Charles Cooper gave me an "F" in software at c|net news (though kindly after giving some nice words to Creative Commons which will go live October 9th, the day Eldred v. Ashcroft gets argued.) He picks up an argument raging in this space about software copyright. I'm for it, as Cooper accurately reports, though I'd only grant it for 10 years, and only upon the condition that the source code for the software be deposited in escrow with the copyright office. Once the copyright expires, the source code would be free.
Cooper grades this idea an "F" because he thinks this would only "formaliz[e] the existing constellation of power that favors the Microsofts and Oracles over the small and independent developers." As he gives no other reason to oppose the idea, I guess that means an idea that would benefit "the small and independent developers" over the "Microsofts and Oracles" would get an A. So I hereby petition the professor (look, he's the one giving out grades here) for a change of grade: My proposal would clearly benefit the "small and independent developers" more than the law today.
My proposal has two parts that it helps to keep separate. One is that the terms for software should be much shorter than they are today: 10 years rather than 95 years (for software written by an author but for a company). The other part is that the source code for that software be kept in escrow, so that when the copyright expires, the source code becomes free (as in both free beer and Free the Mouse!). Both parts, imho, would benefit the "small and independent developers."
I think terms for software should be just 10 years. One reason you might agree is that a short term is what our (at least for those of us who are American) founders wanted. Cooper says its "hard to know what the founding fathers" would have thought about the matter. Actually, that's not so tough. They said copyright terms had to be "limited" so that creative works would pass into the public domain "at as early a period as possible," Pennock v. Dialogue, 27 U.S. 1, 19 (1829) (Story, J.). Why? Because the framers were keen to have others build upon creative work, after copyright assured the "author" a sufficient return for his or her creativity. They believed, in other words, in a public domain, and they required that copyright terms be "limited" so that the public domain would flourish.
Consistent with this theory, they gave copyright owners of "maps, charts, and books" a short copyright term: just 14 years, renewable once if the author survived. Even that protection, however, was more limited than the system today. Unlike today, their copyright allowed others to build derivative works (such as translations, or abridgements) during the initial term of copyright. And unlike today, because the "source code" of those maps, charts, and books could be easily incorporated into new, creative works, it was easier for others to build upon work that had gone before. Their law, in other words, made the creative work of "maps, charts, and books" open sourced, or free, after a very short time.
On their view, the ultimate question Congress should ask is "how long does the term have to be to get lots of great progress" (for, as the constitution says, the purpose of copyright is to "promote the Progress of Science.")
I'm not sure I know the answer to that question exactly, but I am quite confident that it's not longer than 10 years. And the way to see this is to ask yourself a single question: Do you really believe there is any coder or coding company out there that is deciding whether or not to invest in a new coding project on the basis of whether they will have the right to control an unmodified version of their code for longer than 10 years? The bold is important here, because remember, with each new version, you get a new set of copyright protection. So the issue is not whether you want to be in business for more than 10 years. The question instead is, "do you need a head start that is longer than 10 years?" or "If we only give you a 10 year head start, would that stop you from running the race."
I don't have much code on my box that's been unmodified for more than 10 years. I certainly don't know of any I can buy from the original software developer. Thus, it doesn't seem that there's a huge market out there of developers depending on selling 10 year old code. And thus, it doesn't seem plausible that there are many developers who would decide not to code just because they couldn't sell their unmodified code for more than 10 years.
Thus we can conclude from this that there are not many "small and independent developers" who would be hurt by shortening terms to 10 years. But it doesn't follow yet that many would be benefited either. So the argument needs another step.
That's the reason for part 2: That source code be held in escrow till the copyright expires. It is this part, I believe, that would on balance benefit the small developer.
The reason I thought was obvious, until a strong will suggested it was not. But the replies to my question have convinced me again: Having the source code would make it easier for the "small and independent developers" to compete with the "Microsofts and Oracles." (Easier, not easy; compete, not necessarily defeat).
As the open and free software movements demonstrate, it is easier to build on top of something than to build it from scratch. Thus, forcing competitors to build from the bottom up is a nice "barrier to entry" into a market. Having a chance to build on what went before means having a greater chance to compete. If the source code from expiring copyrights were available for others to take and build on, it would make it easier for these others to compete with the likes of the "Microsofts and Oracles." To build a substitute, for example, of an operating system (compatible, but different) would not require building a whole new operating system. To build a competing version of a product would not require building the product from scratch.
No doubt, the "Microsofts and Oracles" won't like this competition much. And it may well be, economically, that on balance, engendering more competition against them is a bad thing (I obviously don't believe so, but to be complete, it's possible.) But on Professor Cooper's scale, the question is much simpler: whether the change would benefit the small guy. And it is fairly clear that lowering the barrier to entry will help the small guy more.
And then there is the further argument about more. Or rather, More. Because while I'm quite certain developers don't need a term of longer than 10 years (some economists argue the term need not be longer than 2 years), and fairly confident shorter terms would help small developers more than larger developers, there is one part of this argument (too often missed in the screed surrounding this debate) that I am absolutely certain of: the benefit shorter terms plus source code would provide for public domain development of software.
For take the case of More. More is an outliner, and my all-time favorite computer program. (From the Department of Irony: guess who helped inspire and build More?). No code ever did what it was supposed to do better, or more intuitively. I use it to this day -- indeed, literally so, because I am working from a More outline while writing the briefs in the Eldred case.
More is more than 10 years old. The last copyright notice on the version 3.1 that I have indicates 1991. Since the early 1990s, there's not been any commercial development of More. You can't buy it from the copyright owner; you can't update it to take account of evolution in software development. It is an orphan within the archives of Symantec Corporation, trapped by an owner that won't let it go.
This is not because Symantec is evil. (Indeed, Symantec is a great company that released to Dave Winer an "antique version" of More.) There's obviously good business reasons not to develop even the best program ever. And there's also good business reasons (given the world of copyright and tax law as it is) not to release the source code of More. And thus it sits, stagnant, awaiting the time when the copyrights expire (2085), or for people to reverse engineer it to make it run again.
But if the copyright were to have expired, and if the source code were available to all, then there is no doubt that some would find it useful to keep More alive. I'm sure there are many who share my affection for this clean bit of code, and I'm sure at least some could do good work bring it back to life. And not just More. Think of HyperCard, held hostage by Apple Corporation, despite a fanatically enthusiastic community of developers and fans who would keep it alive. Or Newton (yes, still protected) (no, stupid, not the man), which also has a following of fans.
If this code were free, then there are many who could, in the tradition of our framers, build upon it and use it, whether commercially or not. And thus at least some orphans would find a home, cared for and fed by those who love them most.
So how about it, Professor Cooper. Shouldn't we Free the Mouse, and More?