A New Model for Software Innovation

An anonymous reader writes "In this whitepaper published at LinuxDevices.com, Matt Asay (former Linux naysayer-turned-disciple) analyzes the GPL, picking apart what it means (and does not mean) for users, and whether it is enforceable. Assay also details how its terms inhibit and foster innovation, and why we should care. In this next generation of software, those who understand 'copyleft' licenses like the GPL will have the upper-hand, and will be best positioned to take on closed-source shops like Microsoft. Assay wrote this paper while attending Stanford Law School, where he studied the the GNU General Public License under Professor Larry Lessig." A thoughtful piece that answers - as well as they can be answered - a lot of the questions about the GPL that we get for Ask Slashdot, as well as examining the economics of it. Good reading for anyone developing or selling software.

Comments

Innovation

[bytesmythe]

From the article link (not the whitepaper):
"In turn, this most 'anti-IP' of licenses is arguably doing more to foster innovation than patents or copyrights ever have."

Patents and copyrights are about making money, NOT fostering innovation.

Re:Innovation

[Milo Fungus]

Interesting attitude. You need to read up on the history of copyright. It was originally intended "to promote the progress of science and useful arts." This is something that very few people understand, even many lawyers. The interpretation of copyright law has evolved into basically what you're describing, but that was not the original intent.

Re:Innovation

[FreeUser]

Patents and copyrights are about making money, NOT fostering innovation.

Very true, but the excuse for patents and copyrights, and the economic and cultural harm such informational and intellectual monopolies cause, is the mistaken notion that they are intended to, or in fact do, "foster innovation." As we've seen demonstrated very clearly in the software industry, and as we've seen evidence of in other areas of endeavor, these government mandated monopolies do not foster innovation or progress ... so their very excuse for existing, and for granting such privelege to so few, is itself a farce.

Which is something the average joe is only now, slowly, becoming aware of.

Re:Innovation

[Bruce Perens]

After you loose your patent, everyone will be able to use your method and build upon it.

That's the way it's supposed to work, but the reality is that after you lose your patent, it's useless to everyone else - at least where most software patents are concerned. The state of the art will have progressed too far, and there will thus be no real quid-pro-quo for the public.

Bruce

Re:Innovation

[Znork]

From the US constitution:

The Congress shall have power ... to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries;

The original intent of patents and copyright was about fostering innovation.

Not that the corruption of the idea we have today is about that tho.

Re:Innovation

[aardvarkjoe]

Well ... the only problem is that the 'rate of technology progression' would be defined by lawsuits, not by any scientific method. (Imagine some judge, not knowing anything about software, ordering that software patents should last 100 years. You know it would happen.) The biggest problem with 'reforms' is that those that it's supposed to effect can usually manage to stall them forever.

Re:Innovation

[Reality Master 101]

Do you think that is all right? Do you think programming a computer is easier or harder than walking down a runway?

"Which is harder" has nothing to do with anything. The only question is which is more valuable in monetary terms. The model isn't getting paid to walk, the model is getting paid to use her physical assets to promote the fashion designer's product. If the fashion design didn't think it was worth it, she wouldn't get paid. Now how many people have the physical assets to do that, versus how many people can crank out code? That's why she makes $50K and you don't. Supply and demand rules all.

And by the way, if you're going to complain that she's "just" using her beauty that she was lucky enough to be born with, and turn up your nose about how superior you are, be sure and realize that you are just using a particular mental talent that you were lucky enough to be born with.

Re:Innovation

[poot_rootbeer]

Making money and fostering innovation go hand-in-hand.

(I would explain my position, but you didn't bother presenting anything that supported your statement, so why should I?)

Re:Innovation

[cmorriss]

Patents and copyrights are about making money, NOT fostering innovation.

I beg to differ. They are about making money AND fostering innovation. It is exactly the possibility of reaping large profits from a patented product that fosters innovation. The fact is, a vast majority of GPLed software is just a copy of a successful proprietary version. Some are good copies, but copies none the less.

Is Mozilla more innovative than Internet Explorer or Opera?
Is Linux (including the various desktops) more innovative than MacOS X?

I would say they are not. In fact, I can think of MANY proprietary software programs that are truely innovative. The open source area produces few programs that fit this bill. They simply play follow the leader.

Re:Innovation

[_Sprocket_]

It might be worth noting that its not just being born with natural talent. Its dedication to maintaining a certain physical appearance coupled with a limited time period in which that is possible. Maintaining this image for as long as possible requires a lifestyle of strict eating behavior and a constant exercise regime. In short, an image worth $50K is not simple to maintain... a point increasing age and a penchant for fully-sugared Coke has driven home to me personally. :)

Re:Innovation

[bytesmythe]

Here's an excerpt from a talk given by Tim O'Reilly:

"History teaches us that as far as innovation is concerned, open beats proprietary every time. You have only to look at the history of the UNIX operating system to see this effect. Many of the innovations that were incorporated into commercial UNIX systems (as well as many of the foundational technologies for the Internet) were developed in universities as extensions to the original work at Bell Labs. Once AT&T took UNIX commercial, under a restrictive license, that work stopped, and didn't burst into flower again until Linux, a free implementation, took over leadership of UNIX operating system development."

When you close everything off, only enough innovation is done to perpetuate the monopoly, and no one else can share in the innovation to help further it.

Re:Innovation

[Stephen Samuel]

Patents and copyrights are about protection of an individual's ownership and reproduction rights of a produced work.

Bzzzt!

That's close to what they do, but it's not what they're about. Patents and Copyrights are supposed to give inventors/authors a temporary monopoly on their inventions so that they have an incentive to create and distribute them.

The temporary 'ownership' of the ideas and the profits that should go with that are simply supposed to be the vehicle for promoting innovation. It's pretty clear when you give a good read to the text of the first amendment.

Re:Innovation

[Stephen Samuel]

Is Mozilla more innovative than Internet Explorer or Opera?

Internet Explorer and Opera are based on the original NCSA Mosaic browsers -- which was an open source project. NT's TCP/IP Stack is mostly 'embraced and extended(corrupted)' BSD (open source) code.

Without open source, Bill gates wouldn't have ben able to program his way out of a router.

Re:Innovation

[Stephen Samuel]

Is Mozilla more innovative than Internet Explorer or Opera?

Is TCP/IP and the Internet more innovative than xns? or netbios (which came much later)?

Is Sendmail / SMTP more innovative than DEC's messaging system from the early '80s, or IBM's? (damned If I can remember their names).

Is NFS more innovative than RFS?? (actually, I'd say that they were about even -- but NFS won out .. possibly because it was more open than RFS).

______

When you think about it: VHS beat out Beta -- not because it was better but because it was more 'open' (just about anybody could produce a VHS machine, but only Sony could produce Beta).

Similarly, many people ascribed DOS's ascendency over the obviously superior MAC to openness -- the fact that anybody could (and did) build a DOS box, and put whatever they wanted into it.

Openness does drive innovation.
With two products of equivalent inovativeness, the one that is more open tends to drive more collateral innovation -- and is thus most likely to thrive.

Part of the reason for patents and Copyrights was the fear that, without them, big-money own-it-alls would usurp the works of truly innovative (usually small) authors and scientists -- make all of the money off of their work and, this remove any incentive for them to make their work known.

Copyright and Patents have now swung to the other end of the pendulum. They've become and end in, and of, themselves. They've been 'strengthened' to the point where we'll soon be 'celebrating' a Century of Intellectual wilderness -- A century during which substantially nothing has been put into copyright that has subsequently flowed into the public domain -- as envisioned by the writers of the first amendment.

Much of what was created in the early 1900s will be lost into this wilderness. Only the most famous (and, in some cases, the most mundane) works will survive. Much of the rest is caught in a legal limbo. Nobody knows who 'owns' it -- thus, nobody can obtain the right to copy it. By the time it's legal to redistribute it, the originals will be incapable of being duplicated.

Consider that recordings of the technical conversations with the Apollo astronauts were only barely recoverable -- and this after less than 30 years in storage. Imagine what another 100+ years of languishing would do to less famous works -- Those who would like to preserve them risk criminalizing themselves in the process.

Had the original works been protected under current Copyright rules, Disny's Snow White would never have been made. Similarly, most of Shakespeares works are known to have had their stories lifted from other authors.

We build on the works of others. Copyright and Patent laws are useful to the extent to which they lubricate the path from creation to public availability. The moment that these laws block that path, they become repugnant to their moral and constitutional underpinning.

Re:Innovation

[extrasolar]

Maybe they think that doing their simple workouts everyday is "earning their worth". But I don't buy it.

It's more likely they are a supermodel because they don't have an eating disorder than they don't have an eating disorder because they are a supermodel.

Or perhaps they have an eating disorder in the opposite way, not eating enough. But I consider this a sickness, not dedication.

GPL Revision

[paladin_tom]

The paper refers to a proposed revision of the GPL. A link to it is on the page.

This GPL revision seems like a completely new GPL. It's language is much simpler, and devoid of legalese. Does anyone else worry about such a complete re-write being enforcable?

Re:GPL Revision

[Bruce Perens]

I don't like Matt's revision for one point - that it excuses things like dynamic linking as not being covered. Given today's COBRA and so on, any linking at all can be dynamic.

Re:GPL Revision

[Bruce Perens]

Do you worry about the old GPL being enforcable? It's never been challenged in a court of law. It's never been tested to see if it stands up. RMS acknowledges that.

Dear AC,

No, I don't worry at all. If the GPL tends not to be valid, the people who have challenged it are then left with the default in copyright law, which is all rights reserved. Even if they win, they lose. Would a court say "yes, you can use and redistribut GPL code without any obligation to anyone else"? If they did, that interpretation would apply to code under other licenses as well - including that stuff in the boxes from Microsoft. Forget about that happening.

Bruce

Re:GPL Revision

[paladin_tom]

I worry less about the current GPL because GNU/FSF has had a bunch of lawyers look it over. You may notice from the wording that it isn't a "simple" license.

Re:GPL Revision

[selan]

The current GPL is full of legalese because the FSF's lawyers go over it with a fine tooth comb. You're probably right that they wouldn't want to change the language.

But it makes the GPL so much easier to understand, doesn't it? If the FSF doesn't like it, why doesn't the writer submit it to the OSI as a new license? I'd use it.

Re:GPL Revision

[paladin_tom]

From the link:

Note: Matt Asay's paper makes reference to a proposed revision to the GNU GPL. That proposal, by Matt Harris (currently CEO of Lineo), is available here . . . Suggested Revision of the GNU GENERAL PUBLIC LICENSE Version 3.0, May 2001

Re:GPL Revision

[mpawlo]

This is actually a real problem for companies being part of a larger development chain. What happens if someone higher in the chain discontinue operations or loose its right to development? What happens if Linus Torvalds dies and his estate withdraws the license for Linux?

