Forgive me for circling repeatedly around this point, but the possibility that the Java Virtual Machine might long outlive the Java language itself is becoming increasingly intriguing to me. More interesting posts are coming out of the JVM Language Summit, including this longish one from John Rose. Rose works on the Da Vinci Machine Project, which is specifically tasked with making the JVM more hospitable for non-Java languages. Rose has this tidbit in his report: “Neal Gafter explained how, as the JVM supports new languages, the Java APIs risk being left behind on the wrong side of a semantic gap, between Java (circa 1997) and the consensus features of the new languages, all of which include some sort of lambda expressions (closures). They also often include richer dynamic dispatch, proper tail calls, exotic identifiers, and continuations. This could be a stressful change, if it requires significant retrofitting of Java APIs. We can do this: A similar stressful change was the retrofitting of the standard Java libraries to generics, and I think that worked out smoothly enough.” Now, the demands for closures in Java has been rumbling along for quite a while. It would be interesting indeed if other JVM languages start driving extensions of the Java language, rather than the needs or wants of Java programmers. Of course, there are any number of Java developers who’d love to see closures, no matter how they got into the language — but agendas might not always be quite so much in tune. The movement to develop non-Java JVM languages seems to have grown organically but has hit something of a tipping point lately — and it’s possible that languages will be competing for attention and pride of place within the JVM at some point soon. For yet more JVM Language Summit summaries, check out DZone’s coverage. Software Development