I NEED to adopt some techniques to make things go faster because I get caught up in the visuals way too easily. I spend days tweaking how an area looks and the map doesn't move any further along.

I also need to start making plans beforehand. Even a vague write up of what I want the highlights to be would help. I wing it way too often. 

.. It takes the time it takes, and that's it.... also mapping runtime is around 7-8 months for me, let's say considering I map 3-4 hours a week best case... though....
In anyway, I'm trying to write a plan on the paper before starting something, else it truns out into a nonsense biult all the time... 

Yeah Jago I would say that the attention to detail is a big one. I've noticed more and more that a lot of big name developers lets a ton of visual bugs slide if the gameplay works - I've noticed some (although very few) in UT3, a lot more in TF2 and L4D and even more in the recent Wolfenstein. Things like missing floors, overlapping/z-fighting brushes or models, assets floating in midair when they shouldn't be, etc etc.

Especially in a game like UT3, and it's kind of ironic, but the visuals actually matter the least as players will be frantically hunting down their opponents and not looking at the scenery.

Willem is right about making plans too, because even a loose sketch can help anchor things. Getting the layout is most important early on, then testing gameplay, then when you're satisfied you start doing detailing. I mean, you can do some details here and there if you have a really vivid idea for an area that has this central pillar with like lightning shooting out of it or whatever, but yeah most people won't notice the 8-unit difference the ceiling had before compared to now, and if they do I almost guarantee they won't care. 

I try and nail down the visual theme as soon as possible, having some kind of back story written down seems to help me, but a sketch or conept photo (not necessarily architecture) can do the trick as well.

Basically I find once the imagination is running the map seems to build itself.

The major block is thinking too much for me. Messing around, trying different variations of stuff before going back to earlier designs is symptomatic of not having the concept clear in mind.

Assuming the map gets finihsed this indecision doesn't hurt and tends to make for a more defined visual theme, but its also the biggest killer of a map.

What you're aiming for, in my opinion, is the turning point where you can't imagine scrapping the map because you're happy with how it looks and plays, and too much work has been done - maybe 50% of the map complete.

At that point I tend to relax and just continue filling in the blanks - work on it becomes more relaxed but also more regular. 

I'm sure many mappers have similar difficulties - and many players would benefit too from mappers being able to finish quicker (and would probably be happy knowing the map was palatable and efficient to make rather than being a huge slog).

As a non-mapper, a couple of things I would guess at would be to have clear, strong concepts before starting, so there's something that's going to make the map worthwhile and exciting without having to refine the details to the nearest pixel. And to have some personal styles and features that one is particularly efficient at so you can use those to speed up the mapping, knowing you're going to get it right (obviously there's a fine balance having a reliable style and have a repetitive style, but it might be worth considering - look how much Tiddles managed to churn out). 

Low-fidelity mockups (layout and scene sketches) <=> analyze/test <=> high-fidelity mockups (build plain layout) <=> analyze/test <=> build actual product <=> analyze/test.

If you're talking about just visuals, then a similar approach would probably be appropriate, but a lot of level designers probably don't have the scene drawing skills to properly flesh out the mockups. So they skip straight to building the scenes in the editor, which leads to all the common slowdowns in development because they're trying to cram too many discrete processes into one step. Therefore we see inconsistency, getting stuck on specific details that are unimportant, lack of vision, etc. 

Pick a usable texture for mapping the bulk of the protitype map. Use 'sketch' brushes with this texture to flesh out the bulk of your layout and the major structures in a simplistic form. You can work item placement and enemies and mechanisms with this sketch map. After that, begin replacing the sketch brushes with the more detailed and properly textured real brushes and add your details. This is a good time to work lighting.

I did this for a couple of my Travail maps, the best example being the forth one in episode I which turned out to be my favourite. The first three were less focused and sprawled out of control and had to be broken down into smaller maps and ate up way too much time.

I also used this method for Fall Cleaning (the Halloween map I made) and that one was built in less than a week - a miracle for me.

Hope that offers some hope for reducing map time. The time required to map is the biggest reason I haven't mapped since Travail. 

was really cool. i was surprised how much was crammed into that little house! i never realized it was made in such a short time. :o 

Sounds like the way hl2 maps were prototyped with just a plain orange texture with grid markings. I suppose if you're using a base texture that matches the theme you can at least save yourself a bit of retexturing.

Perhaps fast prototyping like this could be worked on in a series of speedmapping packs. The first pack is 2-3 hours where everyone creates a sketch of their own map, then the next week do a lighting pass/detail pass/entity pass on the same map to build it up. Only problem is given the ratio of enthusiasm to participation in any speedmapping theme, imagine trying to get people doing it for four consecutive weeks... 

This is an issue that bugs me all the time, what usually happens is that I'll sit down with the editor and make a fully detailed room, and spend time getting this 1 room or area looking just how I want it, with all the textures, lighting etc done and dusted.

2 things seem to happen after I get this area completed.

a) I loved the theme I made, and it inspires me to create more of the map continuing said theme, trying to keep the same atmosphere as the 1st room. This usually results in a pretty finished map after a while, then its just a case of dotting the T's and fixing it all up into something that plays well.

b) That was fun, I wonder whats on tv now?! <<sigh>>

I must admit that I am working with the "orange" dev textures in hl2 alot for blocking out maps and creating basic gameplay, as the game and editor are orders of magnitude more complex than Quake, but I see there definitely is merit in blocking out maps before you start going to town on visuals, even in older games like Quake.

I'll have to give it a try! 

I have trouble with this too. My workflow has been like this:

1. have an idea for a cool environment / setpiece
2. build a small area with details, lights, and monsters, which captures the look and feel I wanted
3. try to build a map around this, maybe by stitching together multiple setpieces with filler sections

This leads to poor gameplay (and other things.. layout, consistency, etc), because I try to work around setpieces which I created with only visuals in mind.

I definitely want to try a more disciplined approach on my next map, with planning, building sketch mockups, and testing gameplay before detailing :-)