Re: On view updating

"Paul" <paul_at_test.com> wrote in message news:41590f52$0$69731$ed2619ec_at_ptn-nntp-reader01.plus.net...

> Having said that though, I agree that it might be too complex for
> business data. But the point is (as made by Costin) that then we can't
> claim that the relational model is a full implementation of first-order
> logic. We are deliberately constraining it to be not so, so that things
> are easier (on both the user and the DBMS designer).
>
> What I'm exploring as well is whether constraints can be used in any way
> to store this kind of information.

Earlier I suggested that partial license plate information in criminal records might be a good example of a database of incomplete knowledge.

Perhaps an even better example might be a fingerprint database. Many of the fingerprints in such a database are incomplete. I wonder how they organize fingerprints in such a way that the FBI can say that a new print does not match any of the prints on file. Surely they don't scan the entire database!

What about a database of biometrics for identification... aren't we going to have to deal with partial information in such a database?

I don't have any ideas on partial databases, but I have two places to start looking.

One of the normal forms, I think it's 4th, deals with partial dependencies. So maybe an exploration of that normal form, and the ways of dealing with it would yield some interesting observations about how to deal with partial information.

The other is Claude Shannon's work on information and uncertainty. This work has primarily been used in data communication, rather than in database theory. But maybe it's more relevant than we think. After all, I've always thought of databases as channels that carry messages across time, to an unknown receiver. It's sort of like the person who seals a message in a bottle and throws it in the ocean.

Unfortunately the messages across time are simplex.

Just thinking freely... Received on Tue Sep 28 2004 - 10:59:04 CEST