Open Source and Free Software:

Just for the record, in a response to my earlier posting about free software and copyright law, Richard Stallman sent me the following:

I am disappointed that you describe my work as "open source", because I disagree with that camp and I constantly struggle against the misinformation which labels my work that way. You might as well call my work "Republican". For instance, this sentence

Stallman understands this thoroughly ­ though the vast majority of commentators on the open source movement have missed this point.

is likely to lead readers to suppose erroneously that I am a supporter of the open source movement, ewhen in fact I disagree with it fundamentally.

See for an explanation of this disagreement.

Would you please replace "open source" with "free software" in this posting about my work?


einhverfr (mail) (www):
Sometimes RMS needs to realize he has lost this one too.

Oh, and I use Linux, so there ;-) (I won't spend the extra 2 syllables RMS wants me to add onto that name either.)

Really, "Open Source" and "Free Software" are flip sides of a coin. They don't exist without eachother ever.
3.31.2009 12:29am
Seriously. If enough people consider two terms to be synonymous, guess what? They are synonymous. The distinction that article tries to make would go against how any person I've ever heard use either term understands their meanings.
3.31.2009 12:39am
einhverfr (mail) (www):

RMS is known for pedantic and meaningless arguments, and I referenced a second one in my post.

BTW, as mentioned in the other thread I am not RMS's biggest fan. There are reasons to dislike his politics which have much more bearing than most of the cheap-shots people make. My biggest concern is the expectation that all GNU project maintainers will politically back him up all the time and his willingness to ask them to resign when they don't side with him regarding other open source projects (Thomas Bushnell's resignation from HURD is a shining example of this-- the issue has to do with Bushnel had failed to back RMS up when the Debian organization decided that the GNU Free Documentation License wasn't Free enough).

Most of it is about seeking power and control, IMO. THis is why although I like SOME of RMS's ideas, I think the man has serious control issues and is often worth avoiding.
3.31.2009 12:49am
E.L. (mail):
They are not synonymous as they have different intensions and extensions.

Open source software is principally based around the idea that software whose source code is available and modifiable by anyone will ultimately be more optimal (secure, powerful, or something else) than closed source software.

Free software is principally based around the idea that software should not restrict the freedoms of its users to do whatever they want on their computer. In other words, software should be enabling.

Accordingly, as RMS points out, there can be open source software which is not free software, such as open source DRM.
3.31.2009 12:51am
no one thinks the two are synonymous. is anyone surprised when something is both open source and for sale? the distinction is sometimes unimportant, which is why people are rarely precise as to which they are talking about.
3.31.2009 12:51am
E.L. (mail):
FWIW, I don't really care one way or the other. But I think RMS does have a legitimate philosophical point.
3.31.2009 12:52am
Kat (www):
But they're not synonymous, and they do exist without each other. Or at least, all free software is open source software, but not all open source software is free software. An audience of law geeks ought to appreciate terms with distinct and specific meanings (despite being popularly misunderstood); liking/agreeing with RMS not required.
3.31.2009 12:56am
Milhouse (www):
DAS, Free software can also be for sale.

Kat, what is an example of actual software that is open source but not free?
3.31.2009 1:19am
I'm not getting the difference between the two terms. If the source is open, isn't it free? The linked article is not terribly enlightening. It almost sounds as though the primary difference is that "free software" has the word "free" in the name.
3.31.2009 1:27am
Soronel Haetir (mail):
I would point out various MS run-time libraries as open-source but not free. The visual c++ run-time, MFC, some others are included with the product. You are even allowed to modify and distribute binaries of those libraries, though not your modified source.

Not that many practicioners would consider that open source, more like source included.
3.31.2009 1:38am
There are many examples of "open-source" but non-free software. For example, the old Mozilla license required any modifications you wanted to release to be sent back to the original developers. A lot of source code is released with the proviso that it may only be used for non-commercial applications (with a paid "commercial license" for other uses). The NASA "Open Source" license used to only allow you to release modified code if you personally created the modifications (rather than, say, you incorporated code elements created by someone else).
3.31.2009 1:40am
einhverfr (mail) (www):
E.L: Why would it be impossible to have Free Software for DRM. The only argument would be whether the definition of Free Software involves technical limitations posed on the use of that software (rather than source-code/programming limitations). So I am not sure that holds.

The 4 freedoms can apply to DRM just like any other program.

The only question becomes whether you think some marginal licenses (like the Qmail license) are even "open source." But these are technical rather than substantial quibbles.
3.31.2009 1:42am
Jim at FSU (mail):
DRM relies on the user not having the ability to dive into the source code of the client and disable it. DRM barely works in the closed source world.

If software has a remote component that is under the control of the DRM provider, that remote component had better be utterly indispensable to the operation of the software. If it isn't, say goodbye to the DRM.

The closest you can get to DRM or copy protection that works in an open source environment is an online authentication based model for products that consist primarily of online functionality.
3.31.2009 1:59am
Bill Poser (mail) (www):

Seriously. If enough people consider two terms to be synonymous, guess what? They are synonymous. The distinction that article tries to make would go against how any person I've ever heard use either term understands their meanings.

