Galen Boyer wrote:
> On 9 Oct 2004, wizofoz2k_at_yahoo.com.au wrote:
>
>>"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message
>>news:<41675d4e$0$10350$afc38c87_at_news.optusnet.com.au>...
>>
>>
>>
>>
>>>Count me in very firmly as a Natural Key man. But I respect
>>>your right to eat l'escargot, too.
>>>
>>
>>Wait until you've had to design, code and get accepted by
>>the users a few dozen systems. Then tell me if you still
>>like natural keys. I used to like them too. Until I had
>>to wear the consequences of incomplete relational models
>>and incomplete implementations. Reality has this nasty
>>way of catching up with the best theory...
>
>
> What does one do when the "natural keys" are actually different
> based on the source?
>
> Maybe someone needs to define a "natural key". Seems to me, it
> is a key that some other system has already defined.
The natural key is well defined by the work done by Date and Codd.
That which uniquely defines a record in one set does not necessarily
define a unique record in another set.
--
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu
(replace 'x' with 'u' to respond)
Received on Mon Oct 11 2004 - 22:43:41 CDT
