ISO Country code content

[Dan Vint <dvint@acord.org>]

We talked today about what the real content of the ISO spec is vs what is 
displayed on-line as a reference and I think teaser to get you to buy the 
full spec. Here is the page with some more details about what is in a 
database version of the spec that you can buy.

http://www.iso.org/iso/en/prods-services/iso3166ma/05database/db-iso3166-1.html

These are the specific fields referenced:
    * Alpha-2: ZA
    * Alpha-3: ZAF
    * Numeric-3: 710
    * English short name: SOUTH AFRICA
    * English full name: Republic of South AfricaPrince-Édouard
    * English remark: Includes Marion Island, Prince Edward Island
    * French short name: AFRIQUE DU SUD
    * French full name: République sud-africaine
    * French remark: Y compris: l'Île Marion, l'Île du Prince-Édouard
It looks like there are 3 different ways to reference a country, 2 letter 
code, 3 letter code and then a numeric value. What is on the website is the 
2 letter code and what is referred to as the English Short name. Notice 
that there are also remarks that really clarify what is included by this code.

First it seems like we would want to capture as much of this information as 
an outside organization was will to provide, which is something we probably 
need to bring back to NDR.

Next about our formatting requirements.

I do agree that this should be appinfo, but I don't think everything that 
is currently in the documentation should all be moved to appinfo. If we use 
the above code you would see that the description should include any 
remarks that are available. We also have to decide what is a minimal 
description that should be included and formatted. In this case the English 
full names seems more appropriate. The documentation section should be as 
friendly and descriptive as possible, I've come to think that the direct 
translation of the code should not come from here.

Instead the translation is for an application to use, so lets move it to 
the appinfo section. Here instead of being constrained with XHTML, I 
believe as Ken that we should have a controlled schema and namespace to 
describe our different types of usages/translations. Here are some ideas 
for what might be needed:
         - <shortCode>
         - <longCode>
         - <shortText>
         - <longText>
         - <symbol>

I would use the first 2 items to control the fact that US and GB are the 
code values but I might want to print USA or UK. The short and long text 
would be various expansions on a name in the actual presentation (camel 
casing), so for US, I might have United States and United States of 
America. With these there might be some additional data to send, such as 
formal/informal and perhaps the length of the content (for easy reference 
instead of calculating. The formal/informal I think is an obvious 
classification, but I think the length would be useful when several options 
might be available - I might know that on average field X will support 25 
characters, I would think it would be nice to find a short or long text 
based upon how well it fit the field requirements. just an idea anyway. The 
symbol element is where I would capture things like the currency symbol.

To be really complete, I would think that many of these might an xml:lang 
associated with them so alternate translations could be found as well.

For code lists we don't own, all of this is just a suggestion as to how 
best to provide this information. It would be nice to take a UBL owned list 
and see how all this would be applied and managed. I think it is useful to 
have this and here with the code list definition seems the appropriate 
place to manage it. But if no one would ever provide it then we may need a 
different way of approaching the problem.

Comments?


..dan