Thanks for your advice. So I should move smoothly away from usage of
id() and idref()? I always thought (in fact I hoped it would be that
way) that elements with ID-attributes are treated in a special way while
they are parsed (some sort of indexing done automatically), which would
enhance processing performance... But Saxon seems to do a lot in
background.
Michael Kay wrote:
I'm thinking of moving away from DTD towards RelaxNG, but
what will happen to my tool chain? Especially loosing the
possibility of fetching elements via id() and idref() would
hurt me most.
Anything you can do with id() in XSLT can be done just as well with key() -
and arguably better, since you're less dependent on external factors to
ensure that the IDs are identified as such.
yes, and I often construct a "more qualified" key access on the fly like
<xsl:key name="column-by-fullname" match="column"
use="concat(parent::table/@name,'.',@name)"/>
(And in Saxon-SA these days you don't really need to bother with key()
either, you can just write //x[p='q'] and let the system take care of it).
Michael Kay
http://www.saxonica.com/
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--