Re: [hybi] consensus call: websocketprotocol baseline

[Greg Wilkins <[e-mail]>]

Martin J. Dürst wrote:
> I'm fine with adopting either draft. It's much more important to adopt a
> draft as a baseline quickly and then work on it than to make a big deal
> out of which draft to adopt as a baseline.

This is a good point.... and as it seams that some browser vendors
are implementing -76 regardless of what is happen in this WG, then
I guess the server implementations will have to follow.

Adopting 75 as the baseline will not sort out the mess of deployment
with unversioned different implementations of 75/76 etc.

So if we can move quickly to start the process, then let's go with -76
as the basis of an ietf-00

> Also, while on controversial issue, it is very important that the chairs
> assess consensus and direct the editor, I think we shouldn't take this
> too far and put up every little wording issue for a WG consensus call.
> 
> Also, I would personally be absolutely fine with the/an editor to put
> some textual proposals into a new draft without necessarily checking all
> the details beforehand with the WG, if such new text is editorial only
> or if it is properly marked in the draft (e.g. with NEW PROPOSAL or
> ISSUE or whatever). I have seen quite a few WGs where that worked out
> fine. Of course, I leave it to the chairs how to handle this.

I agree that adding clearly marked proposal in the draft is a
reasonable way to proceed.

My main concern is making sure that we do not force implementers
to all follow speculative changes.   We need some clearer way
of indicating when a proposal is sufficiently mature that we
would expect implementations tracking the draft to implement
it.

Of course having a version in the protocol handshake would
also help servers enormously.  Servers will now be confronted
with a mix of -75 and -76 traffic being presented to them
and will have to infer which behaviour to follow.

regards