> If it were just spelled 'det' the vowel would start jumping around. You'd have to go 'dett' or even 'dette' in order to avoid 'deet'.

Huh? None of this makes any sense.

There are no markers in 'det' that one can use to figure out whether the vowel is long or short. 'ee' is always long, and 'ett(e)' is always short.

English has far more vowels than letters. 'ee' is an old digraph that always afaik means long-e, and ett(e) is a french borrowing that will always indicate short-e. 'deat' would work for long-e, too, probably, although 'ea' is much squirrelier (turns into a couple of diphthongs.)

> There are no markers in 'det' that one can use to figure out whether the vowel is long or short. 'ee' is always long, and 'ett(e)' is always short.

For one thing, the evolution of the pronunciation isn't going to take the spelling into account unless the word gets a lot rarer.

But even if it did, there's only one way you could pronounce "det". It would rhyme with bet, get[1], jet, let, met, net, pet, set, vet, wet, and yet. The spelling is fully unambiguous. This is the paradigm context for distinguishing "long vowels" from "short vowels" - det must be short just as dete must be long. You could no more indicate a "long E" with "det" than you could indicate a "long A" with "mat" or a "long I" with "hit".

> although 'ea' is much squirrelier (turns into a couple of diphthongs.)

What? What diphthongs? As far as I can think of, the closest you get to a diphthong with -ea- is as in "near", which is still just /i/ plus whatever you get as you reposition your tongue.

[1] "get" is such a common word that its pronunciation has evolved in some nonstandard dialects. But of course it hasn't moved to the FLEECE vowel; it's moved to KIT.