[
Lists Home |
Date Index |
Thread Index
]
Jonathan Robie wrote:
>I actually think a keyword syntax for XSLT might be more accessible for
many programmers.
To echo an earlier point in this thread, I think it's the recursion and
non-procedural approach that trips up programmers accustomed to the most
popular general-purpose programming languages, and replacing some of the
start- and end-tags with curly braces won't change that.
XSLT's XML format was one of its huge advantages over DSSSL, which it
effectively replaced the way XML replaced SGML (and DSSSL had far less
success than SGML). It's much easier to read
"</xsl:if></xsl:for-each></xsl:if></xsl:variable>" and know exactly what
kinds of structures are being ended, in what order, than to look at "))))"
and know the same thing.
Another great advantage of an XML-based syntax is that it's so easy to
automate the manipulation of resources expressed in that syntax. Outside of
XSLT, I can't think of another language in which it's so easy to write
programs that read and manipulate other programs written in the same
language, and that includes LISP and Scheme (")))))))...."). Many, many
production applications out there have automated the reading and writing of
XSLT stylesheets as part of their workflow.
I prefer RNC to RNG, but if I want to automate the manipulation of RNC
files, I'm going to trang them to RNG files first. I wouldn't even look at a
non-XML representation of XSLT until there was a conversion tool as reliable
as trang to go with it.
Bob DuCharme www.snee.com/bob <bob@
snee.com> weblog on linking-related topics:
http://www.oreillynet.com/pub/au/1191
|
