Well, we don't accept *didn't could/might/should/would, and some don't accept ?didn't ought. That seems to be one reasonably sound reason for rejecting didn't used.
I agree with you there. That is definitely a solid point for the opposing camp.
As I see things now, while the spellings
didn't used to and
Did you used to rightly acknowledge that
used, not
use, is the verb in play in modern English—even if this semi/quasi modal etymologically evolved out of a historical full/main/lexical verb
use (separate, obviously, from the
use in
They use chopsticks to eat,
He uses too much salt, etc.) which used to have a full range of tenses along with a parallel meaning—they are nevertheless grammatically anomalous for the reason you have given.
I attribute this anomaly to present-day English speakers' discomfort with using contraction (
usedn't) and subject-auxiliary inversion (
Used you to . . . ?, What/why/where used you to . . .?) with quasi/semi modal
used to. To compensate for our discomfort, we add do-support in the (semantically redundant) past tense, and wind up with the syntactically messy situation of dummy
did sitting alongside a modal on the same branch of the Chomskyan tree, I suppose via head-adjunction.
I'm glad you've brought up
?didn't ought, Piscean, which I wasn't aware existed. By saying that "some don't accept" it, you seem to be implying that many people, perhaps including you, do accept it. One realization that the thoughts which this thread has occasioned have brought me to is that
used to has a syntactic sibling in
ought to, which I now consider its closest grammatical relative, purely in terms of its syntactic behavior. There are interesting differences, too, of course.
For you, Piscean, is it both possible and correct to say
Did he ought to do that? as an alternative to
Ought he to have done that?