I always got the impression Java the language was copyrighted, based on the hoops I had to jump through to get a copy to install.

You never, ever install Java the language on your computer, just like you can't install English the language on your computer.

What you install in the case of Java is the runtime environment (virtual machine and bytecode for the class libraries) and maybe the SDK (compiler, debugger, maybe IDE). Languages only exist in people's minds.

That is a very important distinction. Software is copyrightable, but languages aren't.

Now Oracle tries to lump APIs (which are abstract things and also exist purely in the mind) into the same category as software, and claims that Google infringes their copyright by reimplementing some of the Java core APIs.

(Oracle also has some other claims, like that Google copied 9 (!) lines of code, claims that Google copied documentation, and two patent violations. But those claims are all pretty minor, and won't affect the software world as a large nearly as much as the API copyright claims).

Lets say Google Search algorithm was broken down and somebody writes API's callable to crawl, index and search information.

Will Google still claim that API's are not copyrightable and its perfectly ok for some one to run a Google search engine clone?

Anybody would object to a API clone, if that API is the very revenue source of their business.

Lets say Google Search algorithm was broken down and somebody writes API's callable to crawl, index and search information.

Note thare are lots of APIs already for crawling the web, in fact there are a few on CPAN. There are also APIs for indexing and search (I quite like Lucy).

Will Google still claim that API's are not copyrightable and its perfectly ok for some one to run a Google search engine clone?

You do realize that there are other search engines out there, and the google does not sue them? Try duckduckgo for example. Or bing. Or yandex (if you happen to like Russian sites), or Baidu (if you happen to like Chinese sites).

If somebody wrote a clone of google search in the sense of using Google's logo on their page, they would infringe trademarks, not copyright on an API.

I think Google also has a patent on their PageRank algorithm, but the API and the ranking algorithm underneath are completely separated.

Anybody would object to a API clone, if that API is the very revenue source of their business.

I don't know a single company that sells API licenses at the core of their business.

Google sells ads on search result pages. Oracle sells software, trademark licenses (so that others can call their java implementations "Java"tm). Sun did too. None of them make money from APIs.

Perl 6 - the future is here, just unevenly distributed

But doesn't the phrase "owner of Java" imply that?

Me calling Oracle the owner of Java has no legal significance. It's uncontested that Oracle owns the trademark on the name Java. It owns the copyright on their their code, libraries, binaries, etc. That is what I meant what I colloquially called them the owner of Java.

I wonder if the fact that it's Java will strengthen Google's case, because Java is its API.

That seems to be a major argument of Google's.

You can do a lot with Perl without ever using a module

Any idea how unique that is? Assembler, C, C++, Java and bash separate grammar from functionality. I believe LISP, Scheme and Haskell do too, but I really don't know them. Pascal and VB do provide at least a print statement.

I think Perl would prefer to have the two separate. In the development of Perl5, a lot of emphasis is being placed on features that give the ability to extend perl without changing perl itself.

Ok, so Oracle owns the copyright on the code, but not on the name of the language, and that's why the judge says it's questionable whether "languages" can be copyrighted? That seems like hair-splitting to me, but I guess that's what courts are for.

I didn't mean that Perl is unique in separating (or not separating) grammar from functionality. I just meant that if you take away the support libraries -- the modules, which I'm thinking of as the API, analogous to Java's class libraries -- you still have something that's identifiably Perl. This doesn't look like anything but Perl, as far as I know:

next unless $s =~ m|/foo/|i;

With Java, on the other hand, if you remove the classes, what you have left doesn't look much different from several other C-style languages. (To my knowledge; admittedly I haven't used Java in a few years, and was never an expert.) Since software copyright infringement often comes down to how much you copied the "look and feel," it seems like Google might have an easier time making their case with Java than if they were using Perl. But I'm just spitballing; I don't know anything about the legalities or details here.

Aaron B.
My Woefully Neglected Blog, where I occasionally mention Perl.

Ok, so Oracle owns the copyright on the code, but not on the name of the language

Names aren't covered by Copyright law, but by trademark/servicemark law. The latter is consumer protection law. Oracle does own the trademark on Java and the coffee cup logo.

and that's why the judge says it's questionable whether "languages" can be copyrighted?

The judge has said nothing of the kind, much less for the reasons you gave.

The judge simply said there's no precedent either way. "No decision on point that says that computer languages have not been copyrightable."

I didn't mean that Perl is unique in separating (or not separating) grammar from functionality.

I know. I understood your point and I agree with it. Your comment simply raised a interesting question.