Re: Generalised approach to storing address details

From: JOG <jog_at_cs.nott.ac.uk>
Date: 11 Dec 2006 08:23:15 -0800
Message-ID: <1165854195.051022.25170_at_j72g2000cwa.googlegroups.com>


Marshall wrote:

> On Dec 10, 6:18 pm, "JOG" <j..._at_cs.nott.ac.uk> wrote:
> >
> > Nevertheless, being able to determine which predicates are fulfilled by
> > propositions containing a certain value would be a useful feature that
> > SQL lacks, and which ought be easily implementable relationally in a
> > DBMS such as Oracle, given its master relation.
>
> It's not obvious to me what that would be useful for. Can
> you elaborate a bit?
>
>
> Marshall

To ascertain what relationships an item is involved in. Say for example that I knew there was a relationship (as defined by its predicate) fulfilled by a proposition featuring both Wilma and Fred, but I did not know what that relationship was. It could be Parent/Child, Boss/Employee, Sister/Brother, or some more complex non-binary relationship, who knows. The functionality suggested above (via perhaps a mster_table with a 'relationship name' column and a relation-valued 'contents' column) would allow me to illicit relationally that the proposition concerning them exists in the 'marriage' table. I could then subsequently query that for more information.

However, given that one would then be handling a predicate as a first order object I suspect I might be promoting a system that handles 2OL, over which I understand the mathematical community is split. I haven't thought enough about this to offer much more comment on it than that I'm afraid.

Regards, J. Received on Mon Dec 11 2006 - 17:23:15 CET

Original text of this message