Re: RE:read-write same url in xslt 2 [was appendig to

On Mon, 28 Jan 2002 12:27:41 +0100, Bryan Rasmussen wrote:

>
>>>why does it have to be an error? it works with Saxon if you're using
>>>version="1.1"
>>>
>>How do you mean 'works' ?
>
>I mean works in that it does ouput the same document that was read into the
>parameter, with the appended nodes.
>
Sorry, I got the Saxon processing model wrong in my earlier message -
Mike has corrected me.

>>>Summary- its a mistake to do this, even if it appears to work for a
>>>particular processor.
>I agree that I didn't think this out well enough to realize that in some
>cases one would get errors, I just sent another post on that. I have a
>philosophical disagreement there, if it works with a particular processor it
>isn't a mistake to do this, it's just a processor specific application.
OK, insert the word 'usually' - from my experience, anyway: YMMV.  The
problem is that it can be hard to distinguish between 'processor
specific' and 'processor *version* specific' and even 'processor +
stylesheet specific'.  You are relying on execution order, and that
can theoretically change for example due to some apparently unrelated
modification of the stylesheet causing the processor to adopt a
different optimisation strategy, and hence a different execution
order.

Regards,
Trevor Nash
--
Traditional training & distance learning,
Consultancy by email

Melvaig Software Engineering Limited
voice:     +44 (0) 1445 771 271 
email:     tcn@xxxxxxxxxxxxx

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