Re: two nasty schemata, union types and surrogate keys

On Nov 8, 6:37 pm, r..._at_raampje.lan (Reinier Post) wrote:
> Brian wrote:
> >Under the closed world interpretation, there are no unknown truth
> >values, but under the open world interpretation, only what has been
> >explicitly asserted is known to be true, and it is known to be true
>
> Yes ...
>
> >even if the user that made the assertion is mistaken.
>
> Huh?!  So if I have a database relation 'X works at Y', with the open
> world assumption, if someone updates the relation to say that Jane Doe
> works at Acme Corp., then I must assume this to be true even if that
> person is mistaken?  How does that make any sense?

Yes. If you don't also record who uttered the assertion, then the supposition implicit in its representation in the database under the closed world interpretation is that the assertion is true or under the open world interpretation is that whoever uttered the assertion knows that it is true. This makes sense because it is not known that the person is mistaken. If a bank teller counts his drawer at the end of the day and is neither over nor short, then the assumption must be that every customer received correct change, but it could be that one customer received a dollar more than he should have and another received a dollar less.

>
> >[...]  Just because something
> >is known to be true doesn't mean that it actually is true.  Even under
> >the closed world interpretation, what is supposed to be true may not
> >actually be true.
>
> Exactly.  There is no difference between OWA and CWA in this respect,
> regarding the assertions that correspond to tuples in the relation.
> The difference is in what they imply for assertions that correspond
> to tuples not in the relatoion.
>
> --
> Reinier
Received on Mon Nov 09 2009 - 03:06:43 CET