5333 private links
malor Ars Tribunus Angusticlavius et Subscriptor
jump to post
I think it's important to point out that this case covers only one kind of API, a direct programming API where you're incorporating libraries into your main program. Because of the rules of most computer languages, this means that the header files and user-written program files have to look almost exactly the same in the function definition and call location. If they don't, it won't work. If you're trying to call the C standard library printf(), and you call it as printg(), ain't nothing gonna happen.
Thus, any programmer providing a replacment printf(), if he or she wants it to work with existing code, must generate a printf() header in their library that will look pretty much exactly like every other printf() in every other C standard library. It doesn't mean they copied it, it means that there's only one possible way to express that idea. Thus, a reimplemented library of any kind will, by necessity, have a duplicate line for each function. If it doesn't, it won't work as a drop-in replacement.
However, many other APIs are network-based. It seems unlikely to me that copyright would ever apply there, because each provider of an API will be able to write their own code. There might be a few lines of identical source, but for the most part, the code base will be very different. Both Apache and nginx, for instance, support the http API, but there probably isn't much duplicate code between the two projects. Either can be used as a webserver, but it will be instantly apparent that each is a unique creation.
And then compare those with a hypothetical web server written in, say, Rust..... the new codebase would look absolutely nothing like its predecessors. Copyright protection would even more clearly not apply.
API is not a well-defined term, and I think the Supreme Court probably got this exactly right.... this type of API infringement is not covered under copyright, because it's functional. This is the only kind of API infringement where copyright is likely to apply, so a clear decision that it does not will probably put a stake in the whole idea. Almost any other kind of API 'infringement' would obviously not be a copyright matter.