- From: Chris Lilley <chris@w3.org>
- Date: Fri, 5 Nov 2004 17:11:05 +0100
- To: Robin Berjon <robin.berjon@expway.fr>
- Cc: "L. David Baron" <dbaron@dbaron.org>, www-svg@w3.org
On Thursday, November 4, 2004, 12:49:59 PM, Robin wrote: RB> L. David Baron wrote: >> On Tuesday 2004-11-02 17:22 +0100, Chris Lilley wrote: >>>I don't see that such a statement belongs in the media type registration >>>template. I agree that the statement needs to be in the spec. >> >> So you think that it's within scope of the SVG spec to define the >> handling of non-namespaced elements? Or are you proposing a different >> way of making the same thing a requirement? RB> I think Chris' point is that yes indeed the SVG spec should say that RB> elements in no namespace that happen to have SVG local names shouldn't RB> be rendered (they might not be in error, but they produce no rendering), RB> but no, the media type registration isn't the right place to make that RB> statement. Yes, that is the point I was making. RB> The fact that major SVG implementations will tolerate documents with no RB> namespace traces back to a mistake that was made to declare the RB> namespace as a defaulted attribute in the DTD. That's bad practice, and RB> was therefore removed but UAs somehow kept it. RB> What's most needed here is a test to flag that out. There'll be one. Yes (Robin, struct-frag-05-t.svg ) >> Such a document is invalid (and therefore non-conformant) since it >> doesn't meet the validity constraint that the DOCTYPE declaration match >> the name of the root element [1], since there is no root element. >> (I'm not sure whether that's the point you were trying to make. But it >> was the reason I cited that section.) RB> I don't think that DTD validity is the best way to phrase this either. Nor do I, but David seems to have missed the algorithm in the cited reference and thus seems to believe that DTD validity is being applied to the document as a whole rather than the extracted svg fragment. -- Chris Lilley mailto:chris@w3.org Chair, W3C SVG Working Group Member, W3C Technical Architecture Group
Received on Friday, 5 November 2004 16:11:05 UTC