Re: [xml-dev] So maybe ID isn't a problem after all.

[James Clark <jjc@jclark.com>]

> But I still claim that in a world where servers like cocoon/axkit, even
> IIS are doing server side transforms and sending any number of different
> mime types for the same URI, depending on the requesting client, that
> it's useful to have the _possibility_ of publishing a URI reference to
> within a document that works over a range of mime types. And you don't
> have that possiblity unless you keep the syntax being very simple, and
> in fact just a  name.

In a case where a server might be sending either XML or something else like 
HTML or PDF, then we are typically dealing with a human-readable document, 
which means that the XML is going to need a stylesheet.  One common 
scenario would be to send an XML document with an xml-stylesheet processing 
instruction pointing to an XSLT stylesheet that transforms the document 
into HTML. However, in this sitation with IE5/6, the fragment identifier 
will be interpreted not with respect the XML document, but with respect to 
the HTML document generated by the XSLT stylesheet.  Now, this is arguably 
broken, but it's quite practical.

James