The gods willed it!
As a child, I had an American Flyer S-gauge tinplate set, and it was my pride and joy. A neighbor's kid had an O-gauge Lionel, and of course they are incompatible, except that we found a way to make them so (there really is no restriction on the imaginations of little kids). Later, when my brother asked for a set of his own, I lobbied to get an S-gauge set so that his would work with mine; but, I was out-lobbied by the local hobby-store owner, who convinced my parents to go with HO. So, at that point, the neighborhood consisted of three gauges, with my brother's Athearn set being closest to anything scale.
Well, the neighbor also got a new HO set (Tyco), and then a second hobby store opened in town and had the usual give-away promotions. I entered and won an HO train set of my own, and that was the death knell of the tinplate. I eventually traded it for more HO equipment -- not anywhere near the value, but it was compatible, and it worked (and I still have some of it).
In those days, at first, N was out -- it didn't exist. If you wanted smaller than HO, you went with TT. N first appeared around 1964 and was called "OOO," but it was almost all European stuff with huge flanges and just not very appealing. So, I never really considered N.
When I was in eighth grade, every boy had to take a year of shop. If you were a girl, they taught you how to sew and cook -- no sports, no tools, no fun if you were female. But, I was the right sex, so I was in the shop, which for me was a quarter of metal, electrical, drafting, and wood (in that order). I got an "A" in everything except metal shop, where I got a "B" from Mr. Hill. The school library had a subscription to Model Railroader, and it was full of all sorts of neat stuff to build. Mr. Hill wanted everyone to do a metal project, and in his mind, that meant making a hammer or tin box or something similar (the shop was equipped with, e.g., bending brakes, riveters, and even had its own forge). I took the time to trace plans for an HO Forney 2-4-4T -- I didn't know any better. Of course, things like that are made from brass, but I didn't know that either. So I set about trying to make it out of galvanized steel with a forge-heated soldering iron big enough to use as a murder weapon, which was a waste of time. But, you learn from your mistakes.
Mr. Hartley's electrical shop at the time was more fun. Linn Westcott was publishing a lot of stuff on transistor throttles, so of course I wanted one of those also (another frustrated project which failed, this time for want of funds -- even then, transistors and diodes were expensive, especially for an eighth grader who was shaving his lunch money). But, I found a $5 bill on the ground walking out of shop one day, and believe it or not, I came close to pulling it off (and eventually did build a TAT IV, which worked). Westcott always told modelers never to not do something because "I've never done that before," and I took that advice to heart. You really never do know if you can do something until you actually try it. At that age, I didn't really know what transistors were or how the worked; but, I could see what they could do, and considering that everyone else in the class was learning no more than how to solder a couple of wires together, it wasn't hard for me to get a top grade.
Of course, with my plans for the Forney, drafting was a cinch. And woodworking was learning to make cabinets and the like. I was OK at that -- got another "A" -- but what I built at home was Linn's L-girder framework for my first layout on a side of the garage. My dad, who never really mastered which end of the screwdriver to hold, was amazed.
All that was 40 years ago, and I really never have changed all that much. That first layout had fully interlocked three-color signals using a circuit described in October 1964 MR (use the contacts on a Kemtron switch machine to bypass the twin-T from "C" to "P" to get the interlocking effect when the turnout is adverse). A guy by the name of Fyffe described how to make a snap-action power supply for the switch machines, and of course I had to have one of those too. The local radio shop didn't have any 8000 microfarad capacitors, so I hung four 2000 microfarads in parallel that could throw the points of a single machine into the next county. I spent a lot of time adjusting the turnouts (even code 100 can't take that for long), but at least the lights turned color. Today, I suppose, we'd all be using micro-chips, but not me -- simple circuits one can assemble and service himself are so much easier and cheaper, so why bother with the newer stuff?
Somewhere along the line, I picked up an entire set of Model Railroaders -- I have them all, from 1940 to 2001, am missing only those first years. You know exactly what the "right" paint schemes are because the "wrong" ones haven't been invented yet. Over time, I picked up my share of brass locomotives, which (of course) had to be super-detailed, and that meant I had to polish off those old "B" metalworking skills. There were many articles in the early MRs about how to scratchbuild steam locomotives -- you don't see much of that any more. Still, if you want the best, you have to be able to do stuff like that, because it simply is not commercially feasible for a manufacturer to do it for you. And these early articles actually show one how to construct detail such as cab interiors which aren't covered anywhere else. Today, I don't consider an injector complete unless it is drilled for all three control lines from the cab, and that involves drilling #80 and #79 under a jeweler's loop (I don't think Mr. Hill would give me a "B" today). And yes, I finally figured out that one uses brass and not tinplate or galvanized steel.
Would I consider switching scales at this point? I doubt it -- too much invested in HO. Also, I still think that N is too small, although the quality is much, much improved. I cannot imagine myself trying to do in N what I do today in HO.
I suppose I could go up to S or O, and I've long been attracted to garden railroading, except that I don't own a house. And where I live now there are several active live steamers running on a very large oval at the old depot -- those things can really pull!
But, what would I gain? As long as I still can see well enough to use #80 drills, moving to O would do no more than oblige me to put more pipes on the engine. I'm trying lubricator lines now -- do we next insist upon lubricator control lines? A real steam engine has about 15,000 parts in it, all of them replaceable, so how many does a model-maker have to install before he can claim to be a "scale" modeler?
I figure that if I can get current O-scale detail in an HO engine, most will find that acceptable, and I will be satisfied simply because I wasn't able to do better.
So, I guess HO will have to be it, unless I get to the state where I can't see it anymore.
And then I guess I will have to switch.
The gods willing.