Add to this that the GNU GPL is a very American license. It does not deal with author's rights in a European way which may pose a legal and commercial threat to any company relying on free software.

Regards,

Mikael

Re:GPL Revision

[mpawlo]

I am not sure how US copyright and contract law works in this case, but in Sweden and most of the European countries you can not acquire intellectual property (yeah, I know about Stallman's Wittgenstein take on this term, but I use it anyway) through ignorance of previous rights. That is, should the licensee not be the rightful proprietor of the intellectual property, the licensor has no right to the intellectual property.

Ford and their cars are tangible property and can not be compared to computer programs. In this case we may develop a truck or an airplane of the intellectual property equivalent of a Ford car.

Author's rights are not only a right to attribution, it is also a right to have an impact on distribution. Please bear in mind, author's rights can not be waived.

Maybe I should add that I am a GNU GPL fan, but Stallman does not like me .-)

Regards,

Mikael

Re:GPL Revision

[LMCBoy]

What happens if Linus Torvalds dies and his estate withdraws the license for Linux?

As I understand it, the worst that could happen would be a fork. The codebase at the moment before the license change would still exist, and would still be GPL'd. Guess which fork most people would continue to use and develop...

Re:GPL Revision

[mpawlo]

In order to fork the kernel you must be allowed to do so. In order to be allowed to fork the code you need a license.

From a legal perspective you have to check out whether Finnish or US law would apply. Then you would have to check whether the kernel was a part of the estate or not. Then you would check whether the license could be enforced or not. I guess the estate would have a fair shot at claiming ownership, while the kernel was distributed for free (which adds to the giver's rights).

However, this is most likely only a theoretical problem, while the Torvald relatives seem to be fairly well-aquainted with the concept of Linux .-) Still, the safest thing for Linux developers would be for Linus Torvalds to put the kernel in a foundation, FSF or something else. Maybe this have been done already? Mr Perens?

My point was rather to mention a few European problems that may be interesting to consider, should the GNU GPL be revised. I think there is a European investigation going on, issued by FSF Europe, but I have not got the URL at hand.

Mikael

Re:GPL Revision

[LMCBoy]

Yes, I was assuming that the copyright of the kernel was currently held by Linus, and that it would revert to his estate if he died. However, in the COPYING file at kernel.org, Linus says that the linux kernel is "copyrighted by me and others who actually wrote it", so I guess it isn't so clear-cut.

A different perspective, perhaps?

[DesScorp]

While I welcome any insight or debate into the GPL and CopyLefting, frankly, I'm not getting anything new here. And the fact that he studied under Lawrence Lessig is going to imply some bias on his part, considering Lessig's positions. Frankly, other than speculation and name calling, there's not much real analysis out there on the issue of licensing in open source. What I'd like to see is some expert opposing viewpoints, starting a real debate on the legality of the issue. I've yet to see an article from the other side of the opinion fence regarding the GPL's legality, and this is strange, considering Microsoft's attack against the license, and the fear among many in the Open Source movement that it's unenforcable. How about it, Slashdot? How about some legal opinions from people other than Lessig or his students, pro AND con? That would be an interesting debate.

psych

[sstory]

Whatever the GPL says or omits, being 'best-positioned to take on closed source shops like msft' will not be possible without understanding why people will pay $550 for a word processor and spreadsheet. And dismissing that behavior as just 'lusers being stupid' isn't an answer.
When people will pay $550 for an easy, intuitive word processor that comes with restrictions (Office), rather than a more primitive, ugly, yet free and somewhat as functional word processor (Emacs or other unix WP), there's something begging to be explained. And the licening details are of secondary importance.

Re:psych

[symbolic]

Microsoft decided that they were going into a GUI environment, they devoted a huge amount of resources to the problem coming out with an extremely slick product compared to the Linux desktop that still frequently looks and feels like the product of a hobbyist.

I have to agree. I recognize that the people on the KOffice development team, for example, put a LOT of effort into the KOffice suite, but spit and polish do make a difference. And, there are some apps that, while having immense potential, are crippled by some arcane interface conventions. Dia comes to mind. So, hats off to the open source developers - just remember, that if the goal is to advance the cause of open source software, you have to attract users. In order to attract users, you have to pay attention to issues related involving usability. Simply being able to "use" an app isn't enough - when using it feels like you're riding in a well-engineered car, then you know you're on the right track.

Re:psych

[jonadab]

> rather than a more primitive, ugly, yet free and somewhat
> as functional word processor (Emacs or other unix WP)

Calling Emacs a word processor and complaining that it
isn't Word is like calling your house a car and complaining
that it isn't a [insert favourite brand of expensive car].

People don't use Word as a substitute for Emacs. They use
Notepad as a substitute for Emacs. And if there was ever
anything primitive or ugly, Notepad makes it look advanced
and beautiful by comparison. Using Notepad as a substitute
for Emacs is roughly equivalent to using a wheelbarrow as a
substitute for a HumVee. People can argue that they prefer
the Jeep (vim), and we just shake our heads and say there's
no accounting for taste, but arguing for the wheelbarrow can
only be done out of complete ignorance. All of this is no
reflection on Word -- Word is not the same type of thing.
Word is a word processor. An expensive word processor.
Emacs is not a word processor. I _love_ Emacs, but despite
joking about using it as an operating system, I do not
use it as a substitute for MS Word. (I use OpenOffice
for that.) On the other hand, I would never subject myself
to the agony of attempting to use a word processor (MS or
otherwise) as a substitute for Emacs. I'm not even sure
that's possible; it certainly isn't sane. Not only does
it (Word or any word processor) lack the big high-end
features of Emacs (_complete_ scriptability and automatic
indentation and character insertion according to the mode
determined by the type of file being edited, _complete_
email and usenet facilities, unparalleled extensibility),
but it lacks even the basic features that nearly all modern
text editors have, without which editing is intensely painful
(regular expression search-and-replace and mode-driven syntax
highlighting, for example; it does at least have recordable
macros, but what good are they if you can't record the things
you would do in a real text editor, such as regex searches?).

Besides, Word can't use your ISO standard network coffee
maker to brew coffee :-)

lol

[Quasar1999]

former Linux naysayer-turned-disciple?

That's funny... How come I don't read about former Microsoft naysayer-turned-disciple??? Anyone???

Re:lol

[Anonymous Coward]

because they're too busy playing Sims and Warcraft III, right CmdrTaco?

Sounds great on paper

[Junks Jerzey]

I've read lots of papers like this, going back to Eric Raymond's rants. And you know, it sounds great on paper. It really does. And I love the idea of the GPL. But in practice, we're just not seeing all the innovation that you would expect. The classic examples are always cited--Perl, gcc, Emacs, Apache, the Linux kernel--but for a model that's supposed to be so superior, the success stories are fewer than you'd expect.

One striking example is that the GPLed games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc. You would expect the innovative games to be coming out for Linux: no pressure from marketing, free development tools, big community. But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere.

Re:Sounds great on paper

[Bruce Perens]

The gamers are playing games instead of coding :-)

The lure of Free Software development mostly has to do with useful software that can be developed incrementaly. It doesn't necessarily make much sense for games. The fact that both KDE and GNOME produced great desktops in 4 years is pretty impressive, given what folks have spent on CDE and so-on.

Bruce

Re:Sounds great on paper

[Reality Master 101]

The classic examples are always cited--Perl, gcc, Emacs, Apache, the Linux kernel--but for a model that's supposed to be so superior, the success stories are fewer than you'd expect.

And even those examples aren't really that pure. Perl is (arguably) the work one one loan, mad genius, and is not GPL.

The gcc compiler is by far not the best compiler for any platform. It's sole advantage is that it's portable.

Emacs is, well, Emacs. I'll refrain from commenting because of my personal biases. :)

Apache is not GPL, so it doesn't reflect on the GPL, but it's probably the one application that is truly a success story. The reason, of course is that it filled a need when there was a huge demand for web servers and not many to fill the void. One can even make the argument that Apache represents the best web server, although you find plenty of argument about that in specific cases.

The Linux kernel is certainly successful -- but far from innovative. No one (hopefully) claims that Linux represents even close to the "best" implementation of a Unix kernel. Still, you have to give props to the "usefulness" of Linux.

Re:Sounds great on paper

[kbonin]

That's simply due to money!

A grade "A" title single player game can cost $2-4M to develop. The majority of that figure is salaries for the people involved: Averages might be 2-5 programmers, 5-10 artists, and 3-10 support people, all over 18-24 months, plus overhead.

Comments on game quality aside, most big games (lots of levels, high replayability, lots of art, lots of sound/music) take a heck of alot of money to develop. Someone has to front all that money. And if they think that releasing under the GPL will prevent them from recouping their investment, or have a significant negative impact on the bottom line, then it won't be released under the GPL. Period.

What you have here is a simple incompatability of business models, or at least the perception thereof. Games traditionally have to make a pile of money fast, so investors can make a return, and salaries can be paid for the team to start on the second project. You're looking at $1-2M per team per year before overhead. How can a GPL model provide that?

That's not to say it isn't theoritically possible to release the software under the GPL and sell all the content under another or traditional license, but someone needs to demonstrate that a GPL model can take in that $1-2M per year for a relatively short-lived product like a game before it'll happen.

Re:Sounds great on paper

[haeger]

...games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc.[...]But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere.

Isn't it so that most OpenSource software is constructed to "scratch an itch". You have a need, you write a tool that helps you with that. You see someone who has a similar problem (or think that someone might have) so you give your tool away.
You didn't write the program to make money, you wrote it to help you with a task.

Games on the other hand is pure entertainment. You don't just sit down (most ppl don't anyway) and write a game. If you write a game, you probably want to make some mone off it. Especially if it's a new idea.

.haeger

Re:Sounds great on paper

[Patoski]

One striking example is that the GPLed games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc. You would expect the innovative games to be coming out for Linux: no pressure from marketing, free development tools, big community. But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere.

There's a very easy explanation behind this. Coding is one of the smaller parts (although certainly non trivial) in the effort involved in producing a top notch game. Finding quality artists who grok Open Source and are willing to open their works using an Open Source license is very difficult. This makes creating games like GTA3 and the Sims problematic.

If you've ever gone in search of quality Open Source/Public Domain artwork/music you'll know what I'm talking about. The reason you see so many of these simple games produced by the open source community is that the coder is able to do the art themselves and therefore these are the games that get done.

I would like to note however that free (as in speech and beer) art is getting more common as Linux has matured into a viable gaming platform. At WorldForge we're amassing large collections of music, 3d meshes/textures, sprites and sound fx in our CVS repository to help fix this problem. All of the material in our media modules are either Public Domain or GPL/FDLed (some older stuff is OPL). There is already a great deal of material in our cvs media repositories but much is still left to do before we'll see a vibrant indie gaming scene as you described above.

