OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

ubl-ndrsc message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: RE: [ubl-ndrsc] Ur type discussion


My understanding (and the only way this thing will work without an
almost-wholesale abandonment of interoperability) is if we have a complete
set of Ur-types - one for each UBL type.

Each type that lives outside the UBL tree should be descended from an
Ur-type; all UBL types are descended from Ur-types, but through the UBL
library. Otherwise, entire documents will have to be processed as if they
aren't at least partially (and generally this will be 90+%) descended from
UBL types.



-----Original Message-----
From: Burcham, Bill [mailto:Bill_Burcham@stercomm.com]
Sent: Monday, December 09, 2002 3:10 PM
Subject: [ubl-ndrsc] Ur type discussion

Great to have you back Matt G.!  A few questions about the Ur Type position

** Ur-Types for All ABIE's or Just for "Message Types"? **

Is the proposal to have ur-types only for "message types" and not for all
the ABIE's?  The discussion around line 108 makes me think so.  If that's
the case then a user could specialize the Order (ur-type) so as to remove
the OrderHeader, but the user would be unable to e.g. remove the ZIP code
from an Address (ABIE) (assuming ZIP code was a Basic BIE Property of the
core UBL Address).

If I read that wrong, then is the proposal to have ur-types for all the
ABIE's?  In that case a user _would_ be able to remove the ZIP code from an
Address.  Such an Address would be "loosely" compatible in the terminology
of the proposal right?

** Inheritance Relationship Between a Type and Its Ur-Type **

The proposal around line 78 (in section 2 "The 80/20 Design Principal")
describes LCSC's decision to specify only the 20% of the structures
(components) that will suffice 80% of the time.  While the proposal doesn't
couch it in similarly quantitative terms, I'll volunteer that the
"specialization via XSD extension" described in section 1 "Type Derivation"
satisfies a similar equation, to wit, that type derivation will solve 80% of
the problem (with minimal effort).

Since section 1 "Type Derivation" sets out an XSD type derivation regime as
the dominant specialization mechanism (the one that we expect to suffice 80%
of the time), I wonder what the inheritance relationship will be between a
type and its ur-type.  Intuitively I thought that a type would derive from
its ur-type, but the proposal doesn't say so, and I can imagine certain
problems with that.


To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC