On Wed, 22 May 1996, Skylar wrote: > On Wed, 22 May 1996, Hades wrote: <stuff deleted> > > If you know anything > > about coding you can do mob programs with spec mobs. One trillion times > > faster and usuaslly a bit more stable. > > I agree, but I think it can be taken a step further. If you've got a > special thats used by more than 2 mobs, and its fairly stable, you might > want to move it to a MOB_ flag and put the code right in mobact.c (that > is, for mob specials anyways, and objact.c if you've gone that route). > > -Sky Yeah, the patch called 'mobprogs' may be a CPU hog..it does alot of expensive string compares in some pretty core places, but even its crude system makes more sense than mucking up your base with world-speciffic code. I, personally, don't care to write 20 specprocs for every new zone a builder decides to write. Especially not when I can give them the power to painlessly give their zone life as they build. If you've got a problem with the way mobprogs handles triggers, think about where it's wasting resources and fix it. World speciffic information belongs in world files; I don't want that shit contaminating my base. For example, a simple mobprog like act_prog p pokes you~ glare $n~ Requires probably 10 string comparisons everytime someone does ANY action in a room that has a flagged mob in it. Yucko on performance. Maybe someone (perhaps one of you college students with lots of time on your hands ;) could rewrite it to be string comparison free. (If mob X is the victum of command #204 (say), it reacts with command #145 with the victum being the same pointer as was passed as the ch in the original command.) Maybe in order to get the biggest tweaks, we need to compile world files. (As it is, I'm educating my builders on the c pre-processor). Hopefully at least some of this makes sense, I've got a few beers in me. (I'd definiately recommend Pete's Wicked Ale to interested parties). Forgive my spelling. My asbestos suit is on. Flames to me, meaningful responses to the group. -Kahar
This archive was generated by hypermail 2b30 : 12/18/00 PST