- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Fri, 23 Oct 2009 13:18:57 +0200
- To: Sam Ruby <rubys@intertwingly.net>
- CC: Tony Ross <tross@microsoft.com>, "public-html@w3.org" <public-html@w3.org>
Sam Ruby wrote: > ... >> I realize that there is some broken HTML content out there which uses >> xmlns:* attributes, but doesn't expect them to have an impact on the >> DOM. The question here is: how many namespace URIs does this affect? >> Could we just exclude the big offenders (Word HTML export?) from >> processing? > > My recollection is that the biggest problem was xmlns="". As to Word > export, the biggest problem is finding attributes with names that start > with o:, but with no declaration for the namespace. > ... I also saw <o:p> (as opposed to <p>) being mentioned. It appears to me that if we *wanted* to align HTML and XHTML namespace handling more, we could. We just need to exclude certain constructs. For instance: 1) ignore xmlns='...' and xmlns='' (default namespace declarations) 2) missing prefix declarations are ok, elements/attributes missing them are treated as before 3) certain namespace URIs, such as the one used by Word's HTML export, are completely ignored I don't think these restrictions would be a problem in practice. BR, Julian
Received on Friday, 23 October 2009 11:27:14 UTC