Re: Concurrency in an RDB - another question about recursive definitions

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Sun, 21 Jan 2007 17:43:03 GMT
Message-ID: <HCNsh.3601$1x.61816@ursa-nb00s0.nbnet.nb.ca>

paul c wrote:

> Bob Badour wrote:
>

>> paul c wrote:
>> ...
>>
>>> Given relation : B = { {a,b} | a in A and b in B } /* Using C-style 
>>                                                  ^
>>                                                  |
>> Relations are not domains -----------------------/

>
> Thanks and okay, to put it bluntly, I take it that using B to stand for
> a relation in this way is nonsense.

Let's just say confusing a relation with a domain is a blunder. Using the same name to refer to a relation and a domain is not necessarily a blunder provided everyone agrees that relations and domains with the same name are not the same thing.

Consider a system with Person, Name, Date and Integer domains. Might the following compile?

CREATE RELATION Person (

   Person Person

, Name Name
, DOB Date
, #Dependents Integer

)

It might compile provided the language can always contextually distinguish between relation names, attribute names and domain names. Would it be a good idea to create such a relation? That's a matter of opinion.

>> One can describe the extent of a domain as a relation.
>> The extent of domain B is infinite.
>> ...

>
> And I guess it should go without saying that domain B is
> static/unchanging too, if we want persistent relation values.

It is no more or less static than the domain Integer. I agree it should go without saying; however, it should be said if necessary or helpful. Received on Sun Jan 21 2007 - 11:43:03 CST