- From: Laurens Holst <lholst@students.cs.uu.nl>
- Date: Fri, 09 Jan 2009 11:50:09 +0100
- To: Ian Hickson <ian@hixie.ch>
- CC: Boris Zbarsky <bzbarsky@MIT.EDU>, Christian Schmidt <w3.org@chsc.dk>, HTML WG <public-html@w3.org>
- Message-ID: <49672BE1.8090701@students.cs.uu.nl>
Laurens Holst schreef: >>> Out of 104879 pages with at least one <img src>, from my collection >>> of pages from dmoz.org, there are 529 (0.5%) with at least one empty >>> <img src="">. >>> >>> I don't see any obvious pattern in those pages - there's a mixture >>> of old and new pages, dynamic and static pages, hand-written and >>> various generators, etc. So it doesn't appear to be the result of a >>> single tool. >>> >> >> Based primarily on #2 above and on Philip's research, I've made the >> spec say to ignore <img src=""> if the base URI of the element is the >> same as the document's address. >> > [...] > If you choose to specify this, may I suggest defining it differently: > make the spec say to ignore images where the (resolved) URI of the > element matches the document URI? That seems a rule that simplifies > the logic and where the implementation doesn’t need to perform two > checks (value = "" && this.baseURI == document.documentURI). Here the > explicit baseURI check seems a bit odd and unrelated, and working on > raw values (presuming knowledge about the resolving function) instead > of the result after resolving the URI. > > By the way, for reference, the Bugzilla bug that (recently) changed > Mozilla’s behaviour: > > https://bugzilla.mozilla.org/show_bug.cgi?id=444931 Note by the way that if I understand the patch attached to that bug correctly, the Mozilla implementation again does something different: value = "" && resolveURI(value, this.baseURI) == document.documentURI. (Where value is the attribute value.) Although this is effectively the same, the Mozilla version does not depend on knowledge of the URI resolving mechanism, which I think is better. Both Mozilla and the currently specced behaviour special-case this for empty src attributes only, which I guess is slightly better than applying this generically, so that there is at least a workaround. ~Laurens -- Note: New email address! Please update your address book. ~~ Ushiko-san! Kimi wa doushite, Ushiko-san nan da!! ~~ Laurens Holst, student, Utrecht University, the Netherlands Website: www.grauw.nl. Backbase employee; www.backbase.com
Received on Friday, 9 January 2009 10:50:55 UTC