Owning a Language

[degibson]

Forgive the n00bish question, but I've never been particularly sound in the area of IP law. Here's the question:

Can a company own a programming language?

I can understand how a company can own proprietary software: compilers, runtimes, debuggers, etc. But if I open up a text editor and type some characters into it -- and they happen to be recognizable by that proprietary software -- do they own my source? (No)

So where is the line? In case it's not obvious, I'm thinking of Java here. What precisely makes a FOSS JVM a violation of IP law (or does it?)?

 

[Cogman]

The violation comes in the form of "Your JVM does what our JVM does." I don't think you could truly protect a compiled language (other then to attack those that make the compilers for the language). What you can protect against is people that do the same thing as you.

software patents, etc, are pretty loose however. MS, SCO, etc, have been suing people in the OSS arena for a long time. To dates, I can't think of any software patent suit that was actually held up by the courts (I could be wrong though).

 

[Markbnj]

Copyright protects a particular expression of an idea. Probably the main reason that patents took off in the software arena is that copyright alone was insufficient to keep other people from emulating an original piece of software. Patents can protected an original implementation above and beyond its specific expression in text, and can even protect processes, but they are horribly misused in software, as pretty much everyone knows.

Copyright is a much surer thing than patents, i.e. if I write something and register it with the LoC and you try to register the same or substantially the same written work, you'll be denied. I own that expression of the idea for 25 years, or whatever the period is now. There's no real ambiguity involved at all unless you allege that I stole the work and then registered it, which would be hard to prove. Interestingly, copyright is inherent the moment you create a work. Registration with the LoC just makes the date of creation formal.

Patents are so much broader, and at least in the case of software so much more ambiguous, that it's very hard to say what is protected and what isn't if you are not a lawyer who specializes in that area. Thus the old saw about patents being meaningless unless defended in court, and I guess you can sort of say the converse is true: if you have money to defend and your opponents do not it doesn't matter how meaningless your patent is.

 

[Aluvus]

Oracle has claimed that it (due to its purchase of Sun) has patents which essentially any Java Virtual Machine will tread on. This (and a variety of other legal details) give it something like ownership of the language, although there are a lot of strings attached. It has also claimed that the actual JVM implementation in Android contains code that infringes its copyrights (that is, it's a direct copy of Oracle-owned code). Ars has a quick overview of the main points.

This is really not an uncommon situation; the 802.11 standards (Wi-Fi), for example, are covered by a variety of patents held by the various parties to its development. As part of the standardization process, IEEE now essentially requires companies to agree to offer reasonable license terms. Otherwise all implementors of the standard would have to live in fear of a lawsuit. This system is hardly perfect (Microsoft recently sued Motorola over their license terms), but it mostly works.

software patents, etc, are pretty loose however. MS, SCO, etc, have been suing people in the OSS arena for a long time. To dates, I can't think of any software patent suit that was actually held up by the courts (I could be wrong though).

SCO's big lawsuit was over copyrights (which, a judge determined, they did not own). Microsoft has mostly just made noise about its patent portfolio and various open source projects (mainly Linux), though they have sometimes taken action. And software patents have indeed been held up by the courts, as Microsoft could tell you.