-Jason

Re:Sounds great on paper

[GMFTatsujin]

The examples you cite - Apache, Perl, etc - are also tools which have been lauded for their flexability under a variety of circumstances. It seems to me that once the major component is in place, innovation becomes a matter of what kind of modules you can tack onto a flexable system to cover a specific need in that system's baliwick. Therefore big success stories can be expected to be few and far between - a single flexable success story can cover a *lot* of ground, and the innovation then defers to figuring the niggly bits of how it applies specifically to you.

As far as games go... well. I'd expect that the game programming going in is really done more in the spirit of "I wonder if I can figure out how they did that, then add my own touches to it", such as a CS student might employ to test their skills. These games have simple rules and a narrow environment. The sparky games like Sims and Grand Theft Auto shot for the moon in creating games with unique characteristics of gameplay and theme. The rules and environments are far more complex - "emulate various personality traits in a world full of objects that affect statistics in different ways, which then play off of the personality traits themselves to produce human-like behaviour... and make it skinnable" is a far cry from "blow up missles falling from the sky" or "push rocks down a dirt tunnel."

Unfortunately, that takes a lot of talented and skilled developers working hard toward a declared, managed goal. The fact that these are ground-breaking ideas makes it all the more difficult and time-consuming. The sheer interactivity that *every object* in the sparky games carry with them is a thing of wonder to me, and I wouldn't expect anybody to pull something like that off in their copious spare time. The kind of game takes a fellowship of like-minded dedicated programmers who can put serious day-to-day time into the project -- and that means money.

Not a programmer, just a guy who tries to think about things. I could be wrong.
GMFTatsujin

Re:Sounds great on paper

[NorthDude]

Makes me think...
I really really beleive that the GPL is a brilliant concept.
But, and sorry for that, I fail to see how I could buil a business on top of it.
Sure, it can work for some type of application, you give the app, sell the support.
But there are other type of application which I don't see could fit in this model.
Games for one. How could you make a profit selling open-source games?
First of all, there is no support to offer.
Second, games are generaly obsolete within months.
I mean, it is not like an open-source server which could evolve years after years.
A game, you release it one time, maybe add to that a couple expension pack, a couple patch, and it's over.
And as beautyfull an idea as it may be, and as much as I would like to share knowledge with each others, I have to eat, pay the bills and build up my future.
All that takes real money!

So, do you people beleive (and have a rational explanation) that the GPL can be applied to EVERY single type of software, or it can only be aimed at some categories of software which one can offer services for?

Re:Sounds great on paper

[pmz]

And CDE is still unusable.

CDE is just fine. It is pretty simple, quite configurable, and very well documented. It is also pretty ugly compared to other desktops, but aesthetics wasn't the main motivation behind its design. It is, simply, a workstation desktop.

Re:Sounds great on paper

[Djinh]

Grade "A" operating systems can cost $100-1000M to develop. The majority of that figure is salaries for the people involved: Averages might be 10-1000 programmers, 10-100 artists and 30-100 support people, all over 18-48 months, plus overhead.

[...]

Oh wait, there's Linux, FreeBSD, etc, etc...

I've got no idea why there are no Great Open Source Games With Spark, but I bet money's got nothing to do with it.

Re:Sounds great on paper

[robson]

That's simply due to money!

All of your points are correct, but I want to take it one step further: What's different about application development? Why is it possible to produce high-quality GPL applications under the same circumstances? In other words, is there anything we can learn from open-source app development and then apply to potential open-source game development?

Here's where I'm coming from: I'm a professional game designer. I use Linux at home for (most) of my computing tasks, and that means I'm benefiting from the work of hundreds (thousands?) of open-source developers. I want to give something back, and I'm intrigued by the possibility of doing a GPL'ed, FREE game that isn't subject to the whims of the marketplace. That's where this thread's parent post starts to kick in -- once you get good, passionate people working on a game without having to worry about sales, suits, and the like, you're likely to get a riskier, more innovative result.

Now... you don't need to tell me how long or how many people it takes to make good games; I know. But I can't help but think that there's a solution somewhere that we're ignoring. Again, what can open-source game developers learn from open-source application developers? Do applications really come together that much faster than games? Is it that there aren't enough artists willing to contribute their time and skills?

(Afterthought: My post assumes that open-source games are being developed in volunteers' spare time; I'm under the impression that most open-source application development works this way. If this isn't the case, that changes the arguments.)

Re:Sounds great on paper

[pmz]

But in practice, we're just not seeing all the innovation that you would expect.

What innovation are you expecting? Are you looking to be swept off your feet by the next handsome OS that you see in the elevator?

Linux and GNOME are a unique twist on the old idea of a UNIX desktop. Linux has matured a great deal over the years to rival old giants like Microsoft, and GNOME has become a very good desktop environment for many people. While they aren't earth-shatteringly innovative, they provide some relief from the stagnant mainstream desktops out there.

Then there are the countless tools that made my life livable over the years: GCC, gnuplot, ghostscript, Emacs, LaTeX, ispell, XFree86, Mozilla, and OpenOffice.org for my desktop, and sendmail, apache, and BIND on the servers. There are many many many more tools that I won't take the time to list out, but that doesn't mean they aren't there doing their part.

Many Open Source software projects are not truly innovative but provide best-of-breed implementations of standard protocols. Nearly all of the standard Internet protocols have been implemented for Linux and the BSDs. Ones like Apache steal the show, and ones like the BSD TCP/IP stack are "borrowed" by smart and successful people like Microsoft.

Perhaps the biggest contribution of Open Source is properly moving things into the "public domain" after no longer being commercially worthwhile or after becoming commodities. The protocol implementations I mentioned certainly fall into this category. For example, I seriously think the idea of a commercial web browser is obselete, and ones like Internet Explorer really serve to hold back progress rather than help it.

Another way that Open Source contributes in ways that companies can't is to provide uncorrupted software. By "corrupted" I mean: integrated DRM, proprietary file formats, proprietary communications protocols, unfair EULAs, and any other blatant company or government-driven attempt to limit what computers can do. The frequent advocacy by Open Source enthusiasts for open file formats is one way to drive innovation that really can be earth-shattering (imagine a world without .doc!).

So, I'm not sure what you're looking for, but I'm pretty convinced that there is more innovation occuring than we can shake a stick at. It isn't always suprising, revolutionary, or even obvious, but to say it isn't happening is to be in denial.

Re:Sounds great on paper

[robson]

Translation: I'm one of the people responsible for the glut of FPS games on the PC.

That's a curious assumption. No, I'm not working on a FPS, and I'm not developing for the PC.

Re:Sounds great on paper

[pmz]

Right, if you ignore the focus problems, the motif problems, the apps that come with it and the fact that sun and HP have both ditched it for gnome.

I've been using CDE for years...I guess I just don't notice these problems. Perhaps they really aren't problems but are just differences relative to other desktops?

I have found CDE be a very useful very no-frills desktop. It doesn't really need to have any eye candy, and it is good enough for me to get done what I need to get done.

Regarding GNOME, I do look forward to the migration to GNOME, but only after they have worked everything out to make GNOME a low-risk transition. I'm willing to be patient about this.

Re:Sounds great on paper

[ChaosDiscordSimple]

I've read lots of papers like this, going back to Eric Raymond's rants. And you know, it sounds great on paper. It really does. And I love the idea of the GPL. But in practice, we're just not seeing all the innovation that you would expect. The classic examples are always cited--Perl, gcc, Emacs, Apache, the Linux kernel--but for a model that's supposed to be so superior, the success stories are fewer than you'd expect.

Fewer than you would expect? If by "success" you mean "market penetration", yes, we have problems. But we also lack dedicated marketing departments. If by success you mean "large, useful, stable, worthwhile", we're doing great. The Wine project is making amazing progress on effectively reimplementing Windows. It's to the point where Microsoft Office runs. Gnumeric isn't perfect, but for 90% of the world's spreadsheet needs, it's great. AbiWord and KWord are lagging a bit, but making great progress. I've been using GnuCash for almost a year and haven't missed MS Money in the slightest. Sure, the Gimp has some issues that make it difficult to use for professional image publishing, but it's more than adaquete for your average person, or someone doing web images. (A friend who does photo editing work for a local newspaper says that the Gimp has everything she needs to replace Photoshop.) Playing music: XMMS. DVDs, DivX:-), and many other video format: Xine. Email: Evolution. Highly standards compliant web browser: Mozilla. Every single one of these is a huge success story to me, they've allowed me to replace Windows as my primary work system without missing anything.

One striking example is that the GPLed games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc. You would expect the innovative games to be coming out for Linux: no pressure from marketing, free development tools, big community. But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere.

There are a number of reasons why Free Software games tend to be more simplistic. The biggest is art, sound, and music. Programmers can whip up simple icons for their word processor, but generating the hundreds of textures, dozens of sounds, and music for a game isn't something we're as good at. And the concept of "Free" hasn't infected the artists and musicians to the extent that it has programmers. (This may be because programmers have a great deal of incentive to share code. It's harder to take another artist's work and tweak it a bit for your own needs.)

That said, Free Software games aren't totally dead, just small. Nethack avoids the art problem entirely and remains one of the most innovative computer hack and slash games available. Freeciv is highly derivative, but has an innovating client-server set up. Tools like SCUMMVM and MAME show a great deal of innovation in reverse engineering and improved visual displays of existing content. Crystal Space is an excellent 3D engine mostly suffering from lack of content created with it.

Re:Sounds great on paper

[kbonin]

Its a tough problem, and I'm in the same boat. I'm a professional game programmer, about 15 years in the industry (SSI, Bethesda, Accolade, others...)

10+ years ago I started an VR engine project that I wanted to eventually place into the public domain - I've even continued to get employment contractual waivers from employers to allow me to do so. Over the last 10 years I have invested hundreds of hours trying to cultivate relationships with others who could contribute.

Here's the problems I keep running into:

1a. High-end games are VERY hard, and most people who have the requisite skillset already work in games.
1b. Most people who work in games have signed contracts precluding them from working on other games, including open source peojects.
1c. People who could otherwise help are kept so busy with the sweatshop hours most game houses run on that they have little to no time available to contribute!

2. If work on a game project is done on the side, it becomes very difficult to keep up with the state of the art. I've had to throw away huge sections of my codebase (especially 3d) and begin again to keep it competitive! It gets depressing when you know the code your working on at work is now n>1 years behind the engine you're building on the side. Low tech starting points do not inspire people to join...

3. If development work on an open source engine is done in the open, it becomes trivial for other developers to "steal your thunder" by encorporating your best ideas into their titles, which due to funding, ship before you! It also becomes possible for certain parties to use their access to your work to begin patent abuse! (ala Rambus, etc.) Personal example: In the few weeks after I announced on my website that I was going to publish my VR scalability whitepapers, I had a huge traffic spike from the research.microsoft.com domain. I changed my mind...

