Re: Preliminary agenda for Dallas

On 11 March 2015 at 20:15, Cory Benfield <cory@lukasa.co.uk> wrote:

> we should get some concrete examples of problems and limitations before we
> jump down that rabbit hole again.
>

IMNSHO it is the rabbit hole that is the problem.   If/when HTTP needs
significant reworking again (and I think that it will need it sooner that
we'd like), I think we need to find a better rabbit hole!

I do not doubt the good will, technical expertise or dedication to process
of those who were active in the working group, but the fact remains that a
thorough (if a little rushed) application of the IETF process has produce a
proposed standard that contains a lot of good, bad, ugly and unknown
aspects.

The WG has essentially been design by committee when there were multiple
diametrically opposed views as to what the goals and requirements were, let
alone what the best technical solutions were.    Considering the different
view points expressed in the process it is a  near miracle that a rough
consensus was obtained - no matter how flawed it is!

I think the draft is probably the best that can be produced by the WG
acting in the current process with the charter as defined.   While Bob's
critiques raise valid questions, they are substantially on subject matters
that have already been deeply debated (although I think Bob's clarity of
thought/expression would have been very much appreciated!).    I don't see
that restarting the same process with the same charter will come up with
anything substantially different or more importantly any better!

Thus I reluctantly find myself if the the ship-it-and-see-what-happens
camp.   We do need to gather more experience of how aspects of this
protocol will work in the real world and how developers will use (and
misuse!) them.

However, whilst we are all off gathering that experience, I think the
IETF/WG should also be reviewing how a better charter can be written and if
there is anything in the process that could be improved.   Personally I
think that too much design work was done within the WG and that we should
have seen a more competitive tender approach, with the WG
analysing/critiquing/comparing more complete proposals that are submitted
by teams that share an architectural vision.  Ie the working code should
come before the rough consensus!

regards




-- 
Greg Wilkins <gregw@intalio.com>  @  Webtide - *an Intalio subsidiary*
http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that scales
http://www.webtide.com  advice and support for jetty and cometd.