Hunh? The distinction between "free" and "open source" is very well known in the software world. Lots of people make the distinction. Actually, part of the problem here is that "open source" is used in two senses. Some people use it as a synonym for "free". In this case, RMS does not disagree with the movement, which is the one he founded, but with the choice of terminology, which fails to emphasize the freedom aspect. Other people use "open source" more literally for software whose source is available but which may lack the four freedoms. In this case, RMS disagrees not with the term, which is accurate, but with the approach.

There is a problem with "free software", namely the fact that people tend to take it to mean "free as in beer", not "free as in speech". I likethe term that has gained considerable currency in Europe "software libre".
3.31.2009 2:13am
Bill Poser (mail) (www):

RMS is known for pedantic and meaningless arguments

I don't think that his arguments are pedantic or meaningless. He's just persistent, consistent, and careful. For RMS to be called pedantic on a legal blog is the epitome of the pot calling the kettle black.
3.31.2009 2:24am
That explanation of the distinction as proffered by Stallman is the opposite of enlightening. I still don't understand it.

Nonetheless I find the insistence that the GNU GPL is consonant with free, unfettered software to be odd, since any use of software under the license causes an entire product to be licensed under GPL. I understand the point of that restriction from GNU's perspective, but it's not the same as free.
3.31.2009 2:25am
David Schwartz (mail):
I agree. I cannot understand what distinction RMS, or the people responding here, are trying to make.

Every time you say "X is an example of something that is Open Source but not Free Software" (or the other way around), the same thing you think make it sort of not one, IMO, also makes it sort of not the other.

"Open" and "Free" are synonymous here. "Free" means you can do what you want. "Open" means there aren't restrictions on what you can do with it. That sounds pretty similar to me.

When software is "Free Software", that has to refer to the source. Without the source, you aren't free to even understand it, much less modify it.

The only possible confusion Stallman could possibly be talking about is people misunderstanding "Free" to mean "at no cost". But *nobody* did that. So what is he complaining about?
3.31.2009 3:14am
MichaelJ (mail):
I think this is a pretty clear case of jargon not meaning quite what it looks like it means. "Free software" and "open source software" usually mean the same thing in the software community. What some posters describe, where the source code is available for viewing but not modification, is usually called "shared source." Software available at no cost (but not modifiable or redistributable) is called "freeware."

Just because disambiguation is cool, some call it "free/libre open source software." Yep. FLOSS.
3.31.2009 4:05am
Bill Poser (mail) (www):

"Open" and "Free" are synonymous here. "Free" means you can do what you want. "Open" means there aren't restrictions on what you can do with it. That sounds pretty similar to me.

No, you're using definitions of your own creation rather than the ones in use in the software world. When "open source" is used in the sense that is not synonymous with "free", it means that the source code is available for inspection. You can read it. That is sufficient to make it "open source". But that may be all that you can do with it. In the absence of a license permitting you to do so, you can't redistribute the source and you may not even be able to modify it for your own use or compile it for your own use.

What makes software "free" is the ability to modify it and redistribute it. Proponents of free software differ as to whether to be laissez faire or to insist that software that is once free remain free. The GPL restricts your freedom in certain ways by requiring that that GPL-ed software remain free - you can't take some GPL-ed software, incorporate it into your own project, and release the whole under a proprietary license.
3.31.2009 4:07am
David Schwartz (mail):
BP: I know a lot of programmers and professionals in the software world. I don't know anybody who would ever use the term "open source" that way. This does not match the Open Source Initiative's definition of "open source".

Are you seriously arguing the OSI's definition of "open source" is not "in use in the software world"? Most people in the software world would point you to that definition if you asked them what "open source" means.

The license shall not restrict any party from selling or giving away the software as a component of an aggregate software distribution containing programs from several different sources. The license shall not require a royalty or other fee for such sale. ... The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software.
OSI's definition of open source
3.31.2009 5:09am
Chuck Jackson (mail):

Re FLOSS, see Eric Raymond's blog
for a criticism of that usage.

3.31.2009 7:34am
Prasenjeet Dutta (mail) (www):
> Kat, what is an example of actual software that
> is open source but not free?

PGP (the encryption software) is a very good example. The source code is available at but they're pretty clear that all they are providing it for peer-review only; downloading the source does not give you a license to use the software gratis.
3.31.2009 8:04am
David Schwartz (mail):
PGP (the encryption software) is a very good example. The source code is available at but they're pretty clear that all they are providing it for peer-review only; downloading the source does not give you a license to use the software gratis.
Fortunately, you don't need a license to use a copyrighted work when your copy was lawfully made and you lawfully possess it. This is an essential rule of law -- otherwise, I could drop a million copies of my poem from an airplane and then sue everyone who reads it.
3.31.2009 9:00am
David Schwartz: The DMCA anti-circumvention provisions are at odds with your essential rule of law.
3.31.2009 9:19am
David Schwartz (mail):
billb: To the extent that's true, and it's not as much as many people think, that's a bug that should be fixed.