The other problem is related to tools. Games generally require significant content. Good content generally requires good tools, which generally means expensive. 3ds Max and Maya are very expensive. Artists who have purchased these tools for personal use NEED to make enough $ to recover their investment, not to mention keep food on the table. Audio tools are nearly as expensive...

I think it could happen, but only if a sufficient critical mass was acheived. I haven't seen it happen yet...

Re:Sounds great on paper

[robson]

Hey kbonin.

2. If work on a game project is done on the side, it becomes very difficult to keep up with the state of the art. I've had to throw away huge sections of my codebase (especially 3d) and begin again to keep it competitive! It gets depressing when you know the code your working on at work is now n>1 years behind the engine you're building on the side. Low tech starting points do not inspire people to join...

Okay. I'm glad you brought this point up. Here's the thing: As we know, games aren't like movies -- game technology is constantly evolving, and this makes developers' jobs harder than it needs to be. So... I'm almost tempted to say that we need to resign ourselves, at least partially, to last-generation technology. Something's got to give, and that seems like the best candidate.

This takes a lot of pressure off developers and enables us to focus on gameplay. You recruit volunteers not by displaying cutting-edge technology, but by clearly specifying the project's vision. You attract volunteers who have similar goals. This forces a different question, though -- as an engineer, what are your motivations when you approach an open-source game project?

As a designer, I'm excited by the possibility of making the sort of game that the free market couldn't support -- either because it doesn't have a cute character, or because its themes are too "mature" (not blood-and-tits mature, but grown-up mature), or whatever. Thus, my suggestion above to resign a project to behind-the-curve technology suits my needs, design needs.

However... it occurs to me that an engineer might be attracted to an open-source project (on the side) because it gives him/her a chance to do sexy research & tasks that their "day game" doesn't permit. Thus, resigning a project to old tech may render it undesirable for an engineer to volunteer for.

So where does this leave us? Reduce the scope of the project. Keep It Simple. Don't try to compete with commercial games; just do a small set of things well. Try to build the design in such a way that unsexy art, design, and programming tasks are kept to a minimum.

I think for that, for open-source/free game development to work, it may be necessary to think about it differently than we think about commercial game development.

Obviously, I could easily be wrong -- it's possible that if you remove most of the unsexy tasks from a game's schedule, you end up with half a game :) ("How can we have a game with no UI?!?")

3. If development work on an open source engine is done in the open, it becomes trivial for other developers to "steal your thunder" by encorporating your best ideas into their titles, which due to funding, ship before you!

I don't know how kosher this is in GPL terms, but are you "allowed" to release the source only after a project has reached a certain level of completion? Say, beta?

The other problem is related to tools. Games generally require significant content. Good content generally requires good tools, which generally means expensive.

I don't really have an answer to this. There are free tools out there, but as you know, they don't generally compare to pro tools.

Re:Sounds great on paper

[Eric Damron]

"One striking example is that the GPL'd games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc. You would expect the innovative games to be coming out for Linux: no pressure from marketing, free development tools, big community. But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere."

I run Linux exclusively at home and it is true that popular big name games are few and far between but there are the exceptions. Look at Quake 3. The Linux version has better frame rates than the Windows version if you are using NVIDIA's OpenGL drivers.

Linux has been traveling a long road; uphill and against the wind. Microsoft has done everything to try to stop Linux but it can't be stopped. It can't be bought, bribed, threatened into submit ion, outlawed, or destroyed by FUD. In short, none of Microsoft's usual tools work against Linux.

Does this mean that Linux will be the most popular OS immediately?

No. Linux needs a critical mass of users before proprietary software companies will consider it a viable platform on which to develop products. It also needs a critical mass of products to draw in users. This is a classic 'Catch 22.'

I believe that we are approaching the critical mass of software. The visibility that companies like IBM, SUN, HP, Compaq and others are giving Linux is a big help. The fact that Linux is a great operating system with lots of great free software means nothing if no one hears about it.

So, the growth has been slow but we are nearing those critical masses that will rocket Linux into the public consciousness.

Re:Sounds great on paper

[BadmanX]

It's very simple.

If you use software, but sell products or services, Free Software is your best friend. It reduces your costs while not undercutting what you're really selling.

If you sell software, Free Software is your worst enemy. It takes away the only thing you have to sell.

Game developers make money by selling software.

Re:Sounds great on paper

[flonker]

Nethack.

Not quite GPL, but still, it is open source. And it still has managed to steal quite a few man-months of work from me.

Re:Sounds great on paper

[fferreres]

1) Produce something that can help other make money and you use the GPL (or closed source).
2) Produce something that can't make money for anyone (games, and most home user programs) and you can't make money except using a closed source system.

So under 2) you end up with really low quality apps and products while under 1) you have great or even better-than-Closed-Source solutions (in some cases).

It all can be summed up with: if you can't sell X directly or indirectly, you are left with no X at all. At least, this is was has been hapening up to now.

Innovation and interest

[Bruce Perens]

There's a quote of Lincoln on one side of the patent office building in Washington DC. It says something about the patent system coupling innovation with interest, which I guess means providing a reward for innovation. That's where the money comes in.

But of course there are any number of situations where that reward is not the optimal means of promoting innovation, or isn't even appropriate - one example is the case of public-funded research resulting in private patents, another would be when the monopoly grants not merely methods but entire categories of business, as with today's broadly-interpreted business method patents.

Bruce

Re:Innovation and interest

[SanLouBlues]

I believe you mean this.

The quote: "The patent system added the fuel of interest to the fire of genius."

Re:Innovation and interest

[Bruce Perens]

Yes, but in the times he wrote (said) it, giving things free was unheard of.

Yes, the term "indian giver" was still popular when I was a child. People obviously did not understand a gift exchange.

You know, what you refer to as socialism is almost a side-effect. It's just that the money gets in the way of our collaboration. So, I prefer to refer to it as a commons within an overall capitalistic system.

Thanks

Bruce

A new model for software innovation

[jukal]

I see this as the perfect chance to plug openchallenge and maybe get good comments on how to improve the concept.

Mace Moneta summarized the idea well:
" get those that have the skills needed to tackle a problem together with those that have the problem, and solve it as Open Source. The OpenChallenge web site lets you submit problems; those looking for a project to work on can browse the list of requests for something that they find interesting. "

Stallman must be obsessive-compulsive

[John Harrison]

Either that or a robot. On page 17 he goes off on the "GNU/Linux" thing without provocation. Unless you consider the word "Linux" provocation. Anyhow, I just had to laugh. I guess I should have seen it coming but the FSF guy replied to to question in such a normal way and RMS had been replying so normally that it caught me off-guard.

Re:Stallman must be obsessive-compulsive

[LMCBoy]

Stallman wants fame.

No, Stallman wants his vision to come true. Read "Free as in Freedom". His obsessive behavior is much more readily explained by his strict adherence to his principles, and a near-absolute unwillingness to compromise.

It must be "GNU/Linux", not because he wants to associate himself with the success of Linux, but because he wants to make sure that Linux remains firmly in the Free Software camp, and does not drift off into some less-Free OSS realm.

"There is no OS but GNU, and Linux is its kernel"

Re:Stallman must be obsessive-compulsive

[BadmanX]

Not only that, but in this case, "Linux" was the correct term, since the guy was referring just to the kernel. Didn't stop Stallman, though.

esay

[mobiGeek]

Assay also details how its terms ...

It's a great esay written by Assay.

Re:esay

[Prior Restraint]

It's a great esay...

It's an even better essay.

proposed revision of the GPL

[Dr. Awktagon]

An interesting streamlining of the GPL. However he takes out all the "social statements" which may or may not be a good thing.

A few comments:

You can use GNU GPL Software for any purpose. You can modify, copy and distribute GNU GPL Software, including derivatives, for commercial or non-commercial purposes.

In return, you must agree:

No no, you can use the Software for any purpose. Period. If you re-distribute the software, THEN you must agree to the terms.

The license should emphasize that it only covers "copying, distribution and modification". No one should feeled compelled to accept the GPL, just because they are using the software. This also helps dispel the stupid notion that you don't have a right to use software you buy or download unless someone grants it to you.

This, in my opinion, is one of the greatest thing about the GPL. It lies dormant and only applies to you if you distribute copies. You don't need to read it, you don't need to "click-through" it. You only need to care when distributing copies, which you aren't allowed to do to begin with. In fact, if I were writing the GPL, I'd put that up at the very beginning (after a warranty disclaimer if that's necessary): "you don't need to read this unless distributing copies" or something similar.

if you distribute GNU GPL Software, you include two things: a copy of this license and a ready means of obtaining the source code for the Software.

Define "ready means"..

That if you distribute a derivative of GNU GPL Software, you include a notice explaining the changes you made.

(Nitpick) why isn't this included with the previous requirement? What constitutes a "notice explaining the change"? A diff? A paragraph? Is this important?

The GPL is wordy, partly because it spells out some of this stuff in more detail (for instance, see item #3 in it). The "revised GPL" should also be specific whenever possible.

Anyway, the GPL does need to be streamlined and clarified to a certain extent, and perhaps the FSF can get some ideas from this revision. For instance, cutting out the Preamble of the GPL entirely might be a good idea, and replace it with a link to the FSF philosophies.

Re:proposed revision of the GPL

[naasking]

Unfortunately, it may still be necessary to agree to the GPL even to use software. It frees the developer from liabilities by disclaiming all warranties including fitness for any purpose.

Re:proposed revision of the GPL

[Bruce Perens]

You left out the part about linking. If we are to clear up that part, I'd much prefer to make it clear that some forms of dynamic linking and server-izing (as with CORBA) do create derivative works. I think the key concept of a derivative work in software has nothing to do with linking, and everything to do with the span of the thread of control in a computer program. This can indeed happen across physical system boundaries. The boundary for the copyright should probably be at an API that is explicitly designated as an interface to separate works, such as the kernel system call interface. Not just any API.

Thanks

Bruce

Re:proposed revision of the GPL

[FooBarWidget]

It's only a proposal, not the final version! Don't hammer it, it's not even complete yet.

Re:proposed revision of the GPL

[Zathrus]

The purpose of proposals is to hammer them. If there is no constructive criticism (which the OP certainly gave) then it may as well be final.

Re:proposed revision of the GPL

[Ed Avis]

The licence cannot define what creates a derivative work. That is given by the copyright law and court decisions. Your arguments about where copyright's boundary should lie make sense, but they don't belong in the GPL's text.

