It’s been done
Today I discovered the joy of background research. My research group at Carleton has been working on getting a paper together for VEE 2006, describing a possible language feature that we’d come up with. Essentially, we wanted to add opcode-level macros to Java; potentially for in-memory code-space optimization. We’ve been working on it for the past couple of weeks, checking to see how many repeated subsequences there are in the JRE, if our effort would have any real impact. We were doing the background research in parallel, checking to see if this was an idea that had already been proposed.
Sadly, it had. We just hadn’t run into it until today. At least two groups had thought of this idea in the late ’90s. Both had discovered that GZIPing the opcode stream provided commiserate compression levels, with a relatively small speed hit.
Anyhow, that’s the end of our draft for VEE 2006. And it’s cemented my reluctant belief that Carleton doesn’t have enough people working with the Java VM for me use that as a thesis topic. Barring a miracle, it looks like I’ll be heading back to my work with reconstructing network data.
