My first ever serious layout (ie not on the carpet) was in British N Gauge, built (or, more accurately, started) when I was still at school, as a joint project with a friend that lived 'over the back'. It didn't get that far, if truth be told, mainly because pocket-money weighed down aspirations more effectively than we'd imagined possible. We did have a fair bit of fun with a motley collection of Graham Farish pannier tanks, Peco wagons and both Minitrix and Lima coaching stock, though, and the papier-mache scenery at least was unfettered by financial constraints.
Ultimately the project foundered, the dreadful quality of the Graham Farish panniers with their habit of shedding coupling rods every few minutes was the final nail in the coffin. But I've revisited N gauge quite a few times over the years, both British and (more productively) American outline. Unfortunately the poor running-quality of British N has always killed my enthusiasm pretty quickly, along with some of the most spectacularly awful looking models ever to be released - if the Graham Farish Class 56 isn't the least realistic UK diesel model in the history of the world, then I don't know what is.
I've always said that if decent RTR British N was available off-the-shelf then I wouldn't need 4mm at all. My build it/detail it requirements would be satisfied by my 7mm dabblings and the serious business of operating a prototypical model railway could best be accomplished in N. The two extremes are well illustrated by the above photo - my 7mm Class 37 next to one of my N gauge examples.
The drawback with my 'perfect modelling world' scenario is that quality, off-the-shelf, N gauge stock in sufficient quantity to build even a modest layout is still a long way off. Although Bachmann took over the Graham Farish range some years ago, developments have been few and far between for anyone impatient enough to want a decent layout before they celebrate their centenary. We've had much finer, blackened wheelsets on new rolling stock, giving an effect much more pleasing than Lima 4mm wheels on code 100 track. We've also had a substantially improved chassis design under a good number of models now, and it runs well enough for me to consider N a viable option. But too many of the models still look awful - modelling the D&E years without a decent Class 56 and 47 isn't something I can easily contemplate.
I do have a plan for a small N gauge project, though, based around the limited range of useful products that might reasonbly be available in the short term. This is intended to be a busy urban terminus, intensively worked by DMUs and the odd peak-time HST. The whole thing is designed around the Graham Farish Class 170 and Class 158 DMUs, unfortunately both no-shows so far which kind of scuppers the plan.
Obviously I wouldn't consider building a layout these days that wasn't DCC controlled, but with this one I'd planned to go a step further. Current thinking is it would be completely computer-controlled in order to act as a self-contained demonstration layout for the principles involved. Who knows, it might even get built one day. But, having waited over a year for these promised new Graham Farish DMUs, I'm not currently holding my breath. It'll happen when it happens.