Bruno Lowagie's Online Gazette

During my stay in Silicon Valley, I posted an angry (and very personal!!!) blog about "Evil Google" (it's MY opinion, not the opinion of the company I work for), and today I received a comment about it:

I think you should remember that this case isn't just about settling a dispute. Legal precedent might come out that will negatively affect all of us in the software community. One of the biggest legal questions is whether software interfaces and header files are copyrightable. This could terribly impact the open source community that frequently seeks to create open implementations of proprietary systems. I wouldn't think seeing Google pay for their sins would be worth that to you.

I have been thinking a lot about the legal implications, and I think people who are in favor of Google are shortsighted. As a developer (even a developer who has created an API that is used by many other developers), I'm slowly getting convinced that in the long run, APIs would benefit from being copyrightable.

The problem with interfaces is that they're open, but not standard. I've been working with PDF for a very long time. In the beginning, PDF was a specification of which the copyright was owned by Adobe. However, Adobe allowed every developer to use the PDF specification to create, manipulate, and render PDF files, provided that some conditions were met. The conditions were very reasonable: they were all about avoiding abuse. You weren't allowed to implement the spec in a way that didn't respect the spec. Is this a limitation of your freedom as a developer? Maybe, but I'm sure it's beneficial for the user: if he uses a method "multiply()", he doesn't want the method to "divide()".
PDF was a "de facto" standard, but in spite of the rights granted to developers by Adobe, many considered the PDF spec to be "closed", or more specifically "proprietary", as in "owned and governed by one and only one company." This changed on July 1st, 2008 when the PDF specification was published as an ISO standard. From that moment on, PDF was an open standard, a real standard, not a de facto standard. Is a standard copyrighted? IANAL, but as far as I know, it is. However: everybody is allowed to use the standard, and not a single company is the sole owner of the copyright.
I'm very much in favor of open standards, and I think that what happened to PDF in the past, will happen to interfaces in the future, that is: if the court decides that interfaces are copyrightable. Companies enforcing the copyright on APIs they own, will be able to do so, but the market will ignore them in favor of companies who bring their API to the ISO (or a similar standards' body that doesn't depend on one company).
I would very much like to see such a rupture in the software eco-system.

Summarized:
On the short term, I fear the day when interfaces will be copyrightable, but I think it's an opportunity to turn APIs into open standards in the future. On the long term, I fear the reality that companies such as Google will gain absolute power: in spite of what they say, they don't care about developers, and they have no problem taking away the oxygen from small companies who want to make an honest buck.