For those of you who (like me) are interested in the development of non-Java JVM languages, here are a couple of links that might be of interest. Alex Turner has a longish piece up on tuning the JVM for “unusual uses”; the use case in particular he focuses on is one in which the JVM runs bytecode that has been created from COBOL by an experimental compiler. I suppose tweaking JVMs specifically for running specific non-Java languages is the inevitable next step of the JVM ecosystem. Turner’s techniques involve just invoking compiler flags, but will we someday see JVM’s custom-designed to run certain languages more efficiently? Or would that entirely defeat the purpose of writing code to an all-purpose intermediary layer like a JVM in the first place?Over on Rémi Forax’s java.net blog, there’s an interesting discussion that’s arisen from his post on the upcoming JSR 292, which implements JVM-level mechanisms to support dynamically typed languages. Some of Forax’s discussion focuses on performance of some of these languages, and the comment discussion is in the comments is quite interesting as well, going back and forth on whether having to switch back and forth into “native” Java code to improve performance of certain structures is a good programming practice or a sign of weakness in dynamic languages. Software Development