The Madness of Software Copyrights

My column in the current issue of CIO Insight, href="http://www.cioinsight.com/article2/0,1540,2182447,00.asp">“Battle Over Linux: When a Win May Not be a Win” , drew a lot of froth from Linux-happy bloggers who don't like to hear anything that isn't completely positive about their operating system of choice.

My crime was pointing out that partial summary judgment by a district court (in this case, in a lawsuit by SCO over the ownership of UNIX copyrights) is not the definitive victory the mainstream press made it out to be.

Since then, SCO Group has filed for bankruptcy protection, and the froth will probably come up again when I point out that Chapter 11 does not, likewise, end the threats to Linux, although it certainly starts to look like this particualr battle is running out of steam.

In any case, I only used the SCO litigation to make a larger point, which is the madness of granting the same copyright protection to software that the law gives to novels, music, and other works of art.

Which is not to denigrate the utility or workmanship of software or software engineers. I was one for many happy years.

But granting copyright for life plus seventy years for code that no one could possibly believe will have any use for anywhere near that long (what's the oldest software still in use today, do you think?) is a ticking timebomb.

Protection for software under copyright begins in the 1976 Copyight Act and, being lumped together with other "writings," is the perhaps-unintended beneficiary of all the extensions that lobbyists for large media companies secure for literary works.

Since the half-life of any code, given the rapid evolution of computing technology, is very short, what's the harm in excessive protections?

Well, as the SCO litigation suggests, some code, maybe little tiny bits of it, can live for a while. In the emerging web services architecture, compartmentalization and reuse are more likely. That's good from an engineering and economic standpoint, and ought to be encouraged.

So there's the problem. A small package of software hangs around for years, and its ownership is not entirely clear, and we have a future of these disruptive lawsuits popping up trying to use copyright as a blunt instrument to exact tribute from, well, from everyone.

Open source is, in some sense, a rejection of that scenario, wherein engineers declaim their rights under IP law in the interest of a common good, including their own.

What's interesting to me about the GPL license and other "open" options is that, by necessity, they rely on the existence of the rejected copyright regime to work. Which is to say that the basis of enforcing the openness of open source is that the authors of the open code really do have copyright hanging out in the background. Fail to live up to the open requirements of open source, and the authors of the code you started with can stop you--by arguing copyright infringement!

It's an ingenious idea, but may prove too smart for its own good.

Here's a (much) simpler solution: treat software differently, give it a much shorter copyright protection and let old code into the public domain (where it's truly "open source") much sooner, say twenty years after it's written.

Of course there's utterly no stomach for that reform in Congress or the software community, is there?