Re: Normalization Question

"