Re: Normalization, Natural Keys, Surrogate Keys

In message <ad26f4$pb0$1_at_sp15at20.hursley.ibm.com>, Paul Vernon <paul.vernon_at_ppsion.nnet> writes
>>This is a common problem and a lot of people have got together to fix
>>it. There are ISO standards for country codes. These codes are examples
>>of a surrogate key maintained by a trusted third-party, just like the
>>SSN.
>
>> Using the ISO codes relieves the developer of the need to dream up new
>>codes and ensures compatibility with other databases.
>
>I've not seen many (any) trusted third-parties that provide good
>relational models for their standard codes. (e.g. include domain
>definitions, full temporal history of changes to codes, guarantees that
>the codes will be static for say at least the next 4 weeks, some method
>for automatically receiving updates, ...)

I don't feel the need for a temporal history of changes or a guarantee that the codes will remain static for any specified period. I've run a standards body and we did all these things but I'm not sure how much use it was to our members. The method for automatically receiving updates is important and most standards bodies have some such system. We were moving towards using a published XML schema for that.

>
>I mean, you even have to PAY for most of the ISO codes. Are they sure
>that's the best way to promote standardisation?!

Not necessarily the best, but it's available now. Would you prefer a perfect system manana?

-- 
Bernard Peek
bap_at_shrdlu.com

In search of cognoscenti