> whether these use cases justify the additional complexity,
Working with EDTF is not trivial, which may have some correlations with
the fact that this work have not been done before, even if there has been
a need for it since a long time.
I feel confident that we will manage to formulate a BNF taking into
account most of the complexity of today's EDTF draft.
> I'm really concerned about a workable spec:
I'm too! For the EDTF to be used, it must cover many needs, which in turn
means that most of its users will probably not use all of its features.
The BNF (and maybe future extracting / validating regexes) will help