There are two lines of cases on this. In one line, the courts have done the right thing and argued that if it isn't for an infringing use, it's not circumvention. Storage Tech v. Custom Hardware Engineering (421 F3d 1307) is a key case in this line.

Then there's another line of cases that has held that fair use is not a defense to circumvention. Universal v. Reimerdes (273 F3d 429) is a key case in this line.

These two lines of cases are logically incompatible. If you can't circumvent if the use isn't infringing, then that the use doesn't infringe because it is fair use must be a defense to a charge of circumvention.

A friend of mine, who is a real IP lawyer, knows the people who drafted the anti-circumvention statute. I have it on good authority that this was unintentional and their intention, at least, was that circumvention be clearly facilitating an unlawful use. They were thinking about devices to decrypt cable and satellite tv.

That's what this language was supposed to do, "... to 'circumvent a technological measure' means to descramble a scrambled work, to decrypt an encrypted work, or otherwise to avoid, bypass, remove, deactivate, or impair a technological measure ...".
3.31.2009 9:49am
David: I understood that to be the case, too. I think that Bill Patry even mentioned as much on his blog. Considering the flak that DVD John and 2600 took, though, the safe approach seems to be to ignore your advice and follow your license. :)

Getting back to the substance of the conversation, I like to think of three categories: "Source Available" which is what it sounds like. You can get the source code, but you have no distribution rights or responsibilities outside of copyright law. "Open Source" which follows the Open Source Definition. And, "Free Software" which follows the Free Software Definition.
3.31.2009 10:58am
Stormy Dragon (mail) (www):
The primary difference is that "open source" (e.g. the BSD license) is about freeing up the developers and "free software" (e.g. GPL) is about freeing up the users.

For instance, the GPL requires that if I take your software and modify it, I must make the modification available to the public. That makes me less free as I now have an obligation to provide services to various other people. Under BSD, though, I can take the code and do whatever I like with it.
3.31.2009 11:22am
einhverfr (mail) (www):
Ok, let me give you a hypothetical example of a DRM solution where the software component would meet Stallman's published definition of Free Software, special pleading aside. My reading of the GPL v2 and GPL v3 suggest that such a solution would be allowed under either license (though discussion drafts contained language to prevent it, this language was removed due to industry pressure).

Of course, most DRM solutions would be ineffective if released as either Free Software or Open Source so one can't say DRM can be just Open Source and not Free Software. In this case, I propose a hardware enforcement mechanism which authenticates an executable to a piece of media. In this solution, media is encrpyted with a shared secret key (say, using RC4), and a plain-text x.509 certificate is appended onto the end. The hardware solution reads the x.509 certificate, validates it with the issuer, and then checks to ensure the executable is signed by the issuer of the certificate. If all of this is correct, the hardware side requests the media key from the certificate issuer and decrypts the media, handing it to the executable.

Now, the original author of the software is in the hardware, not the media business, so that business releases the software component under the GPL v2 (could be the GPL v3) and includes documentation that to make the DRM work, one must set up a certificate authority, etc., and we are going to recommend GNU TLS for that..... Media businesses can add features if they want, but they MUST distribute the source code of their versions. The only thing they don't have to distribute are the private keys used to make the signature files.

Now, this meets all of Stallman's definitions as regard free software:

1) You can use the software anyway you want within the limit of the hardware platform. You can even rip your own cd's create your own certificate authority, etc. and drm your own music collection!

2) You get access to the source

3) You can create derivative works from the source

4) You can redistribute the software and derivative works however you like provided that you pass this freedom on.

Now, you may argue that the digital signatures are a part of a functional solution, but I would argue this is only necessary if the hardware refuses to run the software without the original developer's signature. Furthermore the hardware ties the distributed media to the build of the executable, and does not prevent you from creating media that is tied to the executable.

Unless one resorts to special pleading, all of Stallman's four freedoms remain intact on this one.
3.31.2009 11:33am
einhverfr (mail) (www):
Stormy Dragon:

Come on now.... Even Stallman admits the BSD license is a "Free Software" license. The questions of licenses that are "Open Source" but not "Free Software" tend to be seldom-used licenses like those of Larry Rosen, but there don't seem to be many sound arguments for excluding these. The closes thing I have seen to a sound argument for considering an open source software license to be non-Free was the Qmail license, and the differences there were more technical than substantial.

BTW, in the irony department, the Debian Project consideres some GNU licenses "non-Free" but open. The GFDL is a good example.....
3.31.2009 11:37am
Dragon: BSD-licensed software is free software. One should not confuse "Free Software" (an idea due to Stallman) with the GPL (a particular free software license, also due to Stallman).

