Re: SVG 1.1 test missing SVG namespace

Hi, Boris-

Boris Zbarsky wrote (on 9/25/08 3:30 PM):
> The external resource document used by
> is missing an SVG namespace declaration, 

Thanks for pointing that out.  I've fixed it now.

> with the result that the test
> should be failing in UAs that actually require SVG nodes to be in the
> SVG namespace.

Only if you want to be a stickler about namespace declarations, which
are out of vogue these days.  I see merit in allowing for a set of known
root elements and fixed namespace prefixes for Web-centric languages, in
addition to having namespace declarations for inclusion of languages
that haven't yet "made it" into the top tier (and which could be added
to the list as they mature in use).

I would like to think that SVG and MathML could be in that "usual
suspects" list, as well as Xlink, SMIL, and RDF.  (RDF and RDFa make
such heavy use of namespaces anyway that I'm not sure that it makes
sense to drop the RDF NS... unless we also add Dublin Core and Creative
Commons to the list of known friendlies).  With tight cohesion of a few
specifications, like the aforementioned, we could conceivably even have
an "implicit" relationship between known languages, in the same way that
SVG is a host language for SMIL, deferring to that specification for
treatment and functionality of certain elements and attributes while
keeping them in the SVG NS.  (That said, this is not without its
downsides... among others, the confusing clash between the "fill"
attributes in SVG vs. SMIL.)

I would support the creation of a spec (Namespaces in XML 2.0?  XML 2.0?
 Namespaces and Host Languages 1.0?) that would codify these changes in
namespaces.  Admittedly, I haven't looked at some of the problems in
detail, but it would be interesting to explore them.

-Doug Schepers
W3C Team Contact, SVG and WebApps WGs

Received on Thursday, 25 September 2008 23:05:04 UTC