[
Lists Home |
Date Index |
Thread Index
]
RELAX NG is a solid piece of work, thanks mostly to the ideas and hard work
of James Clark, Murata-san, and also I think Kawaguchi-san. I think it is
elegant because it is so intuitive and nicely composable. You just have
fewer things to forget. It is also very implementable, I believe, given the
leg up (formal semantics, source code and algorithms) for coders provided
mostly by Clark, Murata-san, and Kawaguchi-san.
I have written a number of XML Schema and RELAX NG schemata by hand that
match identical instances. The RELAX NG schemata took considerably less time
to compose than the XML Schema schemata. (Another interesting exercise is
converting such schema into an instance by hand.) But not everybody wants to
do this kind of work. Eventually, I think schema composition will be well
hidden from the common user. So why does RELAX NG matter, if all the rigor
is to be hidden under a point-and-click interface?
RELAX NG matters because it has merit. People who really have to do this
kind of work cannot deny that RELAX NG has merit. So I am quite confident
RELAX NG will survive and evolve and grow. I owe it to the people I work
with and for to let them know the good things about RELAX NG. I have found
very little to complain about.
It was a little TC that put it together. The process was not bogged down
with requests or demands from vendors to have this or that feature jammed
in. The mantra was "simple, simple, simple" because the main contributors --
the aforementioned triumvirate -- held the line. They knew that simplicity
with intelligence is something the people who do the real work would
appreciate. It is not for everyone I guess, but I think of RELAX NG and the
tools to process it -- all free -- as a boon. The more I use it, the more I
appreciate it.
It has not flourished yet. It may take a good deal of time. But it has a
solid foundation that I believe will not shatter under pressure.
Mike
|
