Re: Surrogate Keys: an Implementation Issue

Jay Dee wrote:
> JOG wrote:

>> Brian Selzer wrote:
>>
>>> "Bernard Peek" <bap_at_shrdlu.com> wrote in message
>>> news:xn0ep1e3sn11oo000_at_news.individual.net...
>>>
>>>> In comp.databases.theory Paul Mansour wrote:
>>>>

I'm not sure that everybody here is using the same notion of 'surrogate'. I gather Codd said it was not to be exposed and I've seen to that effect in his 1979 paper. However, some of the big names object to the theme of that paper (personally, I wonder whether some semantic model isn't necessarily another 'application' inserted 'under' the 'end-user application').

Date points out that if users have some control over surrogates (eg., if they can create or reference one), then the information principle would seem to require that they be exposed (my words, apologies to Date if I misinterpreted).

Which is it? (I don't know where to find the 'glossary' for this group.)

p Received on Tue Jul 25 2006 - 03:39:33 CEST