Re: [codelist] genericode 0.4 requirements

["Anthony B. Coates (Miley Watts)" <abcoates@mileywatts.com>]

Hi Paul.

On Mon, 09 Oct 2006 10:13:44 +0100, Paul Spencer  
<paul.spencer@boynings.co.uk> wrote:

> I have a couple of questions about the requirements.
>
> * I do not fully understand the difference between R4 (Support for  
> multiple, alternative codes) and F2 (Support for multiple alternate code  
> values for the same code list entry). How do these relate to a list such  
> as ISO 3166, which has alpha-2, alpha-3 and numeric-3 formats (e.g. GB,  
> GBR, 826)? I suspect that, if I don't understand the current wording,  
> some others won't as well.

Agreed, I should have expressed better.  R4 is the ISO 3166 case, where  
the same conceptual code list entries have alternative unique codes.  In  
the tabular genericode model, this means that there can be multiple keys  
for the code list.  F2 is where, within the one set of codes, some of the  
conceptual code list entries have multiple codes within the same key, i.e.  
there is a many-to-one mapping from key codes to a single conceptual code  
list entry.  F2 is a typical requirement where people are merging two code  
lists, e.g. an internal code list and an external code list for the same  
thing, and want to continue to use both the old and new codes (to support  
legacy systems).  They don't want the old and new code lists to be  
implemented as separate code lists for validation purposes, because they  
want to do the checking in a single validation, and don't care whether  
they receive an old code or a new code, as long as they can resolve it  
correctly to the underlying conceptual code list item.

> * Did I miss something, or is there no mention of code list metadata? As  
> well as each row having metadata, metadata such as list owner belongs to  
> the list as a whole. R28 (Metadata-only code lists can be represented)  
> implies that it exists (unless this just means that the column headings  
> exist), but I did not see explicit mention of it.

Agreed, I should mention it.  genericode does have metadata about the code  
list itself, and also supports metadata-only code list files.  It also  
implements an 'annotation' structure similar to W3C XML Schema at level of  
the whole code list, as well as at many lower levels within the code list.

Cheers, Tony.
-- 
Anthony B. Coates
Senior Partner
Miley Watts LLP
Experts In Data
+44 (79) 0543 9026
Data standards participant: ISO 20022 (ISO 15022 XML), ISO 19312,  
UN/CEFACT TMG, MDDL, FpML, UBL.