Re: WG Review: HyperText Transport Protocol Bis (httpbis)

On Oct 16, 2007, at 4:57 PM, Mark Nottingham wrote:
> On 17/10/2007, at 9:25 AM, Roy T. Fielding wrote:
>> In contrast, I think this part of the proposed charter
>>
>>> The Working Group must not introduce a new version of HTTP, and
>>> should not introduce new features or capabilities to HTTP.
>>
>> goes beyond that and makes a technical decision on what is best
>> for the evolution of HTTP before the WG even starts.  I don't
>> think that's why we have charters.  The WG can make that decision.
>
> I disagree; one of the functions of a good charter is to scope the  
> work, so that participants understand what they've committed to do.  
> It's a very effective tool for avoiding known ratholes.

The scope of the work is HTTP/1.1.

> While the established procedures usually end in the correct result,  
> it takes a lot of time to wade through all of the "I'd like to add  
> a X header/method/feature to HTTP" requests (as we see very often  
> on the http list), and that can impede the progress of the WG.  
> Letting people know that this is not the place that you take your  
> favourite new protocol extension to is a good thing for both sides.

We will get those requests anyway.  We get them right now, without
a group of any kind.  The charter proposal says that we can't do
anything constructive regarding those requests.  Without it, the
WG simply needs to make a decision with regard to how to handle
such requests.  The decision could be as simple as "please defer
any such discussion until after ..." or "we don't have that on our
agenda right now, but if you submit an individual draft we will
consider adding it to our agenda at the next meeting (or not)".

I received the same kind of request for HTTP when we were focused
on HTTP/1.0 "current practice".  John Franks wanted to add a Host
header field.  It was obviously needed, though it took much longer
for the WG to agree on the solution (the original one) than it
should have taken.  Should we have told him that it is out of
scope at the IETF until after the HTTP working group had concluded
(three years later)?

> I'm fine with "should not introduce new features" here, by the way  
> (i.e., I'm not holding out for "must not"). I just think the WG  
> needs some guidance about the proposed scope of the group. I  
> haven't yet heard anyone say that HTTP/1.2 or 2.0 is a good idea,  
> and many have said it's a bad idea.
>
> WRT authentication: I agree with most of what you say; it's just  
> that such work would likely be on a much different timescale than  
> revising 2616. Re-chartering to do it afterwards has been discussed.

Chartering a working group is a long and time-consuming process,
as you know well. Most people don't even bother.  In any case, if
you want to charter the "Won't make any changes to RFC 2616 aside
from editorial bits WG", then go ahead.  If you ask for an
"HTTPbis WG", then I expect it to have a charter of "Revising the
HTTP specification", not "Revising only the things listed in this
charter because we already made up our minds".

Once we open the can, the worms will come wiggling out regardless
of the charter.  I'd prefer that the decision on what to do with
those worms be made by the community of vendors within the WG, not
anyone else.

....Roy

Received on Wednesday, 17 October 2007 01:06:50 UTC