David,

please consider this scenario:

You have a stylesheet that is saved to disk somewhere and have a document
you want to include, this document can be described by a relative path to
the stylesheet. The stylesheet is now used to transform xml fragments that
are generated in memory, e.g. from a database query.

With your new default beaviour this would no longer be possible, as the path
now has to be relative to the source tree, which does not have a pyhsical
representation in the file system.

I think that this is a quite common use for stylesheets and the spec should
cater for this !

The second argument allows for futher uses (like the example) that extend
the default behaviour, and therefore I believe that the current form of the
document function has its validity.

Axel

> -----Original Message-----
> From: Paul Tchistopolskii [mailto:paul@xxxxxxx]
> Sent: Tuesday, September 12, 2000 1:10 PM
> To: xsl-list@xxxxxxxxxxxxxxxx
> Subject: Re: XSLT V 1.1

 ----------------------------- snip
----------------------------------------------

> I don't understand what is a problem. I'm dumb and I don't understand 
> things until I see the example.
> 
> I have been provided with the example. I answered : "for this 
> example, 
> changing the deafult behavior of XSLT engine to something 
> natural and easy to understand - should work".
> 
> Now you are saying that "there is another example, which breaks
> your solution". 
> 
> What is your example?
> 
 ----------------------------- snip
----------------------------------------------


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