That said, the main difference between "open source" and "free software" is a matter of philosophy -- it's more of a difference in motivation than in the actual outcomes.
3.31.2009 11:40am
einhverfr (mail) (www):

Stallman's point is AT BEST a question of priorities, and whether one sees Free Software as the best technical means forward (in which case he thinks you shouldn't call if Free Software) or whether you think Open Source is about Freedom (in which case he thinks you shouldn't call it open source). Functionally they are the same.

The problem is that FOSS derives its technical and economic benefits from the freedom components. These are not separate by any and they cannot exist to the exclusion of the other and still remain effective. At most this is an argument about marketing messages, not really about substantive issues of software development.
3.31.2009 11:41am
einhverfr (mail) (www):
Just a note on OSI's Open Source Definition and its relationship to the FSF's Free Software Definition.

The OSI OSD is almost copied word-for-word from the Debian Free Software Guidelines, which were an attempt to provide meaningful standards for whether a work should be considered "Free Software" by the Debian project. These guidelines ended up being MORE STRICT than the FSF's which caused a conflict over whether RMS's GNU Free Documentation License was Free enough (Debian said "no" and RMS asked Debian community members who did not back him up, such as Thomas Bushnell, the main HURD architect, to resign from GNU projects).

The OSI has a different interpretation of the guidelines than the Debian project, but these terms have been intertwined since the beginning.
3.31.2009 11:45am
anomdebus (mail):
Bill Poser,
I am not sure if it was just inartfully worded, but BSD licensed code that is incorporated into a "baby chipper shredder"* is still free. You can download that same software yourself. The question is, "having used it in a software project, should it 'infect' the project with its license?".

Also, I hope that Stormy Dragon was being sarcasic: note the use of quotes and the inherent contradiction between them and the second para.

* - sorry for the inside OpenBSD joke
3.31.2009 11:58am
einhverfr (mail) (www):

Confusing "Copyleft" with "Free Software" is a common mistake though....
3.31.2009 12:05pm
einhverfr (mail) (www):
One more point about DRM and Free Software.

If RMS had succeeded in getting the anti-DRM wording into the GPL v3 (it was dropped in favor of wording that one was required to release any keys required to make the software fully functional), I think it would have been rejected by Debian as "non-Free." They have done this before with other FSF licenses, and there was a remarkable amount of pushback regarding the GPL v3 and questions whether it was Free enough.
3.31.2009 12:37pm
einhverfr (mail) (www):

Your point on the DMCA Access Control circumvention provision would only take effect if David, say, uses ROT-13 to protect the contents from unauthorized access.

Which brings me back to Prof. Kerr's Lex Luthor Hypothetical and I wonder even if Lex wouldn't have a 4A exclusionary claim, if he could still sue the Metropolis Police for DMCA violations ;-)
3.31.2009 1:39pm
Gabriel McCall (mail):
I understand the philosophical stance RMS is arguing, but I disagree with it. If I build a cage, put a comfy bed and an X-box in it, and tell you that you may enter or leave the cage as you like, I have not thereby reduced your freedom. Nobody is forcing anyone to use proprietary software, and if you believe that using proprietary software serves your interests better than doing without then the rational and free choice is to use it.

Certainly, free software proponents are free not to use proprietary software if they prefer so, but it's not the ethical issue they make it out to be. All kinds of transactions include restrictions on the customer's use- from software purchases to hotel room rentals to real estate acquisitions- but that's not an imposition on anyone's freedom unless they are compelled or coerced to agree to the transaction.
3.31.2009 2:53pm
Also, could you please replace "Richard Stallman" with "someone who isn't completely crazy" in the software community?

3.31.2009 2:56pm
In-house IP Counsel:

Also, could you please replace "Richard Stallman" with "someone who isn't completely crazy" in the software community?


As someone who deals with the intended (and onerous) consequences of newer GPL versions, I wholeheartedly second the above. :-)
3.31.2009 3:27pm
einhverfr (mail) (www):

As someone who deals with the intended (and onerous) consequences of newer GPL versions, I wholeheartedly second the above. :-)

Why I won't touch the GPL v3.

Funnily, I have posed the following question to both Richard Fontana (from the SFLC) and Eben Moglen (from the FSF):

Under the GPL v3, does a license have to be reducible to the GPL v3 to be compatible? If a license, like the BSD-license requires that it be passed on as it is until copyright-worthy elements are added, does this run amok with the additional permissions clause of the GPL v3?

Moglen's viewpoint was that the GPL v3 does, in fact, require license conversion without additional changes, but that this isn't a problem because the BSD license allows such license conversion without additional changes.

Fontana's viewpoint was that the BSD does NOT allow such changes absent copyright-worthy changes, but that the GPL v3 should be read to allow BSD-license compatibility nonetheless. To be fair, Fontana's viewpoint seems to be closer to the consensus viewpoint at the SFLC.

Needless to say, this leaves me a bit nervous about ever using the GPL v3 in any project that may rely on dependencies under any allegedly compatible license.....
3.31.2009 4:25pm
einhverfr (mail) (www):
In short if Fontana is right about the BSD license's requirements but Moglen is right about the GPL v3's requirements, then the licenses are incompatible.
3.31.2009 4:37pm

Other people use "open source" more literally for software whose source is available but which may lack the four freedoms

Much like other people use "free software" more literally for software that can be acquired at no financial cost?

