RE: [xsl] XSLT vs Perl

["Jim Fuller" <jim.fuller@xxxxxxxxxxxxxx>]

> But both of those specs integrate portions of XQuery and XML 
> Schema that are really over the top.  As it stands, XSLT 2.0 
> smells of second system syndrome.  There are more freeping 
> creatures in there than I can count.  And understanding what 
> all of those creatures do requires a 
> huge investment in time, paper and effort.

As a random 2p.

I think Xquery is simpler then xslt by any measure...yes I think that
initially Xquery as it *was* being specified could have found themselves
with a dead end language, ...but actually its super simple; w/o sounding
too sugary I think that the existing W3C group responsible did the
language justice....and we will feel its impact at some later point when
we need to do some real work on data...XSLT is lightweight fun, airy and
light....though I wouldn't want to apply it to anything but content
related tasks currently; by comparison Xquery is an industrial strength
solution to many *hard* data problems.

> By comparison, XSLT 1.0 required a good grasp of XPath 1.0 
> and a small number of XSLT element behaviors.  Fewer 
> interrelations, fewer techniques to master, and easier to 
> apply than what I remember from the XSLT 2.0/XPath 2.0/XQuery 
> 1.0/XML Schema 1.0/XML Schema Datatypes 1.0 stack.

More 2p.

XML Schema is dead on arrival, always was dead.... it never
happened......so the W3C had a bad day, so what, lets get over it and
use RELAX NG and move on, or at the very least translate to XML Schema
at the last moment. 

On the other hand XPATH, XSLT and I have no doubt XQUERY will continue
to be successful.

> > Therefore, if people preferred XSLT 1.0 over other languages when 
> > performing these tasks, despite its shortcomings, they will 
> certainly 
> > prefer XSLT 2.0 over other languages.

People prefer other languages because they don't want to learn 5
languages to achieve something, XSLT is good at transforming xml into
other xml...which apparantly is vastly useful....possibly the work of
partial schemata in motion.

> Actually, the way XSLT 2.0 is going, I'd much prefer XSLT 1.1 
> or 1.5: add the grouping and date handling, remove the 
> nodeset/rtf distinction, and fix a couple of other warts in 
> XSLT 1.0.  *That* would be a killer language.

To me XSLT has benefited from the wild success of XPATH, we all get
it....though I cant understand why XPATH hasent been extended to an xml
update language....why make data when we have no facility to update it ?

> The situations that demand input/output validation and XQuery 
> integration are totally separate domains.  That greatly complicates 
> XSLT 2.0, and obscures the fact that part of XSLT 2.0 is 
> XSLT 1.0 - warts + fixes.

Its rare for version 2.0 of anything to live up to its name, so why not
submit your suggestions to W3C before Feb 15th ?

Gl, Jim Fuller



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