions:
1 - What things are necessary and sufficient to encourage innovation?
2 - Are there sources like "The Cathedral and the Bazaar" that I
should be checking out on the subject of innovation?[ Reply to This | # ]
|
|
Authored by: DarlingMcBribe on Monday, April 05 2004 @ 09:48 AM EDT |
This articale mentions that Open Source software
provides freedom but it
is often free cost as well.
I'd like to point out that the word "often" can
be easily
replaced by "always" in the statement.
Many FOSS
advocates often emphasize that the word Free
in FOSS refers to liberty and not
cost,
i.e. free as in free speech not as in free beer. However,
for all
practical purposes the liberty actually
enforces the zero cost, so there is no
point to
distinguish the two.
Let me get the facts lined up
first:
- "You may charge a fee for the physical act of
transferring a copy, and you may at your option offer
warranty
protection
in exchange for a fee."
(from the GPL terms
and conditions,
point 1), see also
th
is
link in the GPL FAQ
- If you distribute code under the GPL,
anybody who
receives it will also have the right to further distribute
it,
see this
link
in the GPL FAQ
- "Each time you redistribute the Program (or any
work
based on the Program),
the recipient automatically receives a license
from the
original licensor to copy, distribute or modify
the Program
subject to these terms and conditions. You may
not impose any further
restrictions on the recipients'
exercise of the rights granted herein. "
(from the GPL terms
and
conditions, point 6)
- "You can charge nothing, a penny, a dollar, or a
billion dollars. It's up to you, and the marketplace, so
don't
complain to
us if nobody wants to pay a billion dollars
for a copy." - from
this
link about selling GPL-ed code.
So, on one hand it is
true that one can charge a
reasonable fee (fact 1 above) for the distribution
(burning
of CD, printing manuals etc), and you may also charge for
additional services, e.g. support. On the other hand,
there is no way
that you can charge sufficient
price to recover development cost. Suppose, as
a
commercial
company, we develop a large application with 100 men years
of
effort (e.g. 20 developers over 5 years) at
a cost of about 10 million
dollars. How can the company
recover that cost if they GPL the product ?
Every user is entitled to put it up on the internet for
free download
with the source code (facts 2 and 3 above).
So, by simple market logic, if I
charge more than the
distribution cost, then my first user can force me out of
the market by selling it cheaper (as he does not need to
recover the
development cost). It seems, the only way to
recover the development cost is
to charge it all to the
first customer, i.e. set the price to 10 million
dollars.
But, who is going to buy the product for that price ?
Nobody (fact
4 above).
Another popular argument is that there are commercial
Linux distributions, which proves that one can charge for
Free Software and
build sound business on it. This is
again a false statement. Red Hat, SuSE,
Mandrake etc.
charges
for a box, which includes a book, CDs (or DVD) and it
has
an associated cost to distribute such. When you buy the
box
you are
paying for those materials included and the cost
of distribution plus the cost
of support. You are NOT
paying
for the software. The software is still free
(as in zero
cost) even in that case.
It is clear that despite the
pupolar urban legends that
a GPL software can be commercial, it simply
does
not make any economic sense. Hobby projects,
University projects and projects
sponsored by companies
that make money on something else, e.g. hardware (see
IBM)
may find it good business to release code under GPL.
However,
commercial software development is impossible
with the GPL.
If you
say that you do not need commercial software
development, because you have
GNU/Linux, Open Office etc.,
then think again, and think about specialized
applications, e.g. in Medical imaging, Computational
Chemistry,
Computational Fluid Dynamics, Computer Aided Design,
Molecular Modelling
etc.
--- IANAL, IANAG, IARASC (I am running a software company)
[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 10:52 AM EDT |
Has anyone finished this yet?
Bit by Bit and I'm almost done and I must say (despite being hard work in
places) it has opended my eyes and I thought I was far from blind on this
topic.
