Re: Foreign key problem

From: paul c <>
Date: Thu, 15 Jun 2006 12:40:49 GMT
Message-ID: <lzckg.29639$iF6.29187_at_pd7tw2no>

x wrote:
> "paul c" <> wrote in message
> news:7RWjg.26165$IK3.10314_at_pd7tw1no...

>> Rich Ryan wrote:
>>> ...
>>> At the end of the day, foreign keys is an  implementation issue. But Ted
>>> Codd had a lot to say about the domain of foreign keys. The "glue".

>> If you are saying that foreign keys are THE "glue", the fact that nobody
>> can seem to apply them to the OP's schema without denormalizing suggests
>> they aren't always so.

> THE "glue" are the domains, not the foreign keys. See RM/V2.

Maybe I should have said that the word glue is a time-waster without any practical use, one of those casual metaphors, so rife in IT, that end up taking on a life of their own. Some people think keys hold a database together, as for what holds a relation together (in a database), I'd say it is the complete definition of whatever theory, eg., relational theory we use. One could just as easily say that the defined operators are the glue, eg., JOIN.

Salesman stole my copy of RM/V2, no idea why. Out-of-print now, not to be found in any library I've searched. Maybe somebody could quote what it says about glue!

p Received on Thu Jun 15 2006 - 14:40:49 CEST

Original text of this message