RE: [ciq] Use of Codelist in CIQ

Contrary wise - embedding the codelists (approach #1) - may result in people being surprised when they have enforced errors they do not want - and are painted into a corner.

Approach #2 is the most flexible (which is why people make all those fields in XSD optional that are not really).

I'm also seeing that most people are used to the fact that XSD does not check their codelist values and that either backend code or second pass validation is needed for that.

Of course I'm a little biased here - since OASIS CAM is IMHO the perfect compliment for this 2nd pass situation...

I'd feel more comfortable if the approach #2 is the default schemas - and that the approach #1 are offered as an alternate. The reverse would not be good.

Thanks, DW

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

-------- Original Message --------
Subject: [ciq] Use of Codelist in CIQ - A thought
From: "Ram Kumar" <kumar.sydney@gmail.com>
Date: Thu, April 26, 2007 6:56 pm
To: ciq@lists.oasis-open.org
Cc: "Max Voskob" <max.voskob@paradise.net.nz>, "Bruehlmeier, David"
<david.bruehlmeier@sap.com>, "David Bruehlmeier"
<typo3@bruehlmeier.com>, "Colin Wallis" <colin_wallis@hotmail.com>

Members,

I sent you a list of questions earlier about the things we need to consider

before using Codelist in CIQ. I have a thought. How about we use two versions

of CIQ :

- On using codelist and the other without codelist, but with the current approach

for codelists.

Though we will have two versions of the schemas, the XML documents generated

out of the two schemas will still be the same. Except that, one will do one pass

validation and the other will do two pass validation.

This gives the flexibility to the users to decide the best approach for them.

Let me know your thoughts. The reason why I thought about this idea is because

I am of the view that when users embedd CIQ specs., as part of their application or

other specs., and if they do not use codelist approach to handle their codes, they

will be forced to jsut validate CIQ related data using codelist methodology which

could be a pain for them.

Regards,

Ram

ciq message