Monday, December 15, 2008

A difference of paper

(A line of trees on the north side of the Kawauchi campus classroom laboratory building)

...And the paper reading continues. Lately i've been reading about Jikes RVM to gauge my interest in virtual machine/compilers research. While I've had a vague interest, I haven't exactly been hacking around in the code. I thought I would read some of the papers, and see how they compare to Programming language/type theory papers.

So far, they are a bit easier to understand on a first reading. Unlike many type theory papers, a typical breadth-first reading of compiler/VM papers seems to work, and I can selectively choose what parts to skim and what parts to digest. Unfortunately, many PL theory papers prevent this at a meaningful level, because they are highly linear. You won't understand the proofs if you haven't read the introduction and development thoroughly.

There are some similar things in VM papers. For example you won't know how to evaluate the graphs near the end until you can look at the algorithm and see what may be the cause of differences. But in general, an illustration representing a dependency graph is more readily understood than a page full of small-step semantics and judgements full of greek (which may or may not be greek to you).

I'll start hacking into the code probably tomorrow or on Thursday, and see what it's like to hack in a JVM written in Java.

In other news, the 忘年会 (bounenkai, or end of year party) is in a few hours, so I had better get ready.

No comments:

Post a Comment