Day 60 - Monday August 19th
Today, I was running more experiments while also looking into the jpf-core version of the Native Peer for java.util.Random(), which of course I was not aware of last week! From the looks of it, it would not have been particularly helpful for the current translations, as it is almost too powerful for the work around that was implemented. But this of course might be better. I spent some time looking at it, and seeing if I can use a modification of it to make translations even more straightforward, or to require to no translations at all. That would be really nice. No progress there yet!
(And I know I said results today, but I was distracted. They will be ready very soon!)
Day 61 - Tuesday August 20th
Today, I was looking at the jpf-core implementation of Random again. I haven't made a whole lot of progress, but I would really like to get this version working, with at least the simulation aspect of the code. The idea of not having to translate many Java programs is just too irresistible, and I feel it will make the the code much more easily usable. I think the problem is that the class tries to enumerate all random choices, and this causes the run-time to sky-rocket, although in most cases we only need a small subset of all possible random choices. And I don't yet know how to control this behavior. I spent some time cleaning up the code just as a change: there was a fair bit of dead code, old comments, and not-commented code, so I cleared many of these issues.
No comments:
Post a Comment