Lists Home |
Date Index |
Yes to declarative is better. The implications are harder when
that declaration has to be shared with lots of previously
unengaged individuals (which is what companies, countries,
etc. come down to). I agree there are multiple means and
that the architecture as spec'd should account for that
by modularity. I disagree with those that assert what is
in XSD is useless. I agree that the spec writers should
be very careful of normative inclusion into other languages
by reference unless they strictly intend to limit the
scope of operation of those languages.
International standards even at the simplest of type
levels, elements and attributes, are already very hard
to create. What do they call area codes in Nigeria,
Britain, France, etc.? We have a fellow here who
insists that everywhere else in the world, these are
called "city codes". On the other hand, I have a
fellow in Britain who insists that is untrue. Who
to believe? I have to wonder if naming the names
is this hard for the same item of data, and the
disparities among things like legal systems
are so large, maybe a common set of declarations
for datatype safety isn't that daunting.
I understand the issues of cost and variety in
the programming languages. The specs should be
modular. How many options can XML buyers afford?
From: Uche Ogbuji [mailto:email@example.com]
Your implication, if I read you rightly, is that it's better to use declarative means to ensure this conformance rather than writing a lot of"code".
Believe me, I preach this gospel every day.
My point is that there is more than one way to declare constraints, and I don't think the way enshrined by XSDL is especially useful, nor do I believe it should receive consideration at the expense of other schemes.