The little codies won’t obey
I did it again. Yesterday I started to whack up some “nice” physics stuff, thinking I’d have it covered quite soon by lifting some collision algorithm code from Real-Time Collision Detection book.
That was a mistake, and it certainly isn’t the first time I have done it. The stuff by itself worked well in single-frame test cases and I have tested the algos beforehand on that level, which fooled me to think this would be quite easy. But making things move and then testing borderline cases, that usually uncovers things to fix which takes a lot of time, and are still necessary to make the whole thing robust enough. So, once again, when in hurry, one should rely a bit more on tried&tested stuff than making something new.
So yesterday I spent most of my day just tweaking that stuff and not working on the other parts. I had built a nice hierarchical level structure, but I was relying that I’d build simple editor stuff for it, so for now it’s more or less unusable as well. After trying to fix the system I just finally got tired (and slightly angry) enough and went to sleep.
So, if I want to continue with the idea, I’d have to integrate some phys. lib in and try to fix the stuff work with it.. and create an editor.. and make some levels.. and all the other polishing. I’d say it’s quite probable that there won’t be an entry for me this time.
I still might tinker around with the pieces for fun though.
(0 days 19 hours 10 minutes left)