Certainly, one of the definitions of open, from the OED is:

13.a. Exposed to general view or knowledge; existing, performed or carried on without concealment or so that all may see, hear, or take cognizance;

However, another is:

14. Not confined or limited to a few, generally accessible or available; that may be used, shared, or competed for without restriction.

To declare that 13a is a "more literal" definition than 14 is nonsense.
3.31.2009 4:59pm
einhverfr (mail) (www):
BTW, I agree with Fontana's interpretations of both licenses, but feel that the disagreement opens up directions in liability (both legal and otherwise) which I just don't want to get into.
3.31.2009 5:33pm
Alan Crowe (mail) (www):
Stormy Dragon is propagating a common but bizarre misunderstanding of the GNU General Public License

For instance, the GPL requires that if I take your software and modify it, I must make the modification available to the public. That makes me less free as I now have an obligation to provide services to various other people.

I would love to know where this idea came from and how it has taken hold. One of the ideas behind the GPL is that people can experiment with the software, making changes to the source code, re-compiling, running the new version. Do you have to make the modification available to the public, uploading every experimental version to a public FTP server? No, of course not. How could that possibly work? The idea is bonkers.

Do you have to make your modified version available at all. No.

GPL code does carry a burden that BSD-style licenses do not impose. You can sell or give away your modified versions of GPL code, but most business models depend on unbundling the various rights. In most business models you give the purchasers the right to run the code but not to distribute copies. If some-one else sees their friend running the code under the usual unbundling of the copyrights their friend cannot legally give away a copy. The GPL forbids this unbundling. If you distribute your modified version of a program covered by the GPL the recipients also have the right to distribute it, either what you gave them, or their own, modified version.

Notice that one of the ideals behind the GPL is that as software is passed from person to person the opportunity to change the source code and recompile is never lost. That is an important part of the ideal. What do you do if the software almost suits your needs? You make a small change. If you do not have the source code you cannot do this, so you either write your own code from scratch or you make do. Neither meets the ideal of a community sharing and improving programs. The GPL could have been written to avoid this problem by forbidding object-only distribution.That would be awkward with embedded systems. In fact the GPL seeks to offer the a route to object-only distribution that still preserves the recipients ability to modify the source and recompile it if they really want to. This is the intent of section 3b

b) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or,

It is important to understand just how technical 3b really is. If you are distributing software you would either distribute the source code, leaving the user to compile it, or you would include both source and an executable, enabling some users to run the software immediately without compilation. That is section 1 or 2. 3b does not apply.

If you are distributing a gadget, you would ordinarily include a CD-ROM with instructions and you would go down the 3a route of including the source code on the CD-ROM. Stallman and the Free Software Foundation encourage people who redistribute free software to charge as much as they wish or can. Again 3b does not apply.

3b is there as an extra option for those gadget makers for whom it is really awkward to include a CD-ROM with the source. Is it the source of Stormy Dragon's misunderstanding?
3.31.2009 5:42pm
Fat Man (mail) (www):
So RMS, hows that HERD coming along?
3.31.2009 5:52pm
gattsuru (mail) (www):
"Open" and "Free" are synonymous here. "Free" means you can do what you want. "Open" means there aren't restrictions on what you can do with it. That sounds pretty similar to me.

Similar -- there's a movement for Free and Open Source Software for a reason -- but not the quite same.

Most software that is either free or open source is both, but not all of them are. Early versions of the Apple Public Source License, for example, was open source (the source could be easily modified and shared) but did not meet the free software movement criteria (private modifications were limited). All modern licenses that match one match the other, but much software will emphasize one over the other.

The difference is like that between Libertarians and Objectivists; lots of practical similarities, but likely to chew the other's arms off over the ideological distinctions. The Open Source Initiative thinks that open source software can beat proprietary software by being better; the Free Software Movement thinks free software can beat proprietary software because they think intellectual property should not exist.

It might seem pedantic, but that's because it is. Stallman's the sort of person who won't take an interview with someone that uses the phrase "intellectual property". That makes it no less significant a difference.
3.31.2009 6:41pm
einhverfr (mail) (www):
BTW, one point discussing issues with seeing the FSF and GNU as champions of software freedom was found ina blog entry I wrote some time ago.

If Stallman suggests that software freedom is like his view of freedom of speech, I want NOTHING to do with his views.
3.31.2009 6:44pm
einhverfr (mail) (www):

I would agree with you if the FSF was looked to on the ground as the primary arbiter of what is "Free Software." However this is not the case. Individual developers are FAR more concerned with what the Debian community thinks is Free Software because that directly impacts distribution.

For example, Debian considers Qmail to be Free Software, while the FSF does not. Debian considers the GNU Emacs Manual to be Non-Free while the FSF considers it to be Free (the issue is forced advocacy in the manual).

