# Re: A pk is *both* a physical and a logical object.

Date: Wed, 18 Jul 2007 09:41:40 +0100

Message-ID: <FYCdnStLB9jYTwDbnZ2dnUVZ8rOdnZ2d_at_pipex.net>

"Brian Selzer" <brian_at_selzer-software.com> wrote in message
news:_Nfni.26756$2v1.2010_at_newssvr14.news.prodigy.net...

*>
*

>> Under the closed world hypothesis, the only sensible reason to do the

*>> update you describe to a row in a relation in which the entire header is
**>> the key would be to retract a falsehood. I have no interest in
**>> falsehoods and I don't see how they are related to "individuals"
**>> (whatever they are). What am I not getting?
**>>
**>
**> What does the closed world assumption have to do with it? The closed
**> world assumption simply requires that if a tuple that does not violate the
**> predicate of a relation is not contained within the particular relation
**> value at a particular world, then the atomic formula represented by the
**> tuple is false at that particular world.
*

Yes, exactly so. But a database exists only to represent what is said to be true about the real world. The real world is the only "particular" world of interest. If any particular database represents a falsehood about the real world or is updated to represent a falsehood about the real world, it is of no use and no interest.

> An update selects which possible world is actual;

>therefore, it operates independent of the closed world assumption.

Notice I used the phrase "the only sensible reason" above. We are not interested in random updates.

> Any possible world can become the actual world, so it follows that each

*> possible world should be closed with respect to itself.
*

Get a grip. We are talking about databases. Databases are not abstract collections of symbols that can be manipulated willy-nilly. They are collections of symbols that we seek to manipulate in the safe knowledge that we always end up with a representation of what we think is true in the real world.

Roy Received on Wed Jul 18 2007 - 10:41:40 CEST