RE: xsl self-documentation on the fly /feature request

["Pawson, David" <DPawson at rnib dot org dot uk>]

 Eric van der Vlist 

>I am not sure of which kind of side effects you're mentioning here ;=)
>
>If you mean side effect in the same meaning than when we say that XSLT
>is a side effect free language to justify that a variable cannot be
>modified, I think that to control the "logging" in the output tree
>should avoid the problems since the order of the output tree is
>guaranteed even if the order in which the operations are done is not.

Yes, that was my concern. Perhaps this is a start to the 'trace' solution
of documenting? I.e. documenting the response of a stylesheet to 
a particular input document.

Minor concern, unsure if its important. Paul mentioned that he
was having difficulty in getting good xml output. 

Is this a real issue? If its trace information its durability must
be very short? Normally trace information is noted, then discarded?
Perhaps for formal testing where a record is kept... but then surely
plain text output is sufficient?

Regards DaveP



>> >4) Alternatively, the logging could be done outside the XSLT

>If the goal is "only" to log in an external file the fact that a
>transformation has been made at time t on input document src.xml by
>trans.xsl with output out.xml, 

No, I was looking at the 'trace' sub-problem of docmentation,
rather than a log of activities.


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