RE: Issue 248: client "Date" requirements

Julian Reschke wrote:
> the spec currently says:
> 
> "Clients SHOULD only send a Date header field in messages that include
> a payload, as is usually the case for PUT and POST requests, and even
> then it is optional. A client without a clock MUST NOT send a Date
> header field in a request."
> 
> (this comes from RFC 2616).
> 
> This is very wrong:
> 
> "SHOULD only .. and even then it is optional".
> 
> So, if it's optional, it's MAY. I don't believe we need to say that
> clients "MAY" send a Date header :-).
> 
> Can we simplify that, or even remove it?
> 
> (see <http://trac.tools.ietf.org/wg/httpbis/trac/ticket/248>)

I read that as "Clients SHOULD NOT send a Date header field in messages
that do not include a payload. The header is optional for messages that
do include a payload, as is usually the case for PUT and POST. A client
without a clock MUST NOT send a Date header field in a request."

I'd be interested to know why they should or not in either situation,
however. I don't really see the use case.


Robert Brewer
fumanchu@aminus.org

Received on Monday, 11 October 2010 19:22:40 UTC