Re: SVG12: SVG namespace declaration

On Monday, April 18, 2005, 12:58:59 AM, Bjoern wrote:


BH> Dear Scalable Vector Graphics Working Group,

BH>   http://www.w3.org/TR/2005/WD-SVGMobile12-20050413/struct.html section
BH> 5.1.1 claims 

BH> [...]
BH>   In all cases, for compliance with the "Namespaces in XML 1.1" 
BH>   Recommendation [XML-NS], an SVG namespace declaration must be
BH>   in scope on the 'svg' element, so that all SVG elements are
BH>   identified as belonging to the SVG namespace.
BH> [...]

BH> This is incorrect, if the 'svg' element is not in the SVG namespace the
BH> requirement does not apply.

Yes, we meant to say on 'our' svg element. But there is no easy way to
indicate that textually. svg:svg could also be misleading. 'The svg
element in the svg namespace must be in the svg namespace' is
tautologous.

The reason this text is in there is because SVG 1.0 allowed al the
namespaces to be defaulted in the external DTD subset, which gave
problems with parsers that do not fetch it.

Mozilla in particular, has problems with svg content without an in-scope
namespace.

We can't say 'on the root element' because that might not be an svg
element. We can't say that the declaration has to be on the rootmost svg
element because it could be on any ancestor of that. So we can only say
that it must be in scope.

So okay,

<svg xmlns="http:example.org/saphenous-venous-grafts" />

*is* compliant with namespaces and is *not* a conformant SVG document or
fragment. But then, that is what we want.

BH> All the draft could say is, when a resource
BH> has been determined to be a SVG document through means of higher-level
BH> protocols such as a Content-Type protocol header,

I don't see that the orthogonal issue of Content-Type needs to be
brought into this. In fact, that could be misleading.

SVG carefully defines a conforming SVG document fragment as well as a
conforming SVG document. If for example someone sends content in a media
type uses mixed namespaces, for example DocBook with inline SVG, then
the conformance to the structure of the svg elements still applies even
though its not an image/svg+xml content type.


BH>  the root element of
BH> the document must be an 'svg' element

No, we explicitly do not want to say that.

BH> in the SVG namespace. Please change the draft to state something
BH> that is neither incorrect nor mis- leading.

We would if we could think of a way to say it. The wording you proposed
introduced other problems. Could you try again?


-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead

Received on Monday, 18 April 2005 18:51:41 UTC