RE: [ciq] Use of OASIS code list specs. in CIQ TC V3.0?

I concur here. This was pretty much my thoughts.

OASIS seems to have a happy knack too - of picking TC names that do not reflect what they are doing. Codelists is one - and then the latest in that saga is "docstandards-interop" - which you could be forgiven for thinking was developing interoperability standards - instead they are designing a common XML representation format for use with DITA for creating technical documents.

Codelists should have been called something like - code values metadata IMHO - but I have the sense they wanted it to be overlapping.

My big concern here is that people will think the codelist stuff is to be used in creating actual codelists for use - because noone is saying not to - and we can already see that happening in UBL here:

notice - tons of processing overhead and tough to quickly verify via eyeballing it - just what codes are or are not there (sorry silly me - I thought one of the key design features of XML was its human accessibility...)

Compared to CAM extension format - where the same values take up just 12% of total bytes - can be viewed in couple of screen pages in XML viewer - and lookup occurs in fraction of CPU cycles.

Lesson learned - it does not look authoritative to have something really simple and compelling - you have to make it have "throw weight" where the printout sits a foot high when placed on someones desk for review.

I'm much happier with following your example - and collecting useful references - and pointing out these things are HIGHLY CONTEXT dependent and people should not feel locked into something because.

BTW - you may have noticed that ISO are terrible at versioning and providing consistent releases of XML versions of their code values from a secure and sensible location - such as sourceforge CVS! And that's because they are not in the software business - but a standards publishing outfit - so they still think books and paper.

I'm not sure what the real answer here is - but I hate to give people the impression that OASIS codelists work is for defining codelists - rather than just metadata.

Very bad "deja vue" attack here!

Hopefully Ram can suggest something good for us to do - that adds value and gives people a heads-up on what their simple obvious options are with CIQ.

Perhaps providing pointers in our CIQ context - we can sift through what folks need - to save them time tracking down useful codelist sources?

Cheers, DW

"The way to be is to do" - Confucius (551-472 B.C.)

-------- Original Message --------
Subject: RE: [ciq] Use of OASIS code list specs. in CIQ TC V3.0?
From: john.glaubitz@vertexinc.com
Date: Tue, April 17, 2007 10:26 am
To: "David RR Webber (XML)" <david@drrw.info>
Cc: ciq@lists.oasis-open.org, "Bruehlmeier,David"
<david.bruehlmeier@sap.com>, Max Voskob <max.voskob@paradise.net.nz>,
ram.kumar@oasis-open.org, David Bruehlmeier <typo3@bruehlmeier.com>

David,

If you could help me out, I'm not sure I understand what you're objecting
to with this proposal and I don't believe it's necessarily true that
"everyone knows what a codelist is and what it contains".   Certainly if
you limit the code list information to simply the code and it's value,
it's
extremely straight forward.  However, the problem is that there are many
code lists that represent the same fundamental values and those code
lists
evolve over time as the values represented change.  So providing a
structure for a list of codes that includes all the meta data that the
parties involved in the exchange of information can use to identify that
they're working with the same lista seems very worthwhile and not all
that
obvious.

To take a simple example and one that's relevant to CIQ, consider country
code.  It would be wonderful if the world only used the ISO 3166 list in
its most up-to-date form, but that's not necessarily the case.  There are
many lists that may be appropriate or at least preferred for a particular
communication.

http://en.wikipedia.org/wiki/Country_code

This is a simple example, but there are codes for everything and ensuring
that there is commonality seems to make common sense.

All that being said, I haven't yet fully reviewed the Code List spec
to say
I would recommend it for our committee and we should be cautious about
endorsing it, but in concept I would welcome having a standard.

Thanks

John

John Glaubitz
Vertex, Inc.
1041 Old Cassatt Rd.
Berwyn, PA  19312
john.glaubitz@vertexinc.com
Tel: +1 484 595-5877

ciq message