- From: Eran Hammer-Lahav <eran@hueniverse.com>
- Date: Tue, 2 Dec 2008 18:36:40 -0700
- To: Mark Nottingham <mnot@mnot.net>
- CC: Julian Reschke <julian.reschke@gmx.de>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>, "www-talk@w3.org" <www-talk@w3.org>
- Message-ID: <C55B22A8.F4A4%eran@hueniverse.com>
(adding www-talk for /site-meta comments) A separate header is easily workable, but will require a different approach to the proposed simplified /site-meta text format. It cannot be a list of Link headers without the "Link: " prefix and still support this kind of extension. EHL On 12/2/08 5:31 PM, "Mark Nottingham" <mnot@mnot.net> wrote: A separate header is by far the best way to do this. On 02/12/2008, at 7:35 PM, Eran Hammer-Lahav wrote: > >> Well, right now there's no extension point; the target does not allow >> templating, and a parameter can't override it (recipients will ignore >> extension parameters they don't understand). >> >> One potential solution would be to state that if a link target that >> is >> not a syntactically valid URI-reference is reserved for future >> extensions (so clients ignore it for now). > > This looks like the more promising option. > >> Another one would be to use a different header, such as Link- >> Template, >> defined in version 00 of the draft >> (<http://tools.ietf.org/html/draft-nottingham-http-link-header- >> 00#section-5>). > > The problem with that is the lack of equal treatment this whole > draft is trying to establish for Link header/element. This will also > stand in the way of the new suggestion to simplify /site-meta to > switch to a simple text document with Link header records (with the > "Link: " stripped). > >> Personally, I'd like to see this move ahead without any dependency on >> URI templates. > > I agree. There is urgent need in getting this spec finalized and the > templates discussion is far from conclusion. But there are enough > compelling reasons to bake into this spec support for such future > extensions. I think the suggestion to ignore non-compliant URIs is > the best option, and I would prefer to see it indicated in the ABNF, > but if not, a simple clarification would suffice. > > EHL > -- Mark Nottingham http://www.mnot.net/
Received on Wednesday, 3 December 2008 01:37:38 UTC