Perhaps all this a reflection of the fact that BR did not come up with a 'future proof' numbering/classification system back in 1948/49, but rather opted for a simple renumbering based on adding some multiple of 10000 for the 'big 4', and leaving some 10000 number blocks free (some of which proided too many numbers, others too few), and no general reclassification ... one of my idle speculations is what such a scheme might have looked like.
It was pretty future-proof. The 1957 D/E renumbering did not have any more available numbers than the 1948 1xxxx/2xxxx ranges (two fewer actually!), although about 200 diesels and 70 electrics (mainly classes 11 and 76) did not need numbers in the 1957 scheme as they kept their 1948 numbers. It would have been perfectly possible to find suitable number ranges in the 1xxxx series for all diesels, at least in the BR era, without renumbering any of the existing shunters in the 11xxx, 12xxx 13xxx and 15xxx ranges.
For example
Pilot scheme uses the unused range@s
Type A/1 14xxx
Type B/2 16xxx,17xxx
Type C/4 19xxx
Later:
Shunters after 13999 numbered in the high 12xxx range (12999-12808)
Class 14 153xx
"overflow" class 25s numbered in the high 15xxx range (15999-15823)
Type 3 - 17500 upwards (note- we need to find a slot for the Hymeks - maybe somewhere in the 15000s?)
large Type 4s - 18xxx, after the gas turbine locos were gone
Type 5 - 10xxx (plenty of room for Classes 56, 58, 59 and 60 after the early prototypes (100xx, 10100, 102xx, 10800) were gone