Re: [hybi] Random Responses

[Greg Wilkins <[e-mail]>]

Jamie,

excellent information and insight in this and your other messages.

I think it is very valuable to look at existing protocols like
BEEP and to identify/enumerate their deficiencies. Straw men
are useful (and others probably appreciate not having me
continue to use websocket for that purpose :)

I've tried to capture a lot of this recent conversation on the
wiki: http://trac.tools.ietf.org/bof/trac/wiki/HyBi

I've added a new section called "Candidates Requirements" where
I list/describe the attributes of a protocol that people have
identified as good (although there is no consensus if all are
indeed good and/or needed).

For each proposal I've listed some pros and cons, mostly
against candidate requirements.    It would be good if others
could sanitize those for any biases coming from me.


I've also added a "Asynchronous HTTP" proposal based on the
description you gave.   You allude to existing proposals for
such a thing - do you have references to those?



cheers