On Mon, 11 Aug 1997, Roy T. Fielding wrote:

> >> in the response chain, how about the following compromise:
> >> 
> >>    Follow the existing versioning requirements as-is, except that if the
> >>    request is HTTP/1.0 (and only HTTP/1.0), then make the response HTTP/1.0.
> >
> >and this is for proxy, server or both?
> Both.

Sigh.  This discussion has often been more strident than necessary
and I am probably as guilty as anyone in that regard.  But I wonder
if at this point we couldn't stop, take a deep breath, step back
and assess where we are.

We have a specification which has apparently been in existence and
unchanged since 1993.  It has been one of the most contentious parts
of the specification if we judge by the mailing list.

We now learn that essentially all proxy implementors have incorrectly
implemented this specification.  Indeed, they seem all to have
implemented it in the SAME incorrect way.

Before we add a band-aid on top of a band-aid to the current spec,
wouldn't it be prudent to entertain the possibility that the
specification might be improved in the light of implementation
experience since 1993?  Isn't it possible that communicating strictly
more version information than is in the current header and making
sure that the information is clearly labelled would serve us better
in the long run?

John Franks 	Dept of Math. Northwestern University

Received on Monday, 11 August 1997 08:44:47 UTC