Re: === Identifying a unique node in a DOM tree ===

I think you'll find that the templates listed are (a) correct and (b)
fast enough for their purpose, since you are presumably interested in
taking *persistent* bookmarks and don't need to do that very often.

However, the bookmarks generated by this approach will break if someone
goes in and adds or removes elements on the preceding-sibling axis of
any element included in the bookmark path. In other words they will only
be useful if the document being edited is locked as read-only until the
editor returns to it.

If you want to make the bookmarks more robust, so that they survive
across serial single-user edits, then I suggest you start off each edit
session by resolving each bookmark path to a specific node, and end each
session by regenerating the bookmark paths for each bookmarked node.
This sounds more like a DOM/XPATH approach than an XSLT approach, but
then editing XML *is* more DOM than XSLT.

Hope this helps -

Francis.


Khalid Asad wrote:
> 
> I just can't help asking again ... why, why, why is there not an XPATH
> function as part of the XSLT spec? All these loops will kill performance and
> it's not even obvious (based on the number of questions on the subject) how
> to get the function right.
> 
> Khalid
>


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