Summary
Sun's Java site, java.sun.com, has an interview with Sun Lab's Misha Dmitriev, the creator of JFluid, a profiling tool that relies on hotswapping for bytecode instrumentation to collect information on the fly.
Advertisement
Sun's Java site, java.sun.com, has an interview with Sun Lab's Misha Dmitriev, the creator of JFluid, a profiling tool that relies on hotswapping for bytecode instrumentation to collect information on the fly.
Dmitriev makes a lot of claims! He claims that "Code hotswapping is probably the most powerful way to address the performance problems of profiling, while still collecting useful data." He also claims that it is "proven" that profiling gets you better software. He says that the the traditional edit-compile-debug cycle should be replaced with edit-compile-debug-test-profile cycle. He has a lot to say about how hotswapping can reduce the performance problems of hotswapping.
The JFluid tool is now a part of NetBeans. Not surprisingly, this all sounds plausible; I'm wondering if anyone has tried the tool and has first hand experience. And can anyone separate the hype from the reality?