I think I hit on a image name structure that will make it easy to know content of image without opening the image.

Going on what I said earlier:

I have a naming scheme that I worked out nicely, fill out as much as possible.
(aircraft type)(identifier)(location)(note)
Note: (location) AF -Air Field, IA -In Air, GF -on airfield with Gear Failure, BL -Belly Landed off AF, WK -Wreck.
ex: Spitfire GBoA AF getting ready for mission

It will be like this:
(Image #)(set ID, 2 letters)(aircraft type)(identifier)(location) the “(note)” part I will discard.
(#####)(aa)-(Do17z)(5U+xx)(AF)

The key is first 2
Image # = 5 digit (or more) sequential code, gives me at least 99999 images in library.
set ID is it ID different scans of same image, as I will very likely take multiple scans of same image, sometimes including back. This will be letter codes as follow:
“12345aa” = full images scan at 600 dpi (maybe 300dpi if subject is very small, then an “ac” is used)
“12345ab” = full image of back side (IF something on back worth scanning. 300 DPI)
“12345ac” = High resolution on area of interest.
“12345ad” = higher resolution on area of interest, and so on.

Now, if there is a series of related images (like same aircraft at same time (likely taken on same film roll))
“12345ba” the first letter of course indicates new image, even though the # is same.
“12345bb”
“12345bc”
Doing that will keep closely related images grouped, and the historical relation is maintained.

OK, why am I doing this?
A few times with other catalogs I made the database / organization programs became obsolete, or the file became corrupted, or had to migrate to new platform, or whatever, and I ended up having to spend a lot of time fixing thing, or starting over.
With this name scheme I have at least some organization even if I loose all the details in a database.