Came across this interesting exchange in xml-dev mailing list.
Norman Gray says:
“I still very strongly suspect that XSLT was a James Clark joke. After being told that DSSSL would never take off because it had the wrong shape of brackets, is it not the case that he came back with a spoof mini-subset of it which had the right brackets, but which was received with such rapturous acclaim that he never had the heart to confess it was intended as a gag. No?”
Having spent good bit of time myself on using XSLT, I couldn’t believe that XSLT was a joke. Anyway, Henry S Thompson was quick to assert that it was not a joke 🙂
“No. I was there. The decisive breakthrough in the DSSSL->XSLT move
was the realisation that by writing XSLT _in_ XML we could make the
‘output’ side of templates iconic — that is, instead of _describing_
the desired result tree fragment, as DSSSL has to do, we could
_manifest_ it.”
And Michael Kay (love his XSLT book) responding:
“No, the use of angle bracket syntax was to filter out undesirable users
whose minds had been corrupted by use of curly braces. This had the entirely
intentional consequence that the language is happily used (a) by people who
would otherwise be writing HTML and can see that XSLT is similar but does a
lot of the work for you, and (b) by programmers who are sufficiently
open-minded to see the deep beauty of the language through its superficial
ugliness, while scaring off the Javascript kiddies who don’t deserve such
good tools.In fact, using XML as the syntactic basis has many benefits. The most
notable one for me is that it is very easy to extend the language: whereas
XQuery goes through anguish every time a new construct is added, because of
the ambiguities and inconsistencies introduced by new grammar, XSLT is
infinitely extensible through new elements and attributes with no problems
at all.”