On 2011-07-11 16:02, The IESG wrote:
...
This is a set of comments on how the specification defines new HTTP
header fields and the upgrade token.
1) In the IANA considerations section, please group subsections by
registrations and new registries.
2) "Specification document(s)" should actually point to the part of the
spec that *specifies* the thing being registered. Just saying RFC nnnn
or "this document" wastes the reader's time.
3) That being said, the header field registrations each contain a
paragraph of prose explaining the header field. *Is* this the
specification of the header, or does it only summarize more information
from somewhere else? If so, from where?
4) Each header field definition needs to say what the header means in
general, and what the syntax is (the latter can be either prose or ABNF,
but ABNF would be preferred). It's not sufficient to describe it as part
of a Websocket-specific algorithm - what's needed is a statement about
what the header means on an HTTP message in general (and yes, saying
it's only allowed in this or that context, and that it's meaningless
otherwise would be ok).
Best regards, Julian
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf