Grenville,
I don't mind that you don't find anything valid in my little bit of our
exchange, that was the point -- WGs, IESG, inexpiry, etc., aren't the
exhaustive sources of "validity".
Incidently, the "isn't revised, and falls into the expired (and hence
not-valid) state" applies to both the HTTP man-mech and trusted-labels
drafts, modulo one wasn't expired even though not revised. Counterexamples
to your "valid due to inexpiry" status check exist, not all with 1 April
time stamps.
Anyway, in keeping with my indifference to fashion, I'll leave validity
to its finders, as the original issue was author's intent.
Cheers,
Eric