Derivation by Restriction

[Fabio Arciniegas A.]

Hi Stefan,

> This principle would be violated if you "restrict" B 
> that a subelement required in A is not required in B. 

Right, at least if we map every restriction to an inheritance during
the translation. This can be not so, but it brings questions about the
usefulness of derivation by restriction.

> If I read the specs 
> right, you are not allowed to widen the [minOccurs...
> maxOccurs] interval, so the problem you address seems 
> not exist.

I understand your point; The key word is "required";Yet, my _real_
concern is not about the congruence of type definition on the Schema
in which you are right: since Derivation by Restriction always *narrows*
[minoccurs ... maxoccurs] or eliminates members of a disjunction, B,
inherited by restriction from A is always a member of the original
type. 

The differences between typing in Object calculi and typing in
schemas may render my original example useless even in the cases where
and schema is used to reflect OO constructs different than data
containers (case in which, we agree there is no problem). Furthermore,
as I mentioned in the previous mail, even for the strangest cases I
could think of when using schemas to define other OO/OA ideas
(e.g.Cross-cutting concerns) there are natural workarounds that avoid
compromising this and other principles...

However, this was just the oblique thought that started the wheel,
maybe the really important thing is that it brought me and others to
question the cost/benefit of derivation by restriction on Schema, as
well as its impact on implementors and users. 

Probably the best thing is to have D. by R. (I have a deep trust on Dr. H.T.
et al.); yet, I (we?) think that specially in the light of its impact
on implementors, more on the rationale/real added benefits of D. by R.
would be nice to hear.  

Best,
	Fabio

ps.Thanks, Stefan.

--
Fabio Arciniegas A.  	                    Viaduct Technologies, Inc.
fabio at viaduct.com			         Software Engineer
Interests: XML, Wittgenstein and just about everything in between.
Oblique Strategy of the day: "Feed the recording back out of the medium"

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1
Unsubscribe by posting to majordom at ic.ac.uk the message
unsubscribe xml-dev  (or)
unsubscribe xml-dev your-subscribed-email at your-subscribed-address

Please note: New list subscriptions now closed in preparation for transfer to OASIS.