Of course the GPL could waive rights: 'even if for some weird reason it turns out that mentioning the name of Richard Stallman makes all future programs you write a derived work of GNU Emacs, we choose not to restrict distributions of such derived works'. But you are talking about strengthening, attempting to claim control over things which the law might not consider to be derived works, and that isn't possible. (Without some unpleasant UCITA/click-through which I'm sure RMS wants to avoid, and probably not even then.)

('IANAL' considered redundant.)

Re:proposed revision of the GPL

[Zathrus]

Well stated.

The proposed revision of the GPL does "fix" some of the murkiness related to dynamic linking -- but does so by implicitly claiming them as derivative works.

I know this is what RMS wants, and perhaps it's what the community as a whole wants, but it's also why most commercial companies won't touch GPL code with a 10' pole.

At the very least an API as you define it should be able to form a firewall between GPL and non-GPL code (note - not necessarily commercial - I may want to distribute my code under BSD license, but it still needs to be untainted by GPL for that to be an option).

Without this provision companies will continue to avoid GPL code for fear of losing core business logic. I code in Unix, use open source tools, and we avoid any GPL libraries even though there's no intent to distribute the code outside our company. It's simply not worth the risk that one day we'll have a component that does need to be distributed and potentially tainting everything else -- which is core business logic.

Re:proposed revision of the GPL

[sacrilicious]

The boundary for the copyright should probably be at an API that is explicitly designated as an interface to separate works.

This thought is good/interesting. Comparing the linking interpretation with the above, I conceive of the scenarios this way:

(a) Linking: if you link to GPL'd code, you must release the entire source base

(b) API Boundary: if you modify code within the formally designated API boundary, you must release your modifications within that boundary

(b) seems more in the spirit of what I'd want, but it does present some problems:

1) what rules are adopted to ensure that the new API boundary remains clearly delineated within the new additional code, with no holes or ambiguities?

2) what if the person modifying the code decides to specify the API boundary at such a shallow depth into their added code that the spirit of the standard is defied, and the released code ends up getting polluted with such intentionally devious boundaries?

.

Re:proposed revision of the GPL

[Bruce Perens]

The licence cannot define what creates a derivative work.

It could make more of an attempt at this if it purported to be a contract, but it does not. However, copyright law is fuzzy enough about what constitutes a derivative work in software - the law says almost nothing about it - that a statement included in the license of what the parties consider to be derivative works might have some force.

Bruce

Re:proposed revision of the GPL

[jpvlsmv]

I'd much prefer to make it clear that some forms of dynamic linking and server-izing (as with CORBA) do create derivative works. I think the key concept of a derivative work in software has nothing to do with linking, and everything to do with the span of the thread of control in a computer program.

However, this moves a major burden to the USER of the software, who is supposed to be free from restrictions on the use of the GPL-licensed software.

By including dynamic linking (or CORBA) into the definition of derived works, the user must be aware of all code that may be imported into the application. A GPL program may be legal to use on one system (a linux system with a GPL'd corba instance) but not on another (a linux system with a GPL-incompatible corba system). Even if these corba implementations are indistinguishable from the application's point of view.

Another example, would installing a non-free libc.a (with the same function prototypes, but not a derived work of GNU libc) on a linux system violate the license of all GPL software on that system? Any GPLd program which (dynamically) links with /usr/lib/libc.a would be combined (in memory at least) with something that can not be distributed under the GPL.

--Joe

Re:proposed revision of the GPL

[Bruce Perens]

it's also why most commercial companies won't touch GPL code with a 10' pole

Both HP and IBM touch the GPL with some simple policies and a review board as their only precautions. They tell their customers about it, too. It happens that I am in the business of selling corporate Open Source policy and practices, so I can say sincerely that the 10-foot thing is ending.

People who want to save their business logic put it in non-GPL applications and don't put GPL code in those applications. It's not such a big deal. Also, if you do end up infringing, your proprietary code isn't "tainted", you have to cure the infringement. Writing out the GPL code and saying "I'm sorry" is an option to cure the infringement. Do you really think any court is going to compel someone to GPL their application?

Bruce

Re:proposed revision of the GPL

[Bruce Perens]

Well, the big question is whether or not someone can arbitrarily add new API boundaries to your GPL code - as in your #2. Perhaps there should be some rules about the purpose of the API? I haven't thought it out, and I see this needs thinking through.

Thanks

Bruce

Re:proposed revision of the GPL

[Bruce Perens]

The question is whether or not a simply warranty disclaimer requires agreement. I'm not sure that Sprecht v. Netscape (on the nature of agreement) is relevant to that. Now, if you want the user to indemnify you, I would think that really does require agreement.

Thanks

Bruce

Re:proposed revision of the GPL

[SN74S181]

It happens that I am in the business of selling corporate Open Source policy and practices,

Somebody I know was recently hassled by a telemarketer trying to sell her aluminum siding for her house.

She told the salesman: "But our house is made of brick!"

The salesman answere: "We can put aluminium siding on a brick building....."

Re:proposed revision of the GPL

[Dwonis]

For instance, cutting out the Preamble of the GPL entirely might be a good idea, and replace it with a link to the FSF philosophies.

No, that would be a bad idea. The preamble is part of the licence, and would be used in the licence's interpretation in court.

Re:proposed revision of the GPL

[spitzak]

None of these are problems, they are allowed under the GPL.

The GPL does not restrict *use* of the software in any way. What it restricts is the *distribution* of software. You can link a piece of GPL code with closed software. So can the person you give your copy to. This is OK, no matter how much MicroSoft tries to claim it is not.

What you are not allowed to do is modify the GPL software so it only works when linked to the closed-source software and then give that modified copy away. You have then violated the rules of the GPL about when you are allowed to give copies away, therefore it reverts back to copyright law and you are violating US (or whatever countrie's) copyright laws.

Re:proposed revision of the GPL

[evbergen]

A thread of control means little if you post another program and wait for its response--it then moves from one program to the other and back.

A more sensible definition seems to me this: if a new work can change the intended behaviour of the original work, then it's a derivative.

In other words, you're always safe interacting with programs that live in other address spaces; by definition you can only interact with interfaces that the program's author decided to make public (or otherwise the program is buggy security-wise).

You're less safe when sharing an address space, eg. when linking statically or dynamically, because you can access every global variable and call every routine, possibly making the original program deviate stronly from its intended behaviour. However, if the original work is written as a library and the additional parts refrain from calling routines or modifying variables that are clearly marked as internal, it would not interfere with the library's intended behaviour, so the combined program would not be a derivative work.

I'm not sure to what extent the concept of 'not modifying intended behaviour' can be sensibly applied to programming languages other than C, such as eg. a complete Lisp system using a single address space.

However, I do feel that interfaces are a fundamental aspect of all computer programming, and that interface boundaries will always be more or less identifiable, regardless of programming model or language.

Indeed, following the practice set by the Linux kernel seems only too sensible now: if you use *public* interfaces, you're not creating a derivative work. Not with RPC mechanisms like CORBA, not even when sharing address spaces.

If you use internal interfaces (or worse), you're creating a possibly unauthorised derivative work. Of course, you're still free to use that, you're only not allowed to redistribute the combination without source or GPL to cover it, giving subsequent users the right and the means to do the same thing you did.

Re:proposed revision of the GPL

[wfrp01]

I think the terms "API boundary" and "linking" could use some clarification. In fact, I don't know that I'd use the word "linking" at all, because of its specific association w/ particular programming languages. It's all about definitions. I think the issue needs to be translated from the language of programmers into legal nomenclature.

Somebody should do something (but not me). ;)

Re:proposed revision of the GPL

[Zathrus]

Either comply or look for another source for your needs

We do. Thus we use Boost, OTL, and other C++ libraries with non-GPL licenses. Realistically it hasn't been an issue for us -- there aren't any libraries that we've wanted to use but were prevented from by the license. If there were, we'd either find an alternative or write it ourselves.

Problem is, a lot of people do write code, hoping for it to be used widely and then GPL it. Yes, LGPL is the right license, but it doesn't get nearly as much press as the GPL and so a lot of people unfamiliar with the particulars of both licenses might not get it.

I'm not asking to subvert the GPL. I'm asking that it be far more clear as to what constitutes a derivative work -- Bruce took a good step towards that IMO.

His next paper...

[return 42]

...will be on the Slashdot effect.

IANAL++

[jon_c]

This guy deserves a new acronym

IANALBISL

I Am Not A Lawyer, But I Study Law.

-Jon

Re:Worldwide monopolies?

[alienmole]

Following the trend set by "oligopoly", where an "o" was stolen from "monopoly" simply to aid in pronunciation, the word you want might be " polyopoly " - or perhaps "poliopoly", but that sounds a bit too much like someone who is infected by multiple polio viruses.

Taking best advantage of IP and software

[MarkWatson]

I have just skimmed his paper (interesting stuff to me started on page 31), so this is tentative: I have been thinking of GPLing my software technology for semantic extraction and natural language processing.

My problem is that I have invested a little under 2 years of work on my NLP tools, and although I only make about $5K a year selling them, that money does help pay the bills and I believe the long term market for my software is excellent.

I have received many great bug fixes and code improvements for my Open Source projects and even more great help with material for my free web books, but these projects have not cost me nearly two years of work.

I guess that the real question is if companies (and perhaps individuals) who use GPLed software to make money in their own closed commercial products are honest about paying a fair license fee to get a non-GPL license to use my work. I do receive many generous donations for my free web books, so I do trust in many people's generousity and fairness.

An alternative to using the GPL is to license software for free non-commercial use and then hope that commercial use users act fairly.

I think that many developers would very much like to have their work freely used by people who make no money off of their work and at the same time collect reasonable payments from anyone making money on one's work. Tough problem - please share any good ideas :-)

In any case, it is tough for me to make a decision here.

-Mark

Re:barf!

[aardvarkjoe]

If anyone posted this garbage here, they'd be moderated to -1 in a hurry.

Are you under the impression that slashdot moderation has anything to do with the quality of a post?

Re:barf!

[selan]

``Free source software has a long history of playing catch-up to proprietary software.''

Just curious: why do you disagree with this statement? As others have noted, the majority of open source software is described as "an open source version of [insert proprietary software name here]". Even GNU was meant to be a free replacement for Unix.

Comment removed

[account_deleted]

Comment removed based on user account deletion

The Microsoft Factor

[Analysis Paralysis]

One item only partly addressed in the document has been the increasingly burdensome requirements of Microsoft licenses (product keys in manuals, product keys stuck to the back of PCs - good idea MS!, online product activation and shotgun EULAs giving MS the "right" to alter your system). In combination with the increased cost of licensing, this is prodding all Windows consumers to re-evaluate their options. Not to mention MS's famed security...

Re:Innovation and patents....

[Milo Fungus]

Case in point: obstetrical forceps were invented by an Englishman in the 1600's. He used his invention to save many mothers and babies from dying at childbirth. The knowledge of his invention was forgotten until the 1800's when someone found his storage trunk. He had jealously guarded his invention, sharing it with only a few family members. If he had been protected by patent law, he would probably have disclosed the details of his invention. This would have resulted in millions of babies and mothers saved from a terrible death.

