To come back to dmcgeoch's question, I always cheer for N scale if a guy has limited space, almost too restrictive in HO, and wants to run modern equipment. I went through the same dilemma using my study; I don't know how many track and bench plans I drew up in HO before accepting I wasn't going to get my "Empire;" now that same space in N has given me the layout size of my dreams.
If I have any limitations now that I have to live with is this; I can't have the number of engineer/operators I would like to have. The study, now train room, will only hold 3 comfortably however the layout could handle more.
I have used an open helix (not shaped like a circular helix but is L shaped) to have about 65% of the track visible in my "nolix" area. The term nolix was a name John Armstrong gave to an area which was a peninsula, used as an open helix so that trains could gain elevation in a believable manner, yet the majority of the track was visible. The first plan John A drew up for the client had three helix's and the client didn't want "no helix's" hence the word "nolix."
I have two decks separated by 20 inches, and I plan to install another deck underneath the first deck for staging only. I will probably semi-scenic the visible areas as a desert region just to make the staging look more believable.
I have something like 120 feet of main line running in a study that is probably smaller than dmcgeoch's area. To get 120 feet of main line running in an HO layout, you'd need substantially more room.
If I was into HO and had dmcgeoch's room, I'd probably do a switching layout with the main line running through the switching area. I'd have it built with a "ports" theme, with a rail barge, interchange track, and a bulk transfer station (modern term for "team track"). I'd probably have a helix with less than 30 inch curves to either access a second deck beneath for staging; or I'd use the helix solely for staging as a friend of my does, double tracked; I'd probably triple track it.
Dmcgeoch's question is an excellent one. Most of us start out with a scale then fit - or try to fit - the scale into the area we have. The smarter approach would be to start with the area we have, then chose the scale to fit the area.
I have another version of Ian Rice's perspective: reduce visibility and the train will seem longer. My observation is this: when running trains realistically (or trying to approach it), the imagination plays a larger role in realism than the layout. Since this is true, it doesn't really matter the scale size, because the imagination is larger. Let me explain; I learned this lesson running a light "op's session" on a friends G scale "garden" layout. I say garden layout because technically that is correct, but his is more a "forest" layout with no garden at all; this makes a great layout by the way and way more believable.
My friend invited me and my son because he knew we would be the most likely to run his layout in a realistic manner, unlike the beer and barbecue running manner of many G scale layouts. So the three of us ran our trains, switching cars in and out of the train pool, etc. Most of his layout, like many of ours wasn't completed, so we had to use our imagination in fixing locations or industries or stations. Running on his layout reminded me of my son's and I's first 4 by 6 plywood central RR; our imagination was larger than the layout. Subsequently on other's layouts, I've found that a healthy dose of child like imagination increases quality time in "ops sessions." In fact, I've found the main ingredient for running on anyone's layout turned out to be the "inner child."
Now to N scale structures: I've found a great deal of myth in this area, usually perpetrated by experts who have never "worked the field." I remember attending a clinic at the N scale convention in the state of Washington put on by a well known chap whom I unfortunately can not recall his name, but he had a number of published articles in the N scale rag about scratch building. His "thing" was to encourage folks to scratch build and scratch build prototypically using the correct scale of "lumber" or whatever the building material was. Now the myth is that if you want to scratch build a lot, HO is the better scale. Horse Hockey! I, who have never scratch built a scratching post, attempted to build a modern Alberta granary whose plan was in an MR about 6 years ago after this chaps clinic. I was surprised at how easy it was; maybe easy isn't the right word but I thought if I can do this anyone can; when I built my first bird house for cubs when I was but a cub, they asked what it was I was submitting for the contest. I still can't cut a straight line with a circular saw, or hand saw for that matter.
Now to buildings in N scale: they should be much higher than commercially available. We need a certain size for something to make an impact on us humans. HO structures are generally one or two stories high; maybe 4 - 6 stories if modelling a town. You need to double these sizes for the same impact in N scale. Small structures are really small in N scale, so why model those? Companies like Walther's put out a structure in HO which makes sense, usually one or two stories high; and then they make an N scale copy of this structure for the N scale market; often too small in my opinion. If your building a yard office, why not build it up to 4 stories, prototypical in larger cities, rather than a smaller pathetic one story structures.
In urban areas for N scale, visually having 8 to 12 story building will really enhance the impact of the buildings - heck why not a 20 story structure. Getting these kinds of buildings for the modern era is really difficult in N scale as again the structures available are really made for the "transition era" HO layout. Heck if I never see a red brick structure on a layout, it will be too soon. So most of the buildings have to be scratch built, but after doing one or two, the N scale hobbyist will find it really isn't that challenging; and Hey!, you HO friends will be impressed. I like to tease my HO brethren by saying - "Heck any wimp can solder track in HO, try in it N! - actually, it really isn't difficult but I'm not going to let them know that.
Good luck on the layout, dmcgeoch!