AlanF
can easily produce a starting point list
Yes, but, to my thinking, what you really need is a TOE list, which will likely need a "by year' sub-listing.
See, in 1940 there wer some decidedly soft-skinned "armor" out there--57mm guns on WC54 chassis; 75mm guns in M3 and M5 chassis, and the like. Tracked armor would include M@ & M3 light tanks and M3 & M4 medium tanks.
By 1943, that TOE is significantly different. The light tanks will include the M5, several varieties of M4 would be in existence. Additionally there are several SPH, the M8 armored car, and even the M36.
Two more years muddies the waters even more, with M18, M24 and M26 just in tracked armor. Several armored cars, and the M10, and SPH of several sorts.
That, at least would give a place to stand.
Then, you'd have to sort out this difference between 'distributed" and "available" (OOP is not, per se, unavailable). But, the distributors are a bit at the mercy of the manufacturers, too. Which will be a huge issue in the database.
As a for instance. The M3 series of half-tracks is fairly extensively kitted. However, a third of all half-track production wa the M5 series, by International Harvester. Which used just enough variation in style as to be visually obvious (and not easy to scratch an M3 over to look like). There has been a resin kit of the M5, but it never seemed to come to fruition. Dragon (IIRC) showed some CAD previews, but whether or not it's in production is a different question.
So, yes, the database could be built, But, I'll wager you, you'd need two others to track "being kitted now" and "available by any means." With probably a thrid database just to track reviews (consider that there are two Tamiya M4A3E8 kits; the reviews for one are very different tham for the other).
Would be an elegant database if created; the lookup tables would be fascinating to build, too.