W3C home > Mailing lists > Public > www-svg@w3.org > June 2003

Re: xlink

From: Tobias Reif <tobiasreif@pinkjuice.com>
Date: Fri, 27 Jun 2003 21:50:38 +0200
Message-ID: <3EFCA00E.603@pinkjuice.com>
To: www-svg@w3.org

Chris Lilley wrote:


> The Adobe viewer covers a lot of different drafts, so tries to cope
> with older content (its not clear that is a good idea at this stage).
> However, it is using the definition in the DTD in that case. You can
> check this by
> 
> a) declaring the xlink namespace with a different prefix, like foo
>   xmlns:foo="http://www.w3.org/1999/xlink"
>   and using foo:href will still work


What else could happen? Any other behaviour would be erratic (besides 
rejecting the doc as invalid SVG since the DTD specifies one single prefix).


> b) declaring the xlink namespace with the wrong URI
>   xmlns:xlink="http://example.com/wrongURI"
>   now your links have stopped working.


... since they aren't XLinks anymore.

Both a) and b) are, in each case, the only possible correct ways of 
behaviour.

What do these tests show?
You say that they show that the ASV is using the definition in the DTD.
I'm afraid I can't follow.

Choosing a different prefix in current versions of SVG makes the SVG 
invalid, which is one of the many strange and quirky symptoms of relying 
on a pre-namespaces schema language for the definition of a 
multi-namespace language like SVG. I think that choosing DTD was a 
sensible since pragmatic choice at the time, but I'm also looking 
forward to namespace aware normative schemas, especially if they leave 
the document's infoset alone.

Tobi

-- 
http://www.pinkjuice.com/
Received on Friday, 27 June 2003 15:50:59 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:25 GMT