Re: [hybi] Protocol simplicity and the "amateur programmer" standard

[Greg Wilkins <[e-mail]>]

I do not think the "we've shipped it so you have to rubber stamp it as an
ietf protocol" is an winning argument.

However, we do have a problem with timescales.

I've previously suggested  that we should give up on the attempt to make the
current version an ietf standard and instead just let the WHATWG get on with
deploying something pretty close to what they already have and they can call
it 1.0.

This WG should then be focused on coming up with a 1.1 version of the
protocol that is an ietf standard and does address the wider concerns that
are more of a concern of the IETF WG.

My reasons for this are three fold.  Firstly I do not think that this WG can
produce a reasonable standard in a time frame that will be acceptable to the
browser vendors (nor community expectations).

Secondly, by getting something out that can start getting wider usage, I
believe that will give more background on which to base our discussions.
Unfortunately I expect that this will result in some painful user
experiences as they find out how much work an application has to do to
realistically use raw websockets... but that will at least drive some demand
for the features that have been rejected to date as too complex.

Finally, I think there are some real significant issues that we need time to
ponder, experiment and solve.

I think this two phase approach has precedent in the way HTTP/1.1 was
developed.

regards