> It loaded the scene pretty quick IMO. Reading the mesh
data from text
> might
> not be that much of a performance hit, and you probably
have to keep the
> secondary UVs in text members anyways. I've found
having many, smaller
> text
> members for data works better than a single big one.
Yeah that's another area I might explore, but I've simply
not had time to
put into making the whole thing build quicker.
As someone pointed out on the shell game a while ago, even
on low quality it
still takes a while to build which suprised them.
It'd certainly not hurt if low quality (and hence crappy
computers) didn't
have much work to do, which is another big reason for me to
switch to using
an actual W3D.
> I am seriously considering Havok again. It's too easy
to wall-hack when
> players fall with my MUR routines (ledges are
particularly troublesome).
> Why did you give up on it? I thought it worked pretty
well with your old
> RTCW demo.
Various reasons.
I remember having to tweak the hell out of it, which annoyed
me no end, and
I never felt like I was in control, I was more kinda
'attempting to guide
the beast'.
It's not really something I like - to base games on code
that I have no real
clue what is going on :s
Other complaints included the rather important (well, for
some games)
non-deterministic nature (it certainly scuppered my online
pool game).
And of course, it sometimes simply not doing what you asked
it to. Getting
caught on 'invisible' cracks was a PITA too.
As far as your routine goes, how does it actually work?
I really don't have huge problems with escaping the world -
I'm sure I can
remove 99.99% of the remaining instances where mine goes
wrong with a
re-write to handle multiple plane collisions.
Yes, there's lots of backup checks (eg if you ever move less
than a certain
amount, it sticks you where you started, and also it
attempts to float you
away from the surfaces at all times, even the gravity ray
check), but then
that's always going to be the case sadly.
The main thing I think my routine does that no-one else'
does is remove the
'volume' check people attempt to do. I treat the player as a
point and
simply expand the collision data accordingly. Less rays,
less chances for
errors, and overall 'simpler' (well, to code - it's a bugger
to set up the
collision data mind).
BTW congrats to almost everyone on this thread, the sheer
level of geekiness
has been a joy to behold 
Barry
gerbil theburrow.co.uk
_______________________________________________
dirGames-L mailing list - dirGames-Lnuttybar.drama.uga.edu
http://nuttybar.drama.uga.edu/mailman/listinfo/dirgames
-l
|