RE: [xsl] [Ann - Zvon] Templatotron

Subject: RE: [xsl] [Ann - Zvon] Templatotron
From: "Bryan Rasmussen" <bry@xxxxxxxxxx>
Date: Wed, 20 Feb 2002 10:09:07 +0100
On Tue, 19 Feb 2002, Evan Lenz wrote:

>> Are you familiar with the "literal result element as stylesheet"[1]
feature
>> of XSLT? It seems that rather than trying to make Templatotron more
>> expressive by introducing instructions like t:choose, t:if, etc., one
would
>> be better off using the simplified stylesheet syntax.

I thought the LRE was supposedly for html also, and at any rate I only know
that Saxon supports it? I always had the impression that msxml doesn't, is
the LRE supported by major processors that you know of Evan, and if so can
it be used to output xsl-fo, svg, etc?

Miloslav   Nic wrote
>I was partially aware of it :) I knew about this possibility but I somehow
>made  a mental shortcut and I thought that this was applicable only to
>html output :)


>I have some user cases where I need most of the Templatotron power
>including namespace handling.  But there is no problem to derive another
>language out of it which will retain the bits you consider useful.
>You can use as much code of the Templatotorn implementation as you wish,
>just change the namespace for your language.


the difficulty for me when I look at Templatotron is that, as opposed to
most specialty applications using xslt as the compiler, such as schematron
or Oliver's loopxsl, is that it seems to do much of what you do with xslt
anyway, whereas the two aforementioned focus on one very specific possible
app of xslt and implement it. I mean that Templatotron, as I understand it,
is a transformation language implemented using xslt, which is a
transformation language. I liked it when I looked at it, but while it is
useful for you, I wonder about the usefulness for other people to learn the
syntax, I've done the same thing in my time, it seems like one of the
perniciously interesting ideas of xslt that one can create a simplified
version of the language so that others don't have to struggle with the full
complexity. The problem is, I've found, that when you don't make it complex
enough - which is the problem in all those tutorial books where they have
"use xslt to generate xslt!" then it isn't very useful, and when it is
complex enough people might as well invest their time in learning xslt, that
said the templatotron design I think is much better than most of these
example implementations one sees, you could no doubt get it into the next
version of Profesional XSL.




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


Current Thread