First of all I always write ideas down on paper. No layout or anything, just ideas: overall direction, texture style, ambience, that kind of thing.

Then, once in the editor I start the map by making a proof of concept room that will dictate the aesthetics and general direction the map is going to take. I also use this room to create architectural modules I can quickly re-use later on. If I am satisfied with this room I carry on. I have a big dislike of greyboxing a map to work on the layout first, I just don't like doing this. I do that in MP games (like TF2) but never for solo. 

My mapping process has become less organized and more chaotic over the years. I used to sketch layouts, and did so all the way up until Soul of Evil; since then, I just start building shit and when an area has boundaries it's done (ie, nsoe3). As a result of this, once I finish the basic architecture most of my work is still ahead of me. I don't have a lot to say in defense of my style, except that it is interesting to have no idea what an area will be like until it's done. 

I boxed up my latest map before doing any real texture-work just to see how it would go.

Layout is great but the amount of work involved in terms of texturing etc is quite heavy. It's probably a much better way of working though and I will definitely do a couple more maps this way. 

with more info

http://www.celephais.net/board/view_thread.php?id=60386&start=37

My mapping method is usually random, it depends mainly of how the inspiration goes. I almost everytime get the theme, textures and ambience from an image in my head or by something i saw. If that inspiration gave me enough ideas, i have the layout done in my head, if not, time to do some sketches or improvise and think room by room as they come, but improvisation usually leads to very bad maps.

About sketches, they are good to make better maps in general, as you can try things out and get more ideas and polish them faster and i tended to do them all the time, but as usually the layout becomes simpler that way, i am trying to restrict its usage to single rooms or things i don't want to forget.

I rarely block out my maps, only if i need to have some precise size for a specific idea, so they usually go to a half-done brushwork state directly, almost final brushwork in one go in the simpler parts. 

I prefer to just go free form when I make a Quake map, because it's hobby work, and some of the fun is seeing how it will develop. I tend to start with just making a cool looking room, and then build around that with whatever thoughts pop into my head to do around that area. This works reasonably for me, but it can also lead you to dead ends, 'I have no idea what to do here' moments, and lots of inconsistent areas if you aren't careful.

I have in the past done some quick blockout-type test areas. Recently, for my jam2 map I made a jam2_test 'level' which is simply an ikblue wall, the red window texture from knave, a floor, and a red sky... which took me a couple hours of experimenting with different things to get to. Everything from ikwhite to ikblue, to a mixture, just to find something that fit my fancy.

My last Q3 map is actually pretty simple too, but that was also a day or two ahead of time playing with brushes and curves thinking about what gameplay space I'd like to try and make. I also once converted a speedmap into a proper finished map, which is sort of like using a blockout.

When I did level design professionally though, it was all preplanned before setting foot in an editor. There's just no getting by this, you simply cannot get half way through a level and say 'well, I've designed myself into a corner, time to scrap this' while on a deadline.

At Raven my workflow was that there would be a general overview of what story or elements were expected in the map given to me from the design lead, and then I'd need to make a concrete plan for it to discuss in a meeting with project lead, design lead, and art lead. Most often this would be me making a quicky sketchy 'block out' in SketchUp and writing up a page or so of the intended action through the level and presenting it. After this meeting there would be any revisions, re-meet, until the layout was agreed upon.

With the first few games I worked on, they were tile based engines so I'd simply dive into making the level and scripting and the tiles would just automagically update as the art team made them. When we moved into Unreal Engine3 stuff, workflow turned into me making a CSG blockout of the level and handing it off to an environment artist to make pretty and I would do the scripting stuff required for the gameplay before/during/after the artist was working with it. 

Then gameplay, then decorate.

Which is probably the exact opposite of you 'true' mappers. Making maps for me has always been a means to an end, that being the overall games design.

Professionally, I'm a games designer who changed from level design since I'm better at one rather than the other. I tend to do a lot of support for the level designers though, as well as supplying the planning because I've got something best described as 'keyboard diahorrea'.

Back to the point - my favorite bit of Q1 mapping is setting up the gameplay, so I focus on that first. 

I have never planned ahead.

Also I've never released a proper map.

My maps tend to flow according to what I'm interested in during the times I actually map. Most of my maps get done in a weird obsessive state that only occasionally sets upon me. Most often I'm more focused on the architecture at first. Generally I start with a stupid symmetrical pillar.
This was the case in sm169_drew. started with the crate pillar thing. then made the overhead crate conveyor belt thing. then built symmetrical pillars around that. then copy n pasted the other symetrical crate pillar thing. then extended the bottoms and built a slime pit. then built a lame box around it. Ta da!

Other times I decide to focus more on gameplay and barely deal with the looks, or intentionally go for something horrendous looking. Often the gameplay in such maps is even more annoying (http://celephais.net/board/view_thread.php?id=60123). 

How do you design your stuff? 

I've found that when I improvise I have more fun and I build neater stuff, but in weaker maps. When I carefully sketch/blockout a level, I get a tighter design that doesn't evolve or grow as imaginatively. It takes either serendipity or a concerted effort to get both from either starting point.

Whether I plan ahead on paper or in simple brushwork really depends on whether I've had ideas or goals before I open the editor yet. If I have, what I draw on paper is still nothing more than a chicken scratch diagram of lopsided overlapping rectangles connected with vague noodles, to maintain overall organization and provide some foresight into what leads where and what overlooks what.

I'll have the barest notion of theme here ("this is the moat, these are the battlements, this is a cistern, probably") but I'll only let that inform the sketch, and not intrude on it. I don't actually sit down and draw battlements and moats, but having them in mind might make me draw different vague rectangles. :)

In turn I get a kind of guide or framework that doesn't intrude on the construction at all, but then I'll have to build a blockout to test the sketch's ideas in first person. Sometimes I'll find I'm just not feeling it any more and that'll inform a different set of scrawls.

I'm starting to feel like once I like the sketch, the blockout should be thrown away, so I can start building again in the fun freeform way that leads to the best "brush feel." That way the construction and the planning don't intrude on each other at all and I get the best flexibility and results of both approaches. 

I always make the start first and then the ideas flow as the wind, :p but the maps, tend to be to boxy or flat :\ 

Trinca is a terrible mapper.
Probably the worst. 

I'm not planning,
I'm an accident.