That's why we have patent laws.

You are looking in the wrong places

[FreeUser]

I've read lots of papers like this, going back to Eric Raymond's rants. And you know, it sounds great on paper. It really does. And I love the idea of the GPL...ut for a model that's supposed to be so superior, the success stories are fewer than you'd expect.

That is because you are looking in the wrong places for your "success stories." You are wondering why some GPLed mapping program hasn't unseated Microsoft's Streetsmart, or why program X under the GPL so resembles commercial program Y (though often it is program Y which mimicks the GPLed program X ... much of OS X's eye candy can be traced back to the first program to ever use images in window decorations and borders: Rastorman's enlightenment, but I digress).

What you are ignoring are the tens of thousands of small firms, like my own, like the dozen's a friend of mine consults for, and thousands of others around the world, who have used free software (GNU/Linux in particular) to build their infrastructure, their inhouse software, and their business out from under the everpresent heal of Microsoft, Sun, or Apple (though the latter has become a kinder master of late than the other two, it nevertheless remains a master). The success stories you are ignoring range from the traders I work for, who earn a sizable chunk of money thanks in no small part to GNU/Linux's superiority over their competitor's infrastructrue, to Pengiun Airlines, to Wallmart, to BMW, to the German government, the Chinese Government, the Brazilian government, and others too numerous to list, all of whome are able to run their IT infrastructure on their own terms, within their own budgets, and free from the coerced upgrades and chronic, forced obsolescence of their software and their hardware at the hands of their operating system vendor, as is chronically the case for anyone subject to Microsoft, Sun Microsystems, and yes, even kinder, gentler Apple.

That is where you will find the innovation and the flourishing of the technology more than anywhere else, amongst the users which have been liberated by the Freedoms expounded upon by the Free Software Foundation, people like myself and many, many others who have been able to flourish even during these difficult times, thanks primarilly to the GPL and the user's freedoms which it helps to guarentee.

Freedoms that allow us to run our businesses and earn a living, without being subject to the whims and deprivations of a copyright cartel or a convicted monopolist.

Re:You are looking in the wrong places

[sheldon]

I believe he was asking for examples of innovative creations of technology, not uses.

Re:You are looking in the wrong places

[FreeUser]

But in practice, we're just not seeing all the innovation that you would expect. The classic examples are always cited--Perl, gcc, Emacs, Apache, the Linux kernel--but for a model that's supposed to be so superior, the success stories are fewer than you'd expect.

Differentiating between innovative creations and innovate uses of technology is semantic nonsense. Even in the purest act of creation, in which you create a new program that does something no one ever thought to do before, you are still making innovative use of the underlying technology, including the hardware, the operating system, the compiler, and any number of libraries.

In house software for trading systems, for accounting systems, for real estate analysis, or for any other purpose, are no less creative for being in house, and no less innovative. And in many cases, what makes it possible for an average, non-wealthy person to do, is Free Software.

All users may not be developers, but every developer on the planet is by definition also a user of technology. The GPL is as much about user freedom as it is developer freedom, and the success stories may be uncomfortable to those of you who like to advocate a return to proprietary bondage, but they are no less real or potent for that.

Re:You are looking in the wrong places

[sheldon]

Differentiating between innovative creations and innovate uses of technology is semantic nonsense.

No, actually it's not semantic nonsense.

But thanks for trying to participate in an intelligent discussion.

Usefulness is important [Re:Sounds great on paper]

[X]

Perl is (arguably) the work of one loan, mad genius, and not GPL.

But Perl is free software, and it is actually a lot more than the work of a loan, mad genius. It certainly started out that way, but the same is true of most good software.

The gcc compiler is by far not the best compiler for any platform. It's sole advantage is that it's portable.

Actually, on many platforms it's the *only* compiler. There were times where it generated the fastest code of any compiler out there for a few platforms, and I believe that is still true today on x86 (I could be wrong because I haven't benchmarked it against MSVC 7.0). In general, even proprietary compilers tend to not perform as well as the specialized, high-performance compilers that CPU vendors write for their own chips. It's true though that gcc's primary goals have been stability and portability. To that end they've been more successful than any of their competition in the proprietary space.

I could go on. Heck, during most of the last 20 years and 90's it wasn't uncommon for admins to replace their vendor supplied Unix programs with GNU or BSD versions because they were better.

Still, you have to give props to the "usefulness" of Linux.

In general, free software doesn't tend to be terribly interesting from an academic standpoint. However, it does tend to be very "useful" (to paraphrase you). I'd argue that every piece of software you put on that list would score high on the usefulness scale. This is actually what you'd expect to be the emergent quality of software that is ultimately controlled by the end user. Don't undervalue "usefulness". That quality is by far the most important in software, and the one which is most lacking. If that were the only quality where free software exceeded proprietary software, it's reason alone to scrap the proprietary development model.

The Author Does Not Understand the GPL!

[Gorobei]

He's confused about licenses versus copyright, e.g:

The GPL's language, then, tries to swallow every piece of software that incorporates a GPL program or portions of it. However, it probably cannot do so under the copyright law it purports to follow. The definition of a derivative work proves highly controversial in the software world, and may not stand up in court. [Page 15.]

This has nothing to do with the GPL!

If I create a derivative work from a copyrighted piece of GPL software, I can release it under my own terms without accepting the GPL. I then probably get sued by the COPYRIGHT holder, and the courts then decide if this was a derivative work, a new creation, a copyright infridgement, or whatever.

Alternatively, I can accept the license (the GPL,) in which case copyright law no longer applies: I have freely entered into a contract that gives me certain benefits and takes away certain others (e.g. my right to claim copyright trumps the license I just agreed to.) If I now start doing evil like violating the GPL, I get sued for violating a LICENSE, not copyright law.

The GPL does not "build on" copyright law - it's a license that I can agreed to or not. If I don't agree, copyright is the operative thing, if I do agree, the license if the operative thing!

Re:The Author Does Not Understand the GPL!

[Salsaman]

by
modifying or distributing the Program (or any work based on the
Program), you indicate your acceptance of this License to do so, and
all its terms and conditions for copying, distributing or modifying
the Program or works based on it.

Seems pretty clear to me. The act of modifying the program, or distributing it, and the act of agreeing to the license are one and the same. In other words, you can do what you like with the software, but if you distribute or modify it, then you have already agreed to be bound by the terms of the GPL. You might imagine that you hadn't agreed to the license, but how else would a court determine whether you had agreed to it or not ?

Re:The Author Does Not Understand the GPL!

[aridhol]

You might imagine that you hadn't agreed to the license, but how else would a court determine whether you had agreed to it or not ?

If you didn't agree to the licence, you have no legal means of distributing the code - you would be breaking copyright law.

Re:The Author Does Not Understand the GPL!

[Gorobei]

The usually "defenses" by the GPL-violators appear to be:

1) A rogue summer-programmer copied the GPLed code into our product. [This probably doesn't fly - he was their agent.]

2) We re-implemented the code (fair enough, though most "reimplementations" are just rearranging and changing variable names, which probably will not stand up in court.)

3) We never read the GPL, so we couldn't have agreed to it. E.g. we found the routine on the web using a search engine, cut&paste, didn't read the rest of the file.)

4) We based our program on a stripped copy of the source we found somewhere - it contained no GPL.

5) The GPL is weaker than a click-through licence, we didn't even click to accept!

2 is completely a copyright issue.
3 & 4, if found credible by the court, may make the whole thing come down to copyright.
5 is interesting, but I don't know enough to judge.

Re:The Author Does Not Understand the GPL!

[bovinewasteproduct]

But remember the definition of dervived software. If I don't accept the GPL as a license for me to use the software, then ONLY copyright law stands.

That means I can look at header files and even the code to see how it works. All I have to worry about are patents, as trade secrets are moot as regard to GPL'ed software.

That means I can see how the authors implemented it, but I just can't use their code.

BTW, I publish under a BSD license, so most of this does not concern me directly...:)

BWP

Please don't read it

[T.E.D.]

Normally we have trouble with folks posting here without even bothering to read the article in question. But in this case I'd argue that would be a good thing.

I read through about the first half of it, but his explanations of the GPL are so confused and horribly unsound that attempting to follow it was starting to muddle my brain.

In the first paragraph it struck me as odd that he claims that he was recently convinced of the GPL's innovation harming qualities, but now is convinced that the complete opposite is true. Executive summary: I know I sounded convinced yesterday. But I was actually full of shit then, and just didn't realize it. But today I really do know what I'm talking about. Honestly

However, after reading his embarassing attempts at logic, its perfectly understandable how something like that could happen to him. With logic that muddled, tomorrow he might be convinced the GPL has manged to retroactively kill the dinosarus.

My suggestion is that anyone who doesn't have a real good handle on the GPL already avoid this whitepaper like the plague. Folks who fully understand the GPL should take the Monty-Python "World's Funniest Joke" approach, and split up into teams with each person analyzing a small bit of it seperately. That way perhaps no one person will read enough of it to become fatally confused.

Re:Please don't read it

[selectspec]

I agree. The article was bad news. What the article fails to understand is that what my opinion or this guy's views of the GPL is is irrelevant. It is the opinion of a judge someday to decide the fate of Intellectual Property that links or otherwises uses GPL (and even LGPL) code. This is the fear that companies face with the GPL, uncertainty.

I wouldn't exactly trust this guy's conclusions...

[aardvarkjoe]

Let's see ... on the page of the paper, the author asserts that the public and corporate perception of Linux has undergone a 180 degree reversal in the last year. (Simply not true ... a year, even two or three years ago, Linux was very much on most of the mentioned companies' radar. He's trying to pass off his change of mind by claiming that it was the industry's) He also states that "Linux and its global network of developers have long sought nothing less than 'total world domination.'" (Again, untrue; 'Linux,' as an operating system, doesn't have any desires. Linus Torvalds, the head of the "global network of developers," has repeatedly stated that Linux is a hobby, not a grand political scheme. The only one interested in world domination is RMS, however, in that case the article shouldn't be about Linux at all; it should be about GNU and the FSF.)

I'd take any paper that starts that way with a grain of salt.

how 'bout a little puppet show

[Eil]

<RMS MODE ON>

Any part of Mr. Asay's article where he implies the GPL is bad or could use some work is merely due to his failing to comprehend or understand the meaning and scope of what the GPL in its entirety.

<RMS MODE OFF>

Isn't that pretty much what we're going to hear in a few days? I'll thank the moderators for at least attempting to have something resembling a sense of humor.

P.S. No offense meant to RMS or his devout apostles.

Re:how 'bout a little puppet show

[T.E.D.]