In short, on the ground, Debian's opinions of what is Free Software are far more important than RMS's and they are far more consistent too.
3.31.2009 6:48pm
fishbane (mail):
If RMS had succeeded in getting the anti-DRM wording into the GPL v3 (it was dropped in favor of wording that one was required to release any keys required to make the software fully functional), I think it would have been rejected by Debian as "non-Free."

Debian is a wonderful stop on the occasional goofiness that emanates. I love the fact that they are more dedicated to functionally free than the FSF at times. Disclosure: I have written code that is in the Debian code base.

Also, could you please replace "Richard Stallman" with "someone who isn't completely crazy" in the software community?

See, some of us consider that a feature, not a bug. After all, you don't need to deal with Stallman or any GPLed code at all. Whatever is the problem?

I say that as a critic of the GPL. I don't like it, for various reasons mostly stated above. That said, I don't see anyone calling Stallman crazy also calling, say, Microsoft's IP attorneys insane. They are far more restrictive in what they will give.

So, for those who think Stallman is crazy for enabling people to release code on a particular set of terms, I wonder why they just don't go deal with some codebase that is more restrictive, more to their liking, or whatever. That's capitalism, right?
3.31.2009 6:51pm
einhverfr (mail) (www):
Correction: Qmail has been donated into the public domain more recently. That should say, before that point, Debian considered its license to be consistent with software freedom while the FSF did not.
3.31.2009 6:52pm
einhverfr (mail) (www):

So, for those who think Stallman is crazy for enabling people to release code on a particular set of terms, I wonder why they just don't go deal with some codebase that is more restrictive, more to their liking, or whatever. That's capitalism, right?

I like the GPL v2, and I like the BSD licenses. The "4 Freedoms" is a useful framework too. However, RMS is still insane.

I say this because he advocates "Free Software" which he likens to "Free speech" but acts in a way consistently against an idea that speech should be free. The forced advocacy in the EMACS manual, where the GNU Manifesto is included as an invariant section ("you can change the technical docs, but you can't change the political propaganda").

Don't get me started on Microsoft execs either. There are some good folks at Microsoft, and there are some really bad apples too.

Also I have written a lot of code that is on Alioth, and Debian concerns regarding software freedom have always seemed to me to be more important than the FSF's.
3.31.2009 7:08pm
Stormy Dragon (mail) (www):
LEt me give a mor concrete example:

Suppose A Corp. writes an open source package called WidgetSoft and releases it under the GPL. B Corp. modifies the source to add new features (which they name SuperWidgetSoft) and sell it to Joe User.

Under the terms of the GPL, if Joe User requests the source code, B Corp must provide it. Which means Joe User can then modify it themselves, hire someone else to modify it for them, etc.

Suppose A Corp. had used the BSD license instead. Now B Corp. would be required to tell Joe that A Corp. made the original product, but not to release the source code to their modifications.

In the first case, freedom is maximized for Joe User, as he can get the program modified anyway he wants without being dependent on B Corp. In the second case, freedom is maximized for B Corp. who can do whatever they want to do with the code.
3.31.2009 7:16pm
einhverfr (mail) (www):
Stormy Dragon:

That still doesn't mean the BSD license isn't free software. Even RMS agrees that the BSD License is a free software license.

You are confusing "copyleft" with "free software."
3.31.2009 7:26pm
einhverfr (mail) (www):
Re-reading my piece (linked above in the "blog entry" link), I wonder if my opinion piece was the source of RMS's comment about calling his work "Republican." Basically I had expressed concern that Stallman was falling into the same trap as the Bush Administration that freedoms needed to be substantially abridged perpetually in order to be nominally preserved.

BTW, the main issue (and why I would still characterize RMS as "insane") has to do with the value of free speech, not free software. My concerns include the forced advocacy provisions in GNU manuals, the willingness to dismiss GNU project maintainers for speaking out against such practices, and the insistence that forced advocacy is the only way forward. This sort of thing really should raise the following question in the mind of any neutral observer:

"What sort of freedom of speech does Mr Stallman believe in? And what sort of freedom of speech is to be associated with his vision of Free Software?" I don't think the answer is very palatable to me.
3.31.2009 8:32pm
Jim at FSU (mail):
So long as the user controls hardware necessary to consume the product, DRM will fail. Hardware dongles have been tried before but they are no more successful.

Anything that asks permission can be modified to not ask permission or the permission granting authority can be spoofed in a variety of ways.

Anything that decrypts media with secret keys can either be cracked, or it can be modified to output the media without the encryption. Don't forget that there are layers of drivers and hardware that are below the reach of the DRM. You either need a completely sealed and closed source hardware and operating system (the Xbox is an attempt at this) or you are going to fail. And even then there will be modders who figure out ways to defeat it. Xbox is all about convincing console gamers to stay connected to the internet so MSFT can push updates, ban modders and enforce DRM.

