The source has been reverted back to the Pygame version. The spectre of performance continues to hang over my head though. I am beginning to plan the switch to C++.
Back to normal design issues, the switch a couple weeks ago to keeping a cell change list broke vision, since visibility propagation was handled through the CA engine. Today I worked on fixing that. Another bug has come to light (ha ha) in the process, which is causing some out-of-vision cells to show incorrectly. Fixing that is the immediate objective.
At some time we all have to grow into a more efficient language to do what we want done. Just be glad you are not in the area when that required assembly level code.
ReplyDelete-C
Stupid idea, but you ever consider ditching tracking the world yourself and putting it into a SQL data store?
ReplyDeletePutting it into a data store won't really solve anything, since so far all the worlds kept in memory anyway. You can't get more efficient than that really without going to C-style arrays -- which, need I remind you, I probably will be going to, when I convert the program to C or C++. The problems with Python compilers and graphics libraries have gotten me to thinking about doing that sooner rather than later.
ReplyDelete