I agree with Paul- to a point.
But first, CONGRATULATIONS to the httpbis working group, Martin Thomson,
and Mark Nottingham in particular. These people have worked quite hard
to produce a specification.
There was some discussion in the working group about including various
versioning bits in various places. At the end of the day, my
recollection was that the group chose not to include versioning WITHIN
the protocol because they would prefer to rev the protocol instead.
After all, ALPN ids are cheap. While I doubt that any of the current
developers really want to think past HTTP2 today, there are built-in
deployment assumptions in every protocol, and on-the-ground
considerations will dictate when and how to update. The developers in
the WG today are in the best position to make this call, and keeping
them together in some form would be useful (be that a WG, a mailing
list, or some other construct).
Eliot
On 2/19/15 10:32 AM, Paul Hoffman wrote:
On Feb 19, 2015, at 10:09 AM, Sean Turner <turners(_at_)ieca(_dot_)com> wrote:
On Feb 19, 2015, at 10:16, Michael Richardson
<mcr+ietf(_at_)sandelman(_dot_)ca> wrote:
I propose that this document skip PS, and go straight to Internet Standard
to
accurately reflect the status of this document.
Six months after it gets an RFC# I’d completely support this.
Good god, no. HTTP/2 is quite complex, and it is likely that at least some
parts will turn out to be non-optimal. Please give the HTTPBIS WG at least a
year to shake out the protocol after wide deployment and constant use.
Rushing the WG just so we can feel good about slapping a near-meaningless
feel-good label on the spec is not a good process.
Counter-proposal: we let the people closest to the protocol, the WG that
created it, decide when to ask for STD status.
--Paul Hoffman
signature.asc
Description: OpenPGP digital signature