Alexander Schofield wrote:

> No arguments there.  For the vast majority of cases, XSLT would be
> better off as better-looking non-XML source.  My only guess as to why
> they wrote XSLT with XML syntax is because maybe they foresaw massive
> automated document templating systems?  Where the XSLT itself was
> generated on-the-fly from fragments based on some conditions.  In that
> case its syntax would definately help, but for the vast majority of
> situations, it's just a pain.

First off, I've been using XSL since, what, 1996? 1997?  Well, whenever I 
wrote docproc, the project that spawned XSL:P, which eventually became 
Xalan.  I never, ever, found the XSLT logic constructs to be convenient to 
use; Dave hit the nail right on the head with his evaluation, AFAIC.  
However, I was mucking about with SGML and XML at the point in time when 
the transition from DSSSL to XSL was occuring (as a "standard" stylesheet), 
and I was overjoyed to see a stylesheet system based on XML.  Having to 
deal with three different syntaxes to do one job (Java, XML, DSSSL) was a 
severe handicap.

The choice to base XSL on XML was because XSL, for all of its problems, is 
programming language agnostic.  That is, XSL has little application outside 
of XML, so if you're using XSL, you're already committed to the XML syntax. 
This isn't true about any other syntax choice for translation.  XSL is 
bound to XML because that is its problem domain.  If XSL was based instead 
on, say Ruby, then Scheme and Haskell users would be inconvenienced.  If 
XSL was based on Haskell, Java and Scheme users wou would be 
inconvenienced.  And if you want a Lisp based system, use DSSSL.


Given that you *need* a "standard" stylesheet language for XML, an 
XML-based syntax, such as XSL, is an excellent choice.  XSL has a much 
lower cost of entry than, say DSSSL, and remember: it is supposed to be the 
non-programmers, artists, who are designing stylesheets, and any stylesheet 
system is going to require translations.

XSL is based on XML because it is the least inconvenient for everybody, 
although optimal for nobody.  Aside from folks like Tobias :-)

-- 
--- SER