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

[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

[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

[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

[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.

barf!

[Anonymous Coward]

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

``The GPL is an incentive killer.''

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

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.

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: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: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: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

[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:GPL Revision

[Anonymous Coward]

why doesn't the writer submit it to the OSI [opensource.org] as a new license? I'd use it.

Don't! This simplified language carries some nonobvious bombs. For instance, paragraph 8 says that

if you sue anyone over patents that you think may apply to the software for a person's use or distribution of the Software, your license to the Software ends immediately and automatically.

Which may very well mean that Red Hat may be blocked from enforcing its patents (i.e. on Ingo Molnar's kernel work), possibly against some bad guy.

At this point, I'm not saying this is good or bad, I'm just saying it's nonobvious and needs to be carefully thought through.

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

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.

This is one of the most insightful posts...

[Anonymous Coward]

...that I've seen on the subject of open-source gaming.

I'd like to add to this that in many cases, this is because many genres of gaming currently lack a critical theory. That sounds sort of pretentious and vague, but what I mean is that many genres of gaming haven't approached the stage where the players, developers, and critics consider the game an art form that follows a set of theoretical ideals. I know that lots of you might say, "no, lots of good games are just fun", but I think that's not necessarily true, and that that attitude is exactly what I mean (it IS possible to develop a critical theory around why Serious Sam is enjoyable, for example).

I would argue that often, having a critical theory itch to scratch is what drives good open-source gaming development currently. Thus, you see more good open-source gaming development in areas where there is a more theoretical appreciation of gaming: interactive fiction, FPS technical design and visual experimentation, etc.

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!

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: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: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.

Leave GNU/Linux & GPL alone!

[Anonymous Coward]

Why don't you people leave GNU/Linux & GPL software alone??

There is so much hypocrisy in your message.

If you do not like GPL software, please do not use it. Use something else, use MS Windows, BSD, whatever you like.

You would like to use GPL software but you do not like the GPL because it will not let you profit from the work of others available to you for free. How is that?

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....."