Java has been on life support for a while now. The majority of the "innovation" over the last few major(!) releases has been playing catchup with other language

Java has been on life support for a while now. The majority of the "innovation" over the last few major(!) releases has been playing catchup with other languages, other virtual machines, and other bytecodes. If C# hadn't played the Firefox to Java's IE, I'm not convinced that many of Java's linguistic advances would have ever seen the light of day. Perhaps the single greatest contribution of Java has been limiting the impact of poorly-written Visual Basic applications on the business software landscape, though the alternative it provides is poorly-written Java applications, which in addition to being slow, unmaintainable spaghetti code look visually distinct enough from each platform's native widget set to be jarring. I have little experience with mobile development so perhaps this is a monumentally bad idea, but a set of well-defined APIs, a clean windowing and widget layer with its own APIs, and native bindings for a handful of the more popular languages (Python, Ruby, C++) plus perhaps one or two more specialized languages the platform developers want to support (Clojure/Scheme/CL, Go) would provide enough flexibility to let developers do what they want. (Although come to think about it this seems to be the approach that Maemo used, and that platform never saw the uptake I thought it deserved. So I'm probably relying on wishful thinking rather than actual reality.)