[dsdl-discuss] Re: Fw: DSDL - Datatypes and Character Sets

From: Rick Jelliffe <ricko@allette.com.au>
Date: Thu Sep 30 2004 - 14:58:20 UTC

> Peter
> Re:
> >Some thoughts on Datatypes and Character Sets.
>
> >*Character Sets
>
> >*Character sets can be considered a type of datatype. They both define a
> > set of characters that are legal within an element or attribute.
> > Therefore, I believe there is a case for combining these 2 areas into
> > one task.
>
> I note the point, but would point out that situations exist where you
> want
> to check character sets independently of datatyping (e.g. to ensure that
> your output devices can handle the input) and that, therefore, we should
> allow this to be a separate process.

I don't think that character repertoire is a type of datatype, at least,
not for the use cases that interest me.

There is a strong movement in software engineering dealing with
"separation of concerns". This emphasises that parallel issues
should be treated individually, not lumped together.

For example, "we use DOCBOOK" is one concern. And "we use traditional
Chinese characters" is another. Should we combine them and have
a single "traditional Chinese DOCBOOK" schema which uses
sinified datatypes everywhere? No, it would be cumbersome, and
in any case, there is no schema language that supports it.

In other words, I beleive the character reportoire schema
needs primarily to address the need for internationalized
versions of standard schemas, then secondarily to meet the
need for publishers to check which publishing characters
are being used, and only then, as the least requirement,
to address the needs of field-level constraint checking.

In particular this is because the data type library will
presumably have some kind of regular expression checking,
while the character repertoire validation should not go
that far: its problem is the convenient manipulation and
naming of large numbers of characters into sets and combining
them.

But certainly it would be sensible if the character reportoire
schemas could be involked as a datatype too. But that is
accidental, not essential IMHO.

Cheers
Rick Jelliffe

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)
Received on Thu Sep 30 17:47:59 2004

This archive was generated by hypermail 2.1.8 : Fri Dec 03 2004 - 14:00:28 UTC