[ Reply to This | # ]
|
|
Authored by: rdc3 on Monday, April 05 2004 @ 10:54 AM EDT |
This equation may be a useful way of grokking and remembering the distinction
between the open source and free software concepts (to a first approximation, at
least). The critical difference is that free software licences necessarily
include a freedom-preserving requirement on the distribution of derivative
works, whereas open source licenses may permit redistribution without such a
requirement. (See the gnu.org and opensource.org websites.)
The BSD
license is a well-known example of an open source license that does not qualify
as a free software license.
Redistribution is permitted but no requirement
that source-code freedom be preserved is imposed. Modified versions can thus be
redistributed as closed source
software.
Thus, I think that Ben
Kremer's article is not quite correct in attributing the need for a
freedom-preserving requirement to the open source model. Nevertheless,
I
believe it is fair to say that many of us in the open-source community
appreciate freedom preservation
as an important and fundamental added
value.
Prof. Rob Cameron
Simon Fraser University [ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 11:39 AM EDT |
We build something that threatens their way of life.
so of course they will attack it.
Dont let the Bas**rds grind you down.
Hey, we didnt like paying for their software that doesnt work so we got together
used our brains and made something that does work. We can give it away if we
want, it is ours.
Hey M$ we dont want to start charging, perhaps you should fix your code, open it
up and give it away free, stop winging.
[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 11:41 AM EDT |
Sorry lost track.
When will be the next action on the IBM or Novell cawes??[ Reply to This | # ]
|
|
Authored by: grundy on Monday, April 05 2004 @ 11:59 AM EDT |
ThisStuff Copyright © 2004 by Me
You may do whatever you jolly well
please with ThisStuff so long as you keep it to yourself.
If you pass any
of ThisStuff along to anyone else, it must be people readable and with
this exact same license, even if you have added, changed or deleted something
in it.
And it would be courteous to put your name to your
modifications.
[ Reply to This | # ]
|
|
Authored by: Anonymous on Monday, April 05 2004 @ 09:19 PM EDT |
GNU General Public License, not GNU Public License or "the" General
Public License.
GNU GPL.
There are other general public licenses, but this is the GNU one.
[ Reply to This | # ]
|
|
Authored by: jbn on Tuesday, April 06 2004 @ 04:18 AM EDT |
I cannot second your recommendation for understanding the GNU General Public
License (GNU GPL) by reading this paper. I'm reminded of Mark Webbink's
intellectually dishonest paper about licenses (for those who don't know,
Webbink, chief counsel for Red Hat, needlessly renamed, reinvented, and took
credit for the idea of copyleft as
expressed in his paper. Webbink's paper bends over backwards to exclude the
free software movement. To read Webbink's paper you would never know that the
folks at the Free Software Foundation coined the term "copyleft" roughly two
decades before Webbink's paper came out and the FSF has been using this very
useful concept (based on a software license's preservation of the freedoms of
free software) to categorize software
licenses).
Getting back to the Freehills paper, reading the first
sentence of the open source definition tells you that "Open source doesn't just
mean access to the source code.". So it is confusing the reader to focus on
source code ("'Open source' software, simply put, is software whose source code
is released openly, alongside the object code."). It's ironic but widely
known that the term "open source" was intended to replace and clarify the
ambiguity of the older term "free software" and yet the "open source" term
raises problems of its own both in terminology and in understanding the
definition of that term.
Readers will not better understand the GNU GPL
by ignoring the language used in that license. The Freehills paper reads as if
this is just a big setup to talk about a different movement. The GNU GPL cannot
be properly understood in the context of the open source movement. Even a
casual read of the GPL will reveal that the document makes no mention of "open
source" at all, nor should it. The GPL (including version 2 which is currently
in use) predates the open source movement and had no authors who work in the
open source movement. The GPL speaks of software freedom and the open source
movement dismisses software freedom (the OSI takes a pause from dismissing it to
ridicule it calling it "ideological tub-thumping"). The Open Source Initiative
merely defined the terms of license acceptance widely enough to include the GPL.
