[tei-council] we don't no need stinkin' classification

[Syd Bauman]

> Then the remarks should say that, surely, and the current remarks
> are not only confused but plain wrong! but did we agree that
> data.enumerated *necessarily* implied the existence of a valList?

Good question. I don't think we've ever really decided that
data.enumerated necessarily implies the existence of a <valList>, but
I do think we've stated, multiple times, that we *encourage* and
perhaps even *expect* there to be a <valList>. But either way, we've
certainly never imagined any other way of documenting a
data.enumerated, so the remarks should say that. This is a bit
complicated because, in order to get the attribute an element just
needs to declare membership in att.typed. But in order to properly
declare and describe suggested or permitted values the element needs
to have an <attDef>.