by Christian Wimmer (University of California at Irvine, USA)
Summary of reviewer comments:
The paper advocates byte code representations specified at the basic block, rather than method, granularity. The growing popularity of JIT compilation justifies an effort to design a bytecode language suitable for a spectrum of languages. The paper goes on to list design criteria for a new bytecode format. This may be insightful to the reader who may learn from this abridged tutorial on what demands are placed on modern bytecode languages. The paper does not yet reflect a broader analysis of limitations of existing bytecodes, including CLR's CIL and LLVM, and lacks a deeper semantic analysis of bytecode limitations.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment