xml-dev - Best Practice - Character sets

[Paul Spencer <paul.spencer@boynings.co.uk>]

Perhaps I shouldn't be raising a new "best practice" issue before I have
caught up with the rest, but this one has really been bugging me.

You have probably picked up by now that I am doing schema work for the UK
Government. We have a number of re-usable complex data types, and naturally,
we want people to re-use them.

The problem we have is that we do not want to restrict the character set for
re-usable types. For example, we have to accept text in English and Welsh,
but some government departments might need to support all European
languages, and others might need to support non-roman character sets.

So now we have, say, a Name type that has no strict character set
restrictions. But say department A cannot accept a German name with an
umlaut. Given that the name is a complex type with sub-elements (from
memory: Title?, Forename*, Surname, Suffix?, PreferredName?), how do we
restrict the character set in Dept A's schemas? I can't see an alternative
to deriving each sub-element separately, in which case, I might as well
define a new simple type (a:RestrictedString) and redefine the complex type
from scratch for Department A.

Thoughts?

Paul Spencer
Author: XML Design and Implementation (Wrox Press)
Co-author: Beginning XML (Wrox Press)
Boynings Consulting - Delivering XML to industry and government
http://www.boynings.co.uk/