RE: Proposal: i105 Classification for Allow header

Robert Siemer wrote:
> I haven't seen a response to my comment: why is it important 
> or valid under the current charter to remove the possibility 
> to use "Allow" in a request?

The change does not forbid the use of "Allow" in a request. It just
makes its semantics undefined for requests. That makes sense, because in
practice people are not following the RFC 2616 semantics for that case.

This is also a good idea, because it provides an opportunity for another
RFC to define an "Allow" mechanism that actually makes sense. There is
no way to make "Allow" (in a request or a response) generally useful
without a corresponding "Disallow" header, and such an addition is out
of scope according to the charter.

- Brian