From: Seth G. <sga...@li...> - 2001-03-26 01:16:39
|
On Sun, 25 Mar 2001, Joseph Carter wrote: > > They will have to identify themselves as different mods. > > Given that mods identify themselves by subdirectory, just how much of > the directory namespace do you plan to pollute? One directory for each incompatible version. This is why multiple mod directory inheritance would be nice, but it should be enough to specify just two directories - a base directory (i.e. oq1 instead of id1) and a mod directory (i.e. oq1 for standard models and animation code, oq2 for skeletal models and extra animation code.) > Uh, I have yet to see a skeletal model take more memory than a vertex > model. The memory savings is almost always in skeletal models, and the > only case that's not true is when the model does not animate at all. You > lose something on the order of about 16 bytes in that case, I'd extimate. What I meant was that vertex animated models would use lots of memory if you combined the capabilities of both vertex and skeletal animation in a single format. I know you don't want to keep backward compatibility, but it has to be done, and it can only be done efficiently by having two types of models. This talk about "state changes" sounds like pure unadulterated bullshit. Switching between several different model rendering systems is trivial compared to the time taken to interpolate, animate, and render. > > All I'm saying is I'd like to see what happens when can adapt your > > controls to whatever version of up you want. > > Your way has a problem.. Say I point myself straight up (ie, pitch = 90), > and start flying off in that direction.. What should I see? > Say I do it again (pitch = 180, essentially I've turned around and am now > upside down heading back the way I just came), what should I see? > > That doesn't work well.. If you try to make it happen you'll wind up not > being able to see what's in front of you because your camera is going to > be watching behind you because you're upside down. 3rd person view would > almost be mandatory because 1st person would get you TOTALLY confused and > you wouldn't have any idea what direction your keys would move you in at > any given moment.. If you automatically flip the movement axes when you turn upside down, then it's once again intuitive. [Snip section on VIS stuff] Thanks for the info about BSP and PVS stuff. I have some ideas about that, but I can see that it's a bit of a sore point for you, so I'll wait until I have a chance to look into the issue in depth before asking any more ignorant questions :-) > > If you can keep it down to three variations with one merged code base, the > > problem won't become exponential. On the other hand if every permutation > > of all the good ideas spawns a new CVS module, then you are going to have > > an exponential problem :-) > > We already do have that problem. And three different, incompatible > sets of most everything isn't better just because it's all in one > tree. It needs to be one set of everything wherever possible or > things get out of hand. Yep, but I think there are a few things that QF will have to support two or three versions of (such as software and opengl rendering, NQ and QW network protocol.) __ __ _ _ __ __ _/ \__/ \__/ Seth Galbraith "The Serpent Lord" \__/ \__/ \_ \__/ \__/ \_ sga...@kr... #2244199 on ICQ _/ \__/ \__/ _/ \__/ \__/ http://www.planetquake.com/gg \__/ \__/ \_ |