Re: How to derive a consistent set of FOL constraints

On 3/5/07, John F. Sowa <sowa@bestweb.net> wrote:
>
> Yet every SQL database
> permits arbitrary first-order constraints.
>
> Q: How is possible to prove that the constraints are consistent?
>
> A: Trivially.
>
> The point is that no database designer *ever* begins with
> an arbitrary set of constraints.  They *always* begin with
> some actual data -- a sample DB that shows what kind of data
> they expect to work with.


Simple idea that comes to my mind immediately: the DB is a *closed  world*,
i.e. if something is not there, this does not hold.

It is different for the DL, which is an *open world*. That means if some
fact is not known to be true, it can be either true or false (in a
particular world). This means, that the search space is exponentially larger
in case of description logic comparing to DBs.


> John


Best,
Dmitry.

Received on Tuesday, 6 March 2007 20:38:53 UTC