Re: Surrogate Keys: an Implementation Issue

>> I doubt that there are any such rdbms's today. It is like criticism of
>> RT because of SQL shortcomings - the lack of faithful implementations
>> only serves to encourage naive extensions to RT.

The idea of hidden surrogates and associated myths is mistakenly nurtured by many who are intimately familiar with current crop of SQL products. For instance, a cursory inspection of some sql programming newsgroups reveals several blatant claims like:

-- Surrogates are immutable
-- Surrogates should not be exposed the user
-- Surrogates help performance
-- Surrogates belong to the physical model
-- Surrogates are never verifiable in reality
-- ...

In most cases, claimant misses the fundamental that a surrogate key is nothing but a simple, stable and generally an irreducible candidate key. And that it has nothing to do with the physical model and that once used outside the DBMS, it is verifiable as well.

-- 
Anith