3 ms·
> Why, exactly? Imagine you are developing a new language to run on the JVM, and it turns out that to get a 10x performance improvement, it would be really gre
by DougBTX 16y ago
> Why, exactly?
Imagine you are developing a new language to run on the JVM, and it turns out that to get a 10x performance improvement, it would be really great to have this new bytecode instruction added to the JVM. To get it into every JVM, you would have to go through Oracle anyway, so nothing changes there. But if you wanted to distribute your own variation, a hypothetical openJVM+, Oracle is saying that they may well go after you with patent suits.
Good for people happy with the current JVM, looking for stability. Bad for people wanting to bring innovations to the JVM without explicit consent from Oracle.
- cemerick 16y agoBollocks. People in the alternative JVM language space have been hankering for all sorts of VM changes/additions for years, including TCO, fixnums, method handles, continuations, etc etc etc. How many openJVM+ implementations have ever been implemented, nevermind used out in the world? None, AFAIK -- the progress, as it is, happens in the MLVM, which is under the umbrella of OpenJDK. So as I said, the exceptions to this might be highly VM vendors, such as Azul, GAE, and various people that have worked on (maybe deployed, I don't know?) full-stack Java operating systems and/or Java-on-a-chip (for embedded applications I suppose). Anyone developing applications, libraries, et al. are entirely unaffected AFAICT.