Re: ISSUE VARY: Proposed wording

• Messages sorted by: [ date ][ thread ][ subject ][ author ]
• Next message: Koen Holtman: "Re: SAFE, UAHINT and other things not part of HTTP/1.1 now"
• Previous message: Graham Klyne: "Re: Content negotiation requirements"
• Maybe in reply to: Henrik Frystyk Nielsen: "ISSUE VARY: Proposed wording"
• Next in thread: Koen Holtman: "Re: ISSUE VARY: Proposed wording"

Henrik proposes:

=0=0=0=0=0=0=
HTTP/1.1 origin servers MUST include a Vary header field (see section
14.43) in any cachable response based on server-driven negotiation and
SHOULD include a Vary header field with a non-cachable response based on
server-driven negotiation.
=0=0=0=0=0=0=

While I agree that no harm is done if an origin server sends a Vary header
field with a non-cachable response, I cannot understand why a stronger
requirement (ie, a SHOULD) is asked.

I think it means that such a header could be useful in some way, but
since the resource cannot be cached a negotiation has to be done anyway.
Even section 14.43 of RFC2068 does not tell me much. Could anybody
enlighten me, please?

ciao, .mau.

• Next message: Koen Holtman: "Re: SAFE, UAHINT and other things not part of HTTP/1.1 now"
• Previous message: Graham Klyne: "Re: Content negotiation requirements"
• Maybe in reply to: Henrik Frystyk Nielsen: "ISSUE VARY: Proposed wording"
• Next in thread: Koen Holtman: "Re: ISSUE VARY: Proposed wording"