Putting the permission granting authority on a network doesn't make it any more difficult to crack the DRM. So long as the permission granting authority does nothing more than grant permission, the elaborate dance of asking and receiving permission can be defeated because it doesn't contain the essence of what the consumer is interested in. No matter how unbreakable that handshake is, the end result of it is that the consumer receives some non-encrypted consumable that isn't dependent upon the DRM for its enjoyment. That consumable can be rendered into a non-protected form.

To put it in the form of a simple analogy, DRM is like a spiky shell around a delicious fruit that is designed to be eaten. If you make the spiky shell too effective, the fruit doesn't get eaten. Once you've admitted that you intend the fruit to be consumed by a maximally wide audience, it's obviously only a matter of time before someone copies the fruit before eating it.

The only way around this is to ensure that a significant chunk of the experience at any given time takes place on a piece of hardware you control. MMORPGs are a great example of this. Ditto the later games in the quake series, the BF games, etc.
3.31.2009 9:45pm
einhverfr (mail) (www):
Jim at FSU:

My proposed Free Software DRM would use a hardware subsystem to handle the encryption/decryption (similar to but different from what Stallman calls "Treacherous Computing"). It would be possible to break it, of course, but more expensive than a software solution alone would be (I think this would involve modchips).

However, suppose you got rid of the modchip possibility... I would assume someone could STILL make compatible hardware with the exception that the ACL-enforcement hardware could be redesigned to always grant access in certain cases. So, yes, DRM can always be broken. It is just a question of the cash outlay required to do it (reverse engineering and manufacturing hardware is a lot more complex than it is for software). Such a system though given proper peer review would be more robust than closed, non-Free versions, quite paradoxically.
3.31.2009 10:48pm
einhverfr (mail) (www):

To put it in the form of a simple analogy, DRM is like a spiky shell around a delicious fruit that is designed to be eaten. If you make the spiky shell too effective, the fruit doesn't get eaten. Once you've admitted that you intend the fruit to be consumed by a maximally wide audience, it's obviously only a matter of time before someone copies the fruit before eating it.

I appreciate the Durien analogy....
3.31.2009 10:51pm
Eric S. Raymond (mail) (www):
I am the person who promulgated the term "open source", and the senior founder of the OSI. As it happens, I'm also a semi-regular Volokh Conspiracy reader.

Alas, there is a fair amount of misinformation in this thread. Beginning with the misuse of "open source" to mean code with source that is available for inspection but not freely redistributable and modifiable; this is incorrect, and everyone who self-describes as an "open source" or "free software" programmer knows it's incorrect. There is no universal term for accessible/non-redistributable/non-modifiable code, but I like to use "source under glass".

"Open source" means code with a license that complies with the Open Source Definition (OSD). If you try launching a source-under-glass project on any of the community project-hosting sites such as SourceForge, Berlios, Alioth, gna, or Savannah, they will reject it. Even Microsoft -- which is, to put it mildly, no friend of open source and would love to see the term neutered-- recognizes the OSD as authoritative, having submitted licenses to OSI for approval.

As to whether "open source" and "free software" are synonymous: If you're talking about software, the answer is "yes, for all practical purposes"; ever since Apple revised the APSL ten years ago, the exceptions have been minor and technical, involving licenses that are very little used.

Furthermore, there is no boundary in the developer community. No "open source" advocate refuses to work on "free software" projects, or vice-versa. RMS insists that the free software community is separate unto itself, but the actual behavior of hackers falsifies this claim.

If you're talking developer philosophy, the difference is mainly one of marketing - what kind of arguments you use to evangelize to people who are not yet part of the community. "Free software" advocates tend to follow RMS's lead and argue in a prescriptive, moralist vein. "Open source" advocates tend to follow my lead in arguing in a consequentialist, pragmatic way.

However, the situation is not quite as symmetrical as that might seem to imply. RMS claims (rightly) to be the founder and sole ideologue of "free software". I make no corresponding claim; I consider myself part of a hacker tradition of open source that long predates "free software", and only one of a collegium of leaders, theorists, and culture heroes which, in fact, *includes* RMS. I neither have nor want the normative authority over that larger tradition that he does over his faction, and in fact have spent a considerable amount of energy *avoiding* becoming the larger movement's "indispensible man".

RMS likes to maintain that there is an "open source" camp opposed to his "free software" movement. I think it is more accurate to describe his "free software" movement as a purist faction within the larger open-source community. This description better covers the actual working behavior of the people who self-describe with these labels.

In conclusion, I will note that the "open source"/"free software" distinction, to the extent it's actually meaningful at all, matters a great deal more to the "free software" advocates than it does to the "open source" advocates. People outside the community may safely write it off as the standard sort of zealot-vs.-pragmatist hoo-hah that you see in reform movements of all kinds; as usual with such things, it is under most circumstances a dispute that can safely be ignored by everyone else.
4.1.2009 2:48am
David Schwartz (mail):
I fully agree with what ESR said. I'd also add that the "open source" / "free software" distinction is also to some extent just different emphasis.

If you're an evangelist, you may pick the term you expect to get a more positive response. If you're an opponent, you may pick the term you think can get a more negative response. For example, it's hard to give "open source" a hippie or communist slant, but "free software" seems a bit easier.

That's why the argument that there is some technical way that you can have something "open source" but not "free software" is misguided. It's like saying you can have something that's "pro choice" that has nothing to do with abortion and therefore the "pro choice" movement is not the same as the "legalized abortion" movement.

Sure. Supporting school vouchers can accurately be described as "pro choice" too. but that's not how anyone in the "pro choice" movement uses the term "pro choice" and nobody who knows anything about the movement confuses these things except perhaps deliberately to score debating points.
4.1.2009 4:55am
Jim at FSU (mail):

That's a good point about DRM success being heavily influenced by market forces. But I feel that market forces create far more severe barriers for DRM than for crackers. Ultimately I feel that the success of any copyrighted work is going to be determined by the ease with which consumers can acquire it compared to the free, cracked versions, which are going to exist for the foreseeable future.

Anyway, the market obviously creates barriers to DRM. The more onerous the technology is for pirates to overcome, the more onerous it will be for consumers to use in ways they have become accustomed to. In a market where unencumbered media is readily available (either from legit sources or other), there are definite limits to what level of DRM will be tolerated. Even if you manage to sneak something through in hardware and make it successful, you're just increasing the payoff- encouraging someone to crack the DRM.

Valve's Steam was widely seen as a an unbreakable copy protection scheme organized around online games. And it does have many strengths in terms of preventing cheating and piracy of online-only games. Yet its success and reputation has lead to its use for distributing games with significant offline components, which has led to games becoming easier to crack- you've now got a cracking technique that works for every game released on Steam.
4.1.2009 9:48am
einhverfr (mail) (www):
Jim at FSU:

I completely agree with your points. The purpose of my hypothetical was to show that it IS possible to extend, in all meaningful senses, the concept of software freedom to DRM without making it fundamentally MORE vulnerable. I agree that DRM will never be viable in the market and that there are fundamental problems with assuming it will.

As I say, the point of the exercise was not to show how well such a DRM would work, but simply to show it is possible to design a DRM, meeting all reasonable standards of Free Software, which is not significantly undermined by the freedoms that Stallman articulated. Whether one SHOULD build such a platform is an entirely different question, and whether it is viable in the marketplace is yet another one.

Personally, I think that businesses which treat their customers as criminals have bigger problems, so we are probably agreeing in more than we disagree.
4.1.2009 11:26am
Soronel Haetir (mail):
DRM not viable in the marketplace? I find that very implausible given that it already appears to have succeeded. The cost of non-compliance appears to be higher than the cost of DRM for most users. Just because a fringe breaks it doesn't negate the fact that most people do pay.

Note that file sharing does not strike me as a DRM issue for
Services like iTunes show that people are willing to put up with DRM in exchange for media access.
4.1.2009 11:39am
einhverfr (mail) (www):
Soronel Haetir:

Yet cracks in that viability are already showing up. There are a number of DRM-Free music download services which legally license the music they sell, for example.
4.1.2009 1:33pm
Russ Nelson (mail) (www):
This posting is written in ROT-26. If you attempt to decrypt it, I will sue your sorry ass off for violating the DMCA.

Let's all sing it now: D ... M ... C ... A, it's fun to violate the DMCA-a.
4.1.2009 3:09pm
hymie (mail):
The FSF's web site has the GNU Emacs manuals, and I do not see where those manuals designate any section of themselves as "Invariant".
4.1.2009 5:53pm
einhverfr (mail) (www):

For some reason the copyright statement on the online version is missing. See the PDF version for a list of invariant sections. Apparently this is because the preface was not included on the web site version.
4.1.2009 6:47pm
einhverfr (mail) (www):
I wonder if the lack of a copyright notice on the HTML version would be sufficient to essentially licence the whole of the HTML documentation under the GFDL without invariant sections though.....
4.1.2009 7:17pm

Post as: [Register] [Log In]

Remember info?

If you have a comment about spelling, typos, or format errors, please e-mail the poster directly rather than posting a comment.

Comment Policy: We reserve the right to edit or delete comments, and in extreme cases to ban commenters, at our discretion. Comments must be relevant and civil (and, especially, free of name-calling). We think of comment threads like dinner parties at our homes. If you make the party unpleasant for us or for others, we'd rather you went elsewhere. We're happy to see a wide range of viewpoints, but we want all of them to be expressed as politely as possible.

We realize that such a comment policy can never be evenly enforced, because we can't possibly monitor every comment equally well. Hundreds of comments are posted every day here, and we don't read them all. Those we read, we read with different degrees of attention, and in different moods. We try to be fair, but we make no promises.

And remember, it's a big Internet. If you think we were mistaken in removing your post (or, in extreme cases, in removing you) -- or if you prefer a more free-for-all approach -- there are surely plenty of ways you can still get your views out.