W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2011

Re: Content-Disposition cardinality / priority [was: TICKET 259: 'treat as invalid' not defined]

From: Mark Nottingham <mnot@mnot.net>
Date: Sat, 5 Feb 2011 10:39:37 +1100
Cc: Julian Reschke <julian.reschke@gmx.de>, httpbis <ietf-http-wg@w3.org>
Message-Id: <04CF88E1-688B-4A4D-8A95-9003155A6C5F@mnot.net>
To: Adam Barth <ietf@adambarth.com>

On 04/02/2011, at 7:37 AM, Adam Barth wrote:

> On Wed, Feb 2, 2011 at 8:30 PM, Mark Nottingham <mnot@mnot.net> wrote:
>> doesn't seem like there's much consensus among implementers here; unless one forms, I don't think we can say much.
> 
> As I've written before, this seems backwards.  How are implementations
> to come to consensus if we don't write down a specification for how
> they should behave?


If people with implementations were actively participating in this discussion, we'd have a much better chance of doing so. 

I appreciate the effort you've put into this, Adam, but I don't see them coming to the table. I continue to be uncomfortable taking the specification too far down a road where we're speculatively adding text. 

Remember that we're chartered to *not* break existing implementations, except where doing so addresses security issues or materially improves interop. Error handling of cases like this <http://greenbytes.de/tech/tc2231/#attmissingdisposition3> are hard to see as a material improvement in interoperability, especially when there's only one person on the mailing list actively pushing for them.

What I'd suggest instead is for us to concentrate on making the current draft workable without defining error handling (e.g., making sure the BNF isn't too restrictive, checking the requirements), and simultaneously you can submit your error handling profile as an Internet-Draft. 

That decouples error handling from the constraints of our charter as well as the high bar of the standards track; even if all implementations aren't on board with it, it can still be published as an Informational RFC. And, if you do come up with something that is workable for everyone, I think it could be folded into C-D as it progresses down the standards track.

Make sense?

--
Mark Nottingham   http://www.mnot.net/
Received on Friday, 4 February 2011 23:40:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:36 GMT