Any part of Mr. Asay's article where he implies the GPL is bad or could use some work is merely due to his failing to comprehend or understand the meaning and scope of what the GPL in its entirety.

Actually, it would be more accurate to say "Any part of Mr. Asay's article where he implies anything whatsover about the GPL, good or bad, is merely due to his failing to comprehend or understand..."

Or better yet, borrow a disclaimer from TV:
"This whitepaper is a work of fiction. Any resemblance to actual facts found in this paper, either living or dead, is purely coincidental."

You'd honestly be better off learning about the GPL from reading Microsoft whitepapers. That's how bad it is. At least Microsoft will use decent logic after spreading the initial foundation of mischaractarizations. This paper is about as accurate and well-reasoned as a statement from Lyndon Larouche.

I can tell you didn't actually read it before posting, but in this case I won't gripe at you for that, because on the balance that's probably a good thing. I read about halfway through, and just now got back from the subsequent bathroom trip (I feel much better now).

Re:how 'bout a little puppet show

[Eil]

My RMS MODE pseudo-future-quote was, of course, merely poking fun at RMS as that seems to be his standard disclaimer whenever someone has a legitimate gripe about his opinions, projects, or writings.

I don't think the GPL is bad in and of itself, but I hardly think it to be the One True License for Everything either. For instance, if I'm sitting at home writing up a neat little bit of code in my free time and I want it to be used as well as improved then the GPL is perfect. Just about any other scenario will cause you some amount of problem, however.

No, I didn't read the article but for good reason: hanging around the OSS community one gets a pretty good familiarization of the ins and outs of the GPL. Familiarization, dead-horse style, that is. I just wanted to do a little good-humoured RMS bashing, that's all.

Re:barf!

[Exotabe]

You apparently didn't read the papers. Both of them.

The author makes it very clear in his second paper that he has completely changed his opinion on the GPL as a viable licensing model.

He spends a lot of time sorting out both the myths and truths surrounding the GPL's effect on closed-source companies, OSS developers, and consumers/end users. Especially interesting was his analysis of how dynamic linking to GPL code effects proprietary software development/licensing. In short, there was a lot more to the documents than the topic sentences you grabbed from the first couple of pages.

As an aside and a personal opinion, the first of your quotations is actually a truth, not garbage. OSS has been playing catch up for quite some time, as it had a much later start. Granted, it's come a lot farther in a lot less time, but that doesn't change history.

Anyways, if this sounds like a rant, it probably is. After reading all 47 pages, the last thing I wanted to see was a blatent troll cluttering the forums.

Modelling

[Animats]

Fashion modelling isn't a big-money business. Ask any woman who's done it. A few hundred people make decent money on a regular basis; everybody else doing it needs a day job. There are more good-looking women who want in than there are jobs.

You meet enough actress/model/waitress types in SF, NY, or LA that it's a cliche.

Re:Yes, but...

[Stephen Samuel]

"The people own it. There is no patent. Could you patent the sun?" (although granted, the likes of Jack Valenti [slashdot.org] probably would love to have IP ownership of everything and anything).

Or, as my (very right wing) sister once said (after a tornado hit Edmonton),

• Too bad you can't sue god.

She was in law school at the time.

An FSF Response to Asay's article

[bkuhn]

We at FSF have spent a few minutes reviewing Mr. Asay's paper. We are glad to see that someone who once took a clear anti-GPL stance now claims an "admiration" for it. We noticed the paper contained a number of email quotes from our founder, Richard Stallman, and our General Counsel, Eben Moglen. The quotes appear to be taken from various email conversations that Asay conducted with Moglen and Stallman separately. Quoting email out of context is always tricky, and the conflict and confusion with our official statements regarding the GNU GPL are apparent rather than real.

We do, from time to time, have occasion to clarify our views on the GNU GPL, and we work hard to be responsive to questions from the community on them. This does serve to generate a lot of email traffic concerning the Foundation's opinions on GPL, and out of context pieces of the whole of such correspondence can be confusing. When we get pieces together that are particular good and clear, we post them on our philosophy page, as we did with Eben Moglen's essay, Enforcing the GNU GPL, and in our GNU GPL FAQ.

As for the GNU GPL, version 3, we had indeed long ago seen Mr. Harris' proposal, but we did not feel it was representative of our plans for next version of the license. The copyright of the so-called draft of GPL Version 3 distributed alongside the article, thus, is incorrectly attributed to the Free Software Foundation. We do not hold copyright on it, nor did we have a hand in drafting it.

Efforts for development of real GNU GPL, Version 3 continue within the Foundation. We of course expect to have extensive public discussion before any draft is finalized as the official new version of the license. In March 2002, We approved the Affero GPL, which includes a draft of one major provision we are considering for the next version of the GNU GPL. Of course, when we do announce that a draft copy of actual "GNU GPL, Version 3" is available, it will certainly come directly from the Foundation.

Bradley M. Kuhn
Executive Director, Free Software Foundation

Still Not Clued

[Euphonious Coward]

Remarkably, after all this time, he's still only partially clued. As might be expected when somebody was totally wrong and announces the fact, when he explains his new understanding there's little more reason to take him seriously than before.

From such little details as getting the names of the licenses wrong -- leaving the word GNU out of the title of his paper, and elaborating the LGPL as the "Library" GPL rather than its correct name, the Lesser GPL -- to totally missing the true competitive value of the GPL to businesses, as well as the very simple economic motivation for developers to participate, he still has a lot to learn.

If there's anything new or interesting here, I didn't find it. He's articulate, though, and evidently open to reason after he has been pummeled sufficiently, so he might even get it right on the third try. In any case he's no worse than ESR.

Re:WTF is a whitepaper?

[volpe]

Look here.

tooth fairy

[epine]

That reputation argument is a crock. It sounds like some kind of monetary surrogate in the minds of people who can't organize reality in any other way.

I participate in open source because I still believe in the tooth fairly. I put a tooth with a cavity under my pillow at night, I wake up the next morning and the cavity is gone.

Once upon a time it was possible to have incredible wealth, yet still have bad teeth. You don't brush your teeth to improve your reputation. Perhaps those who bleach their teeth are in it for personal profit, but the rest of us just want good teeth _without_ spending a fortune on dental work. Is it really that hard to understand?

As a recent Linux adopter...

[Mithal]

I completly lost the author when I read this quote:

Very few applications exist in the open-source world, at least those that would be useful to the average user.

Completely false! Especially for "average user". Tons of applications to suit all my need were installed automatically when I installed Mandrake. And I could easily complete by looking on the web (gimp, OpenOffice).

The average user needs a browser. Konqueror, or Mozilla, or derivatives will do.
The average user need a mail reader. Mozilla, Kmail, will do.
The average user needs a word processor / spreadsheet. KOffice, or OpenOffice, will do.
The average user needs a MP3 player, and instant messaging client, .... There is some on linux already.

What else do you REALLY need? The author certainly didn't look much before making this affirmation. There are plenty of applications for Linux. And more are coming our way every day.

The author also says that the GPL is an incetive killer for developper. For COMMERCIAL applications only. The way I see it, there is a lot more programming done with no commercial intentions. I like to code; and I always though of releasing some of my stuff as shareware or freeware. GPL is even better! Someone might improve it! What I think the author doesn't get is that the GPL intends to change the buisness model of software makers. And that is a good idea.

Tools vs Art

[_Sprocket_]

One striking example is that the GPLed games for Linux always tend to variations of Tetris, Boulder Dash, Missile Command, etc. You would expect the innovative games to be coming out for Linux: no pressure from marketing, free development tools, big community. But the games with spark, like The Sims and Grand Theft Auto 3 are coming from elsewhere.

As others have pointed out, modern games are becoming more and more a blend of technical and multimedia artistry. Part of this may simply be exposure and time will bring more artists in to the Open Source community. But some of it is also the culture and nature of artistic work.

On the technical side, we already have some great coders working on Open Source tools. Part of this may be that coders are the first to benefit from other Open Source projects and feel compelled to "give back" to the community. But also part of this is definitely the positive return from finding others to collaborate on your project. Development of tools benefit from multiple developers. It might be worth noting that there are some very interesting Open Source projects involving toolsets and engines for games.

The artistic side comes from a different world. First, most of their tools are proprietary... although with tools like the GIMP and Blender (knock on wood), this may change. Secondly, art tends to have a limited collaborative nature. And finally, art does not need an equivalent of source code to improve from one example to another. In fact, artistic influence is largely unaffected by IP laws. For example, while reading many of the better fantasy novels available today... one will often see the influence of J.R.R. Tolkein - despite a rather aggressive protection of copyright by the Tolkein estate (just ask TSR/WotC or whatever they're called now). This artistic influence is felt within other aspects of artistic endeavor also. Simply put, artists do not have as much a need for Open Source ideals as coders do (although "look and feel" and other issues are very similar technical examples of influence).

But that doesn't mean artists will never find themselves in the Open Source fold. As more and more interesting tools begin to appear, Open Source may begin providing an environment that is attractive to artists' desire to experiment. And this experimentation is often produces some of the most compelling content.

It may simply be a mater of time.

Re:Yes, but...

[LMCBoy]

In the case of pharmaceuticals, they typically get the bulk of their R&D free from publically-funded university research. The government then gives a particular company a legal monopoly on bringing the fruits of the research to market, garnering huge and easy profits for the company.

Your reference to Microsoft is apt; they would like something similar: to be able to reap financial profits from the "R&D" work of the community...too bad they have the darn GPL standing in the way! I bet they're pretty jealous of the sweet deal the pharmies have going...

Holes in his proposed GPL...

[anthony_dipierro]

• AS IS isn't capitalized. The UCC requires for it to be.
• It doesn't define "source code".
• It doesn't require a distributor of a derivative work to distribute that derivative work under the terms of the GPL unless that distributor is the same person who created the derivative work.
• You only have to include a notice describing the changes you made, not the changes made by others. That effectively makes the requirement meaningless.
• A distributor cannot impose license restrictions, but the creator of the derivative work can.
• Clause 5 is equivalent to a restriction against removing the "do not remove" sticker on a mattress. I should be allowed to do whatever I want with my copy of the software.
• I'm sure there are others.

Derived Work, Re:proposed revision of the GPL

[angel'o'sphere]

You left out the part about linking. If we are to clear up that part, I'd much prefer to make it clear that some forms of dynamic linking and server-izing (as with CORBA) do create derivative works. I think the key concept of a derivative work in software has nothing to do with linking, and everything to do with the span of the thread of control in a computer program. This can indeed happen across physical system boundaries. The boundary for the copyright should probably be at an API that is explicitly designated as an interface to separate works, such as the kernel system call interface. Not just any API.

Thanks

Bruce


What the term "derived work" means is defined in the appropriated laws. E.g. copyright law.

According to the law using CORBA and such a like is not creating a derived work.

