Re: [codelist] Genericode 1.1? - a request from CEFACT for embeddedcode lists in code list set documents

["Anthony B. Coates (DES)"<anthony.coates@documentengineeringservices.com>]

Apologies that I've been so slow moving on this.  I mentioned previous  
that UN/CEFACT ICG had request some additions to the genericode Schema.   
In particular, in the 1.0 Schema, a code list set can contain

  * code lists
  * references to code lists
  * references to other code lists sets

UN/CEFACT ICG would like us to add

  * code list sets (not just references)

To be complete, I think we could also add

  * column sets (external definitions of columns and keys)
  * references to column sets

I have a prototype Schema that supports this, and am working on a set of  
examples to demonstrate the different ways code list sets can be  
constructed.  While doing that, I have found a bug in the genericode 1.0  
Schema.

In the 1.0 Schema, a code list reference contains

  * a mandatory canonical URI
  * an optional canonical version URI
  * any number of location URIs

A code list set reference contains exactly the same.

However, a column set reference contains

  * a mandatory canonical version URI
  * any number of location URIs

This is an oversight on my part.  When I made an update to the content  
model for references, I apparently fixed only the code lists and code list  
sets, not the column sets.

I think we should fix this, but that would make the next release of  
genericode backwards incompatible.  It would be a 2.0 release instead of a  
1.1 release.  I suspect it would only be "technically" backwards  
incompatible, because I have no evidence that anyone is using column set  
references at present (but I would certainly like to hear from anyone who  
is).  The purpose of this e-mail is to let you all know about this issue,  
and to get your feedback about fixing it in the next release of genericode.

Thanks very much in advance,

Cheers, Tony.

On Sat, 05 Apr 2008 11:58:38 +0100, Anthony B. Coates (Document  
Engineering Services) <anthony.coates@documentengineeringservices.com>  
wrote:

> On Fri, 04 Apr 2008 10:51:09 +0100, G. Ken Holman  
> <gkholman@CraneSoftwrights.com> wrote:
>
>> Can you please propose a 1.1 document and associated schema?
>
> OK.  The updated document will take the time, the Schema changes are  
> easy.  Note that I have given UN/CEFACT ICG an interim Schema (with a  
> differenet namespace) with the changes, so they can progress their plans  
> to move to genericode.  My hope is we can get this official update out  
> somewhat quickly so that we can migrate ICG over to an official Schema.   
> That said, my updating of the specification will be the rate-determining  
> step, I suspect.
>
>> Given that this does not make instances of genericode 1.0  
>> non-conformant with the 1.1 schema, I suggest we keep the namespace URI  
>> string the same (as we are doing in UBL with UBL 2.x versions).  Does  
>> anyone feel that we need to change the namespace URI string?  Recall  
>> the versioning policy in the namespace document itself:
>>
>>    http://docs.oasis-open.org/codelist/ns/genericode/1.0/
>
> OK, I'll leave the namespace the same, but update the version number in  
> the Schema itself to 1.1.
>
> Cheers, Tony.

-- 
Anthony B. Coates
Associate Director
Document Engineering Services (Limited)
UK: +44 (20) 8816 7700, US: +1 (239) 344 7700
Mobile/Cell: +44 (79) 0543 9026
Skype: abcoates
anthony.coates@documentengineeringservices.com
http://www.documentengineeringservices.com/