Re: feature request

Sebastian Rahtz wrote:

> I entirely take your point that XSLT cannot do this, but I wonder
> whether in practice many people will care. I suspect people will just
> write
> 
>  <!ELEMENT foo EMPTY >
>  <!ATTLIST foo
>         img CDATA #REQUIRED
>         type "gif|tif|png" "png"
>         >
> 
>  <foo img="testimg.gif" notation="gif"/>
> 
> which is almost as useful. them thar notations will go in the pool of
> SGML things are nice, but which one would not go to the stake for.

I would be perfectly willing to recommend abandoning notations and the
like over time if there was a viable alternative built into XML 1.0, but
there isn't. The example you provided is fine for a hard-coded
application that always looks for the type attribute, but it fails when
you apply it across various doctypes authored by different people.
Granted, you can usually deduce the notation type from the extension on
the URI, the mime type, or the contents of the file itself, but each of
these is beyond the scope of XML 1.0 itself (i.e. they require a pattern
matching engine or something like it). 

As for people not caring about notations and entities in general, I only
care about them because they abound everywhere I look. There are many of
XML doctypes and applications that depend on them, and I have always run
into them in every job I have had in this field. 

To summarize (since I've spent way to much time on this thread), I hope
that XSLT will finish the job with XML 1.0, and allow for a way to
create a serialized XML tree from any XML 1.0 document that will parse
the same as the original (no, I don't care if it isn't serialized the
same way). 

Rick Geimer
National Semiconductor
rick.geimer@xxxxxxx


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list