- From: Robert Brewer <fumanchu@amor.org>
- Date: Thu, 7 Jun 2007 11:46:07 -0700
- To: "Nicholas Shanks" <contact@nickshanks.com>
- Cc: <ietf-http-wg@w3.org>, "Web-Kit Dev" <webkit-dev@opendarwin.org>
Peter Speck wrote: > On 07/06/2007, at 19:18, Nicholas Shanks wrote: > > An asterisk would be implied because there will always be > > a closest match, even if the only resolution is lower than > > a client requested. (i.e. There would be no 406 errors > > because of this header) > > If we allow a range, why not? Then the UA would be able to say: "I > dont want any media for dpi > 100". Cell phones could use this to > avoid download of huge files. If the proposal exists to minimize bytes transferred, wouldn't a new "Accept-Length" header be more appropriate and useful for a wider set of media types? The server could then vary the representation returned in order to meet the restriction given by the client (or return 406 Not Acceptable). If the issue is actually image resolution and not octets, a media-range parameter in the Accept header seems to be a more appropriate place to implement it; for example: image/jpeg;max-dpi=100;q=0.5 Regardless, I don't see how either one should involve any changes to the request for a CSS file. The user-agent will fetch the CSS resource, execute the CSS rules, and then fetch additional resources in separate HTTP requests. Those latter requests seem to me to be the appropriate place to specify restrictions on the representations returned. I can't think of any precedent for using the headers of request A to influence the characteristics of the response to request B--that's too tight a coupling between HTTP and CSS, in my humble opinion. Robert Brewer System Architect Amor Ministries fumanchu@amor.org P.S. Hello, ietf-http-wg list, I'm the author of CherryPy 3, which includes an HTTP server that is arguably the most used in the Python web-framework space these days. I'll try to keep quiet and learn from here on out. ;)
Received on Thursday, 7 June 2007 18:49:39 UTC