Re: AW: A Generic template for multi-pass processing (

Griebel, Peer wrote:

> Hi Dimitre,
> 
> (sorry that it took some time to answer)
> 
> Your solution uses a special msxsl extension. So it's not pure XSL. I hoped
> to be able to do it without an extension. So I think I'll simply call the
> processor two times...
> 
> But nonetheless thank you!
> 
> connection reset by
>   Peer

Hi Peer,
I'm also a bit'o purist and try to avoid using extension functions as much as
possible.

However, this is a special case -- the xxx:node-set() function is implemented in
most (every?) XSLT processors -- be it msxsl, saxon, xalan, ... or other.

You prefer to "call the processot two times", but this will require even more
implementation-dependent and specific operating system/programming language
environment support than simply using the "standard extension function"
xxx:node-set(). And we also know that starting from XSLT 1.1 we'll just mechanically
replace the xxx:node-set($someVar) with $someVar.

Also, it will be probably considerably less time-efficient, considering XSLT
processor loading, initialisation and parsing time.

In my opinion this is not only a matter of taste.
When one analyses the consequences of not using the xxx:node-set()function -- and
starts to see that this is too limiting and restrictive, then a compromise to use it
is quite reasonable. 

It is a practice, widely agreed upon in the xslt-list, to use the xxx-node-set() in
the code offered as solution to various problems.

There are a number of cases when we ***have to use*** non-XSLT techniques -- passing
parameters to a stylesheet being the most notable example -- and we have to live
with this reality.

Cheers,
Dimitre Novatchev.

__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail
http://personal.mail.yahoo.com/

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