Hi Joe,
Thanks for prodding me. I should have been more clear on my
description.
I'm not sure from your description whether the PI's are in the input
document or are created by the transform -- and I'm not sure who expands
them.
If you could sketch a small example, it might be easier to make suggestions.
The way the XSLT works is that it returns an appropriate page to
insert the base content. This content is read off an XML DB, and a
Xhtml page is returned. So, as far as the content is concerned, is is
placed in the correct place.
The PIs are inserted by the transformation. What I am looking for is
the ability to then send the transformer output to a filter, which
will look at the PIs' target and data and use it to fetch additional
information and insert it. Think of adverts, popular pages list,
random text blocks and other text been read off a XML DB server.
Currently, the inserts are been handled by XPath objects that point to
the appropriate location for the insert. It works, but in not
flexible in the fact you need to do some coding to add or subtract any
insert points. (We want to support different platforms in a more
consistent manner.)
Hopefully, this makes it clearer.
(Personally, I'm wondering why you're going with PIs rather than letting
the transform fully construct the output document... but maybe that will
become obvious when you provide an illustration.)
Am using PIs as it allows the page to be more flexible. We can change
the way these blocks are fetched by changing the target and data.
Joe Kesselman,
http://www.love-song-productions.com/people/keshlam/index.html
{} ASCII Ribbon Campaign | "may'ron DaroQbe'chugh vaj bIrIQbej" --
/\ Stamp out HTML mail! | "Put down the squeezebox & nobody gets hurt."
Alasdair.D