Form a programmers standpoint of course one could think it would.

However, keep in mind: copyright is about creating and distributing copies.

As long as you can "use" a copyrighted work wihtout the need to copy (or incorporate in parts) that work, you have not created a derived work.

That means by loading(and this is a copying!!) a CORBA server into main memory of my server, I need usualy a license to be allowed to do so.

Clients interfacing with that server are not copies, nor do they usualy include/use any portions of the server as copies or adaptions or variations. So the clients are no derived work.

Of course your work still depends in its functionalyty on the original one ... but that is not an issue in copyright, probably it is in patent law.

So your "key concept", see above, is (fortunatly, IMHO) not the way copyright law sees it.

Regards,
angel'o'sphere

Re:Derived Work, Re:proposed revision of the GPL

[Bruce Perens]

According to the law using CORBA and such a like is not creating a derived work.

Go read the law. It does not say yes or no about this linking issue, it says almost nothing about software. Then go read the only relevant case, Nintendo v. Goloob Games. Note that the judge said the case is not definitive.

Bruce

Re:Worldwide monopolies?

[Stephen Samuel]

polyopoly " - or perhaps "poliopoly", but that sounds a bit too much like someone who is infected by multiple polio viruses.

But -- as was pointed out: Jonas Salk placed his vaxine into the public domain -- and thus saved millions of lives and made polio almost extinct (with the exception of 'defensive' bioweapons stores).

From one story about Salk:

The success of the vaccination effort won Jonas Salk unsought fame. The March of Dimes, hoping to boost publicity and donations to fund vaccination programs, lionized Salk to the point of offending his colleagues. He had applied the findings of others in a successful bid to prevent disease. Other researchers and doctors grumbled that he hadn't found anything new; he had just applied what was there. But the timing of his successful vaccine at the peak of polio's devastation made the public blind to that.

Had todays IP laws been in place back then, much of the work that Salk depended on would have probably been patented. He might not have been able to create the Smallpox vaccine, and many of us here today would have thus been dead.

Re:You don't HAVE to give away your GPL'd apps

[Eric Damron]

Actually, everything he said was 99% true. Not eloquently stated but defiantly true.

If you create a piece of software that is based on another GPL'd software then your software must be covered by the original GPL license. You are not under any obligation to release the binaries or source. However, if you do release the binaries you must make the source available upon request.

This effectively prevents someone from making changes to a GPL'd product and then sell the changes for a high price. They could do it but the buyer would have the right to distribute the code for free thus killing the seller's market.

The only part that I disagree with is that because the programmer is being paid to product the product his employer would have the exclusive right to initially release the code. So unless the employer released to code to the programmer, the programmer couldn't release it to the general public.

If I am incorrect, perhaps you could point me to the paragraph in the GPL license that supports your position.

Re:Is linking closed code to GPL library "fair use

[spitzak]

This is a good question. It does seem that it is impossible to prevent this because that would clearly violate free-speech laws.

However it requires distributing the code so the receiver can see how it works. I think anything where you precompile the code could be defied as a derived work (because the compiler incorporated information from the copyrighted header files) so it is almost impossible to obscure your code.

But it does seem you can put rules on the code like "you can't give it to anybody else" which seems not in the spirit of the GPL.

I would think that because of this reason, ALL GPL code can be considered LGPL for programs that use it and are distributed as source code.

GNU advocates stealing Apache credit

[joneshenry]

Yet another GPL/GNU advocate steals credit from The Apache Software Foundation while claiming that Apache's success supports the GPL. Ironic how the same movement that is trying to advocate the name GNU Linux is shamelessly citing Apache's success as its own. Of course the Apache Software License is incompatible with the GPL so that code from the Apache and GNU projects cannot be mixed whatsoever, but that doesn't stop the GNU advocates from citing Apache as support for the GPL.

How many more of these articles are we going to see that make bizarre claims like Apache is GPL software? (Note the passage containing "today thousands of GNU GPL software packages are already available including operating systems (like the famous Linux OS), office packages, Internet servers (like Apache running on 30% of Internet Servers".

Re:GNU advocates stealing Apache credit

[istartedi]

Argh! You're right. After posting my usual snide remarks, I decided to download the article and read it. Right there on the first few pages is Apache's market share graph. That Apache isn't GPL'd (and that claiming it for the GPL camp is one of the oldest lies in the Free Software advocate's handbook) is bad enough. What's even sadder is that it plainly shows Apache to have "peaked" and lost ground to IIS (note, I'm not saying that's good or bad, it's simply an observation).

I stopped reading the essay when I saw that.

So, for once I am not going to offer any calm objective arguments to follow my bombast. I stand by the troll.

That would be horrible

[Sloppy]

The boundary for the copyright should probably be at an API that is explicitly designated as an interface to separate works, such as the kernel system call interface. Not just any API.

Imagine this scenario: Hacker reverse engineers an internal interface in a Microsoft OS, perhaps to write a MSIE replacement or a DOS TSR (heh, remember those?). His program is a derivative work?!

Re:You don't HAVE to give away your GPL'd apps

[Eric Damron]

"I think that's what I got out of the ref'd whitepaper, but also doesn't it say that you only have to release source upon request within 3 years? Or did I just misunderstand."

Well, if you mean by "within 3 years" that you can wait three years before turning over the source then I believe you misunderstand that part. The offer to turn over the source must be good for three years after you release the binaries. So if you get a request for the source the very next day after releasing the binaries you must honor that request within a reasonable time.

"Also, the linking, kern module, and other issues are an extremely big question mark, IMHO, even though the author cites precedent as answering the question for us, ie, the existence of binary only kernel drivers, and programs linking with GPL'd shared libs."

Yes, this also concerns me. I could see it going either way. I think that Linus should make it clear that for the kernel, statically linked hardware device drivers are acceptable.

We need to allow hardware vendors to support us and yet we must not allow some jerk to take GPL'd code and compile it into an "so" file on Linux or "DLL" on windows and feel that they don't have to follow the GPL license.

Copyright makes derived works quite broad

[smiff]

As long as you can "use" a copyrighted work wihtout the need to copy (or incorporate in parts) that work, you have not created a derived work.

I can go off and write a completely new novel set in a world where Jedi Knights swing light sabors around, use something called "the force" to telekenetically move objects and influence weak-minded people. I could do this without copying a single screenshot or character from Star Wars, but George Lucas would still sue me for creating a "derived work".

Best let the courts decide what's 'derivative'

[smiff]

If you spell out exceptions in the license, you simply create loopholes. Consider your example:

The boundary for the copyright should probably be at an API that is explicitly designated as an interface to separate works, such as the kernel system call interface.

What is to stop someone from modifying that API to make it exploitable? Certainly, if one were to "embrace and extend" ext3 so that every file has two forks, the result would be a derivative of ext3.

Suppose someone adds a new plug-in API to the ext3 module. They'll call it ext4 with plug-ins. They then sell a plug-in that allows people to have two forks for their files. In my opinion, this is a derivative work, but under your definition (and the paper's definition) it is not.

The same situation comes up with communication protocols and file formats. I think it's best to let a court decide what is not a derivative work. It would be nice, however, if the GPL gave examples of things that do and do not constitute derivative works.

Personally I think anything that breaks compatibility should count as a derivative work.

OT: Re:Innovation

[extrasolar]

What you say is true.

Someone who scores high on the SAT test doesn't necessarily deserve all them scholarships either. This is just another example.

But pride is a different issue. A person is not proud of being beautiful, just as a person is not proud of being smart. If the person thinks he or she is proud, then I can only attribute this to delusion of some kind.

What is worth being proud of is what you accomplish. The supermodel is at a great disadvantage in this regard. Already she (or he) has decided not to be rewarded for what she is proud of. Lots of people envy supermodels for their beauty and lets face it, for the sex and popularity they have. But there have always been hotties in the world. The difference is that our allegiance to basic human rights gives beautiful people an amazing social advantage. In other times, the opposite would be true--the beautiful people would be the most often raped and enslaved for the worst reasons.

But the great thing about these principles of human rights is that with work you can do things you are proud of. This is an amazing virtue. This is true no matter what genes you are born with.

The point is, in your life, to accomplish something. Don't bullshit yourself either. Raising kids isn't, in itself, an accomplishment. People have had children in all ages, under all sorts of circumstances and privilidge. Scoring high on the SAT, in itself, isn't and accomplishment either.

Life is without meaning, otherwise.

Complete bullshit

[fferreres]

Supply and demand explain PRICES (and wages for the fact). The don't measure VALUE. What is more valuable, 100 tons of rice or a big diamond? As you can see, they are not equal though their prices may be the same.

Abstract:
VALUE = subjective
PRICE = supply and demand

Demand in turn depends on how money is distributed among society. If africa had more money, rice prices would rise. If the world was about 30 bill gates and 20 Soros and 5B poor, diamonds could be worth all the food in the world).

And values do exist. It is not for nothing that you get a life sentence for murder. Also, apparently humans can't be sold for a price (slavery). But if you don't earn enough money to buy food and shelter, you die. So you could say the right to live is arround U$S 300 a month. That is about $30.000 a human. That's our (base) price, though apparently, we are priceless. But of course, VALUE != PRICE, so you can't kill someone and pay $30.000 (nor ANY ammount) to get away from prison.

Re:Not GPL

[pmz]

you clearly argue against GPL...

I disagree. I believe I am arguing for both GPL and non-GPL Open Source.

Let's say I write some software that has a proprietary set of algorithms, but I want to make the file format and its API open to everyone. Simple solution: keep the algorithms closed source and release the file format and its API under the GPL. My business wins and the community wins.

Let's say I write some software that I want businesses and hobbyists alike to use in their own software, but I'm not in it for money. Simple solution: I release my software under a BSD or LGPL license.

Let's say I write some software that I want any one to use for free, but I want to ensure it can't be made proprietary. Simple solution: I release my software under a GPL license.

Any of these paths can lead to a success story, but success isn't necessarily defined by huge popularity or seven-digit incomes. Here, the author chooses a desitiny for the software and, if the author is satisfied with the outcome, then the software is successful. What drives the author's satisfaction is something we can't define here.

Re:Consideration and the GPL.

[Royster]

The GPL is a license, not a contract. Some licenses are contracts, but they do not have to be. The GPL is more powerful in that it is a license and not a contract.

I agree, but

[Per Abrahamsen]

it is actually nice to see how a person despite a lack of understanding of the GPL and general muddy thinking have come to apreciate it, simply by observing its effects.

Apparently we can win over people who don't think like a programmer, by the simple fact: It works.

Re:barf!

[selan]

I based the comment on this statement from the GNU site:

"The GNU Project was launched in 1984 to develop a complete Unix-like operating system..."

I think that RMS named it GNU's Not Unix to make sure no one would get confused.