SPLASH

An increasing number of high-level programming language implementations is realized using standard virtual machines. Recent examples of this trend include the Clojure (Lisp) and Potato (Squeak Smalltalk) projects, which are implemented on top of the Java Virtual Machine (JVM); and also F# (ML) and IronPython, which target the .NET CLR. Making diverse languages--possibly even adopting different paradigms--available on a robust and efficient common platform leverages language interoperability.

Vendors of standard virtual machine implementations have started to adopt extensions supporting this trend from the run-time environment side. For instance, the Sun standard JVM will include the invokedynamic instruction, which will facilitate a simpler implementation of dynamic programming languages on the JVM.

The observation that many language constructs are supported in library code, or through code transformations leading to over-generalized results, has led to efforts to make the core mechanisms of certain programming paradigms available at the level of the virtual machine implementation. Thus, dedicated support for language constructs enables sophisticated optimization by direct access to the running system. This approach has been adopted by several projects aiming at providing support for aspect-oriented programming or dynamic dispatch in general-purpose virtual machines (Steamloom, Nu, ALIA4J).

The main themes of this workshop are to investigate which programming language mechanisms are worthwhile candidates for integration with the run-time environment, how said mechanisms can be declaratively (and re-usably) expressed at the intermediate language level (e.g., in bytecode), how their implementations can be optimized, and how virtual machine architectures might be shaped to facilitate such implementation efforts. Possible candidates for investigation include modularity mechanisms (aspects, context-dependent layers), concurrency (threads and locking, actors, software transactional memory), transactions, paradigm-specific abstractions, and combinations of paradigms.

The areas of interest include, but are not limited to, compilation-based and interpreter-based virtual machines as well as intermediate-language designs with better support for investigated language mechanisms, compilation techniques from high-level languages to enhanced intermediate languages as well as native machine code, optimization strategies for reduction of run-time overhead due to either compilation or interpretation, advanced caching and memory management schemes in support of the mechanisms, and additional virtual machine components required to manage them.