W3C home > Mailing lists > Public > public-powderwg@w3.org > October 2007

Re: assuming HTTP Link will get ratified?

From: Phil Archer <parcher@icra.org>
Date: Fri, 12 Oct 2007 11:52:02 +0100
Message-ID: <470F51D2.10308@icra.org>
To: Dan Connolly <connolly@w3.org>
CC: public-powderwg@w3.org, Harry Halpin <hhalpin@ibiblio.org>, Mark Nottingham <mnot@yahoo-inc.com>, Tim Berners-Lee <timbl@w3.org>

Other interested folk added to cc line.


Sorry it's taken me a week to reply to this. I need to probe a little 

I wonder what evidence for implementation is required? Through Mod 
headers, Apache allows you to set a Link header and Microsoft's IIS has 
its own way of letting you do that too. Rightly or wrongly, I've been 
advising people that this is the most efficient way of adding a link for 
some time [1].

As an example of a widespread implementation, Perl's LWP module makes no 
distinction between a link/rel defined in HTML or HTTP Headers. See, for 
example, [2]. This is the ICRA label test result for an adult site that 
has configured its servers to include the link as an HTTP response 
header, processed using LWP.

As for demand, your link to the GRDDL shows that POWDER is not alone in 
wanting this. Other documents discussing this are linked to from the 
POWDER doc (from Mark Nottingham and Tim BL) [3].

Is that sufficient demand and implementation experience to either get 
this added as an issue in the HTTP draft or for Mark's draft to be 
resurrected? If not, I would be very grateful for more specific advice 
on what else has to be done.

We need to be able to:
1. Apply a POWDER description to any resource, not just HTML documents;
2. Be able to make a HEAD request to a URI to see if it links to a 
Description Resource that may then affect how, or indeed, whether, the 
resource is later fetched.


[1] http://www.icra.org/systemspecification/#apache
[3] http://www.w3.org/TR/2007/WD-powder-dr-20070925/#assoc

Dan Connolly wrote:
> I read...
> "This section makes two key assumptions that, at present, may be regarded as unsafe.
> That the HTTP Link header, defined in RFC 2616 and still in the registry but dropped 
> from later RFCs, will be reinstated through the new draft."
>   -- http://www.w3.org/TR/2007/WD-powder-dr-20070925/#assoc
> You're clearly aware of the current draft of HTTP, since you link
> directly to it...
> http://www.w3.org/Protocols/HTTP/1.1/rfc2616bis/draft-lafon-rfc2616bis-03.html
> That draft has no Link header and no open issue regarding the link header.
> As far as I can tell, there is no reason to believe that spec will include
> the Link header.
> Something like mnot's draft seems more likely...
> http://www.mnot.net/drafts/draft-nottingham-http-link-header-00.txt
> If you need the Link header ratified, you should put energy into making
> that happen: convince implementors to implement it, make noise about
> implementors that have implemented it, get somebody to re-issue
> that draft in the IETF, etc.
> See also...
>   http://www.w3.org/2001/sw/grddl-wg/issues#issue-http-header-links

Phil Archer
Chief Technical Officer,
Family Online Safety Institute
w. http://www.fosi.org/people/philarcher/

Register now for the first, annual Family Online Safety Institute 
Conference and Exhibition, December 6th, 2007, Washington, DC.

Go to: http://www.fosi.org/conference2007/ today!
Received on Friday, 12 October 2007 10:52:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:06:03 UTC