Pushing aside the author of this most important license by citing a
johnny-come-lately movement is disturbing, to say the least. It's fair to say
that the FSF knows the GPL a hell of a lot better than most people and when they
give seminars on interpreting the GPL they don't forget to explain why it says
what it says.
Then there are some smaller points, two of which I'll
mention here:
- What the Freehills article describes as the
"relevant" source code is much better spelled out in the text of the
GPL--"complete corresponding source code". This language is critical to
understanding which source code can satisfy a request for source code.
An earlier or later version of the program (irrespective of a version number) is
insufficient to satisfy this language in the GPL. It's not hard to comply with
the GPL here, merely make a copy of the source files that built the binaries
being distributed and keep those exact files around for at least 3 years. Or
don't distribute binaries at all
(only distribute the source code); there are
plenty of talented packagers waiting to distribute binaries for
you.
- "To some people, this seems to be a novel use of copyright law.
It isn't really: replace 'accompany the work with source code' with 'pay the
copyright owner $x per copy', and you have something approaching a standard
royalty contract." I don't see how the novelty of an idea can be legitimately
diminished by replacing chunks of the idea with something else which is less
novel. The reason the GPL is novel is that no license before it defended the
freedoms this license defends; no software license before the GPL created the
community this license had a hand in building. Why was this whole "[the GPL is
not a] novel use of copyright law" language in here at all? Why not say that
these parts of the GPL remind the author of a royalty
contract?
All in all, I think you're better off following
discussion from RMS, reading the GNU GPL FAQ, and signing up
for the GPL seminar the next time the FSF offers it. [ Reply to This | # ]
|
- On novelty... - Authored by: Anonymous on Wednesday, April 07 2004 @ 12:50 PM EDT
|
Authored by: Anonymous on Tuesday, April 06 2004 @ 10:52 AM EDT |
It seam kind of crazy, that the close source fear its rival the open source
could in fact be the cl;ose source that is committing a crime free and clean
because it illegal to check their code. While the open source brand has no one
to fear because they know their code it will be check. Plus, they would not
release the code if the did something illegal.
Open Source is easy to verify. That is anyone can check the code to see if the
author is the one who wrote the code. But in close source. A person must
illegally decode the binary program to verify who has ownership of the code.
I wonder how many people or businesses own close source code that are not truly
100% owned by them.
[ Reply to This | # ]
|
|
Authored by: Anonymous on Tuesday, April 06 2004 @ 12:00 PM EDT |
Fogive the question...I'm new to FOSS and a recent convert from M$ and closed
source (I still can't bring myself to chuck my Windows box though...although I'm
working on it).
This article was very helpful in understanding the GPL but I don't understand
Redhat's implementation and therefore I am left still confused as to how the GPL
works.
By my understanding of this article, (and in summary), if you release software
under the GPL, you must also release the source code. Okay, I get that...but
how then can Redhat no longer offer a FOSS version of their Enterprise RedHat
products? Is not RedHat Enterprise Edition a modified verion of a GPLed work,
and therefore are they not required to release the source? I know they offer
Fedora, but Fedora isn't the same code, (not even the same kernel version). How
can they do this, and still be GPL compliant? I need to switch to WhiteBox or
some other distro! I think what RedHat is doing is bogus.[ Reply to This | # ]
|
- I need help! - Authored by: Anonymous on Tuesday, April 06 2004 @ 02:54 PM EDT
|
Authored by: Anonymous on Thursday, April 08 2004 @ 06:50 PM EDT |
Funny, contract law doesn't seem to agree with you.
This sounds more like the
weaseling excuses that kids come up with to try and bail their butt out of
trouble with their parents.
A contract is a contract, a license is a license.
Nuff said.[ Reply to This | # ]
|
|
|
|
|