As a person involved in the nitty gritty work of writing programs that
use both these technologies, I readily reach for XSLT over Perl for
transformation tasks. At the theory level, it may just look like a
preference thing. If you work quite a bit with both you'll find yourself
happily using XSLT for most transformation tasks. It just makes my job
easier on many levels.
Let's make it clear. I do prefer XSLT 1.0 over Perl for
XML transformations. The question has been about XSLT 2.0, which
is quite different.
Michael Kay asked me to provide a code equivalent for the XSLT 1.0
construct (I do not know why it is relevant). I did that because
it is as easy to do primitive XSLT operations in Perl with appropriate
modules as in XSLT (Though I admit that would use XSLT 1.0 for XML
transformations and not Perl, performing other manipulations using
other tools if necessary -- and it is not only because I do not
like perl).
The question is entirely different. I know what is the advantage
of XSLT 1.0 over Perl. It is the same as advantage of sed, or awk,
or grep over Perl. These are small dedicated tools optimized for
well-defined narrowly bounded tasks, and they do their tasks very
well, much better than perl does any of them.
But XSLT 2.0 brings a lot of burden; the question is what is
advantage of XSLT 2.0 in comparison to other scripting languages
which have the same burden, can do the same operations using as
short or shorter syntax for that.
XSLT 2.0 will be worse than Perl or Python in almost everything
which distinguishes it from XSLT 1.0; given support for XML, including
XPath, in the scripting languages, I would choose the other languages
just because they also provide much more flexibility.
Perl is not my favorite language. It is just the most popular
representative of a family of languages XSLT 2.0 belongs to.
That is why the question was to compare them.
David Tolpin
http://davidashen.net/
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list