Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: curiousity of sets of no relations?

Re: curiousity of sets of no relations?

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Sat, 10 Jun 2006 18:35:15 GMT
Message-ID: <DhEig.20323$A26.465381@ursa-nb00s0.nbnet.nb.ca>


paul c wrote:

> Bob Badour wrote:
>

>> paul c wrote:
>>
>>> ...
>>> That seems a practical motivation.  In terms of relations and/or set 
>>> theory/predicate calculus can anybody give a more theoretical one?
>>
>> He simply defined them as the identity elements for the specific 
>> operations just as one defines any aggregate/fold of zero elements 
>> using the identity element for the base operation.

>
> In other words, I'm guessing, "to make it all work/to guarantee the ops
> always return relations", i.e., the motivation is practical only,
> analogous to notions such as the "empty product" (such as in
> http://en.wikipedia.org/wiki/Empty_product)? If so, I think I'm content
> with that, e.g., thinking of the empty set as just a gizmo to enable
> various operations we desire.

Exactly so. And exactly the same way Graham, Knuth and Patashnik define aggregation in _Concrete Mathematics_. We define the iteration over zero items as the identity element of the base operation. Sum() = 0, Product() = 1, Min() = max_value_of_type, Max() = min_value_of_type, etc.

Mathematicians define aggregations that way on utilitarian grounds. Received on Sat Jun 10 2006 - 13:35:15 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US