W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2008

Re: I-D ACTION:draft-nottingham-http-link-header-01.txt

From: Phil Archer <parcher@icra.org>
Date: Tue, 06 May 2008 12:38:09 +0100
Message-ID: <48204321.6080204@icra.org>
To: Julian Reschke <julian.reschke@gmx.de>
CC: Brian Smith <brian@briansmith.org>, 'Mark Nottingham' <mnot@mnot.net>, 'atom-syntax Syntax' <atom-syntax@imc.org>, 'HTTP Working Group' <ietf-http-wg@w3.org>



Julian Reschke wrote:
[..]

> 
> What I'm missing a bit in this discussion is the fact that the Link 
> header already is specified, deployed and supported by some UAs (such as 
> Firefox (*) and Opera AFAIK).

I decided to try and test this out. To which end I created a simple page 
at [1] that does not have an HTML link to a stylesheet but then 
configured the server to include such a link using HTTP.

The results show that IE and Firefox do not implement HTTP Link but that 
Opera 9 does. However... the link you give to the Firefox Link 
Pre-fetching FAQ is interesting.

It posits a single relationship type of 'prefetch' - simply a hint to 
the browser that it might be worth fetching whatever it points to 
because the page that's about to be parsed probably includes a link to 
it. Hmmm... that's not quite the same is it? It still needs the link 
within the HTML to actually use whatever it prefetched.

IMHO, that _is_ a case for Brian's suggestion of a specific header.

It fits the semantics of link in that it does point to a related 
resource, but the relationship type is stretched a little I think. 
Suppose that, for whatever reason, you prefer to use your own stylesheet 
  - the prefetch might have exactly the opposite of the desired effect - 
i.e. using bandwidth to fetch something irrelevant.

Meanwhile, by using rel="stylesheet", my HTTP Link header pointing to 
the stylesheet is ignored by Firefox.

 From a POWDER point of view it's important that links encoded within 
HTTP or HTML should be semantically identical and therefore present 
publishers with alternative ways of producing the same result - simply 
because not all publishers are the same.

> 
> Thus I'd really prefer to *extend* it in a backwards-compatible way (as 
> proposed by Mark N.) instead of inventing something new.

+1 to that.

Phil

[1] http://www.fosi.org/archive/httplinktest/

> (*) for instance: 
> <http://developer.mozilla.org/en/docs/Link_prefetching_FAQ>
> 
> 
Received on Tuesday, 6 May 2008 11:38:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:47 GMT