My definition is that 'validity' simply warns interested readers
of the limited conclusions one can reach about an I-D's
relevance to IETF activities once the I-D has passed the
magical 6 month marker. Since work is done in WGs, and I-Ds
foster work, then being WG-centric is a pragmatic definition
of 'validity'. And thus the definition doesn't have to speak to
the issue of whether other non-IETF interested parties exist.
I picked the HTTP WG example because work isn't done in (this) WG,
and Individual IDs may share the "foster work" property (counter
examples exist), hence being WG-centric means overlooking the case
of draft-jaye-, and were the 6 months expiry without exception, of
fixing rfc2109 also.
WGs are not procedurally necessary for a draft to reach publication,
so the pragmatism of WG-centric metrics for validity is ... novel.
Cheers,
Eric