Re: [ubl] CLSC: Draft paper on code list data models: draft-coates-codeListDataModels-0p1

[Chin Chee-Kai <cheekai@softml.net>]

On Fri, 28 Nov 2003, Tim McGrath wrote:

>>we must avoid building the NDRs for code lists before we have the
>>semantics right.  

I tend to agree.  I'll go a bit stronger even, that we
don't call it Naming & Design Rules anymore, but instead, call it
Code List Construction Specification.  

What we need are constructive specs rather than constraining 
rules to tell us how, given a finite set of code values, to 
render the code values in a systematic, unambiguous and 
consistent manner into a code list schema for that given set
of code values.



>>we should be
>>following the LC idea of a syntax independent conceptual data model and
>>then a set of rules for XML expression in UBL schemas.  The CLSC are
>>tasksed with the syntax issues of the latter whereas LCSC should be
>>investing time in the former.  As i said on the LC call, the semantics
>>of codes should remain an LC task - this is the home of the domain
>>experts who know what these semantics are (at least in theory  ;-) ).

Agree as well.  I think one most important outcome of CLSC is
to tell us how to construct a code list schema when given a 
finite set of code values, whether they are from external (to
UBL), or internal.  One starting point would certainly be the
Code List Catalogue, which is a master list of various code lists
used in UBL.   Knowing in advance that UBL itself will not register
the entire universe of all code lists, the same "master list" model
could be re-used and implemented in other organisations managing
several lists as well.




Best Regards,
Chin Chee-Kai
SoftML
Tel: +65-6820-2979
Fax: +65-6743-7875
Email: cheekai@SoftML.Net
http://SoftML.Net/