> Then what will happen to custom mapping
> when level designing becomes much more
> complex and way more time-consuming? It
> already takes quite a while for one person
> to build a level, but that one person may
> yet need to learn texturing and modelling
> just to make a map.

It will, as soon as Doom3 comes out. It already is, really, for the Unreal2 engine. Half of any UT2/U2 map is models and extra junk they stuck in - so is Doom3. We're all going to have to turn into modellers to keep up, or start collaborating a lot more often on things, or both. Mapping is starting to turn into modelling: shaders, high poly mapmodels ...

In 1998 we had Quake 2. Next year we'll have Doom 3. Now, think about five or six years ahead of THAT ... boo! :) 

Gamespy last year did series of articles and interviews regarding future of games. Featuring Sweeney, Spector, Molyneux and something else

http://www.gamespy.com/futureofgaming/engines/ 

i'm actually better at modelling than mapping (quite good at the first, crap at the second) so that keeps me optimistic. i dont like animating tho so BOOOOOO! 

the joys of being a coder... 

There's joy in being a coder? 

Us left-brained people don't have to do any real work. It doesn't even matter what end of the body it comes out, managers discard it equally. There's a bit of satisfaction in being permanently doomed but assured it will never get any worse. :)
Dilbert type senario. 

if you're a masochist. ("What new and exciting things are broken today?" et cetera) 

This is where part of the discussion appears to be heading.

Now, I've been playing a couple of Doom/Build engine games recently, but there, instead of textures being complete images, they're constructed from piles of "patches", or even "decals" (like hazard signs.)

The impression I'm getting is this:

In the future, texture-specific shader files can be overridden by surface-specific shader files.

The question, of course, is whether this is a good idea or not... 

Shaders are going to have to be texture specific to a certain extent (if we're ever going to get any serious shadering done) because the textures are mostly responcible for determing material. Concrete doesn't damage the same way that iron does. And steel doesn't weather the same way copper does.

What I see as most useful is textures defining a series of characteristics about themselves. Then the mapper would say "scraped damage here," "scorch mark here," "water damage there" etc on a per surface basis.

That work share work evenly between texturer and mapper, while putting a lot of it on the cpu. I always believe in making the computer do all the bitch work. 

We really need to get more people that work at game companies to read and post in discussions like this. Since they are the ones making the technology, it would be excellent for us, the community, to bring up points of mapping (tool designs etc) to developers. Developers could also benefit because they could get direct feedback about design ideas or even questions about topics they might not have thought of. This is of course provided they want feedback, keep the discussions from mentioning anything specific that would fall to NonDisclosure Agreements, and have the time to even bother with this minority community.

bleh. 

I disagree with what you wrote. 

http://home.worldcom.ch/~negenter/031AA_Tx_E.html

I'm guessing that most Quake levels are semantic architecture. 

I really hope peej finds a way to at least recover all of the qmap posts/threads for archive purposes -- there were some meaty conversations on there. 

I mostly agree with the stuff Nane talked, but I totally disagree that small embossed textures should be replaced to brushes, because by small sizes there is barely a difference between 2d and 3d, and textures can contain much more details than a light entity can effect an embossed brush. (Example: compare the TECH04_6 texture with a round-cornered TECH04_2 texture.) And even if this detailed-brushes-stuff looked better (but it doesnt), then okay, here you are, you got beatiful brush-rivetted door-jambs and stuff, and you only see the grey void instead of the rest of the map :).
In my opinion there is certain threshold (according to the size of the object) between texture-detailing and brush-detailing, it is not a fix value, depends also from the amount of emboss. This threshold is very easy to find, if one doesnt, and uses textures instead of a big ceiling-decoration made of thick bars or puts small 2x2 pixel brush-needles on a metal surface, then that one is stupid, and needs more experience about Quake mapping. 

of course, you have a point.
otherwise one goes mad in search for detail and each map takes months to do.

there is always a quick and efficient way (or various) and not so efficient ways.

a bad mapper can always use his texturing skills for example. its good that each mapper has his own view.

we are lucky there are so many good texturers around and we depend on them (not me personally but...) 

IMO that puzzles shall be as natural as possible,you can find examples in Halflife. But for Fantasy game s as Quake or Unreal(tech fantasy)you can even a new world and make the players bilieve in the Natrual things THERE. For example you can't jump more than xx meters up,but if the game occurs on another world (matrix i.e.) it's possible to jump that high,so you can make pouzzles like that. 

Dear God. 

Indeed, i now have a team of like 10 environment artists, lighters, texturers, and vfx people, who take my blocked-out gameplay spaces and turn them into fully realized locations. 

I'm doing this wrong... I'm making everything myself!!! 

Dear God

i know, who ends a post with "Boo!"? 

Much worse. It was "boo! :)" 

That's an interesting article with some interesting terms.

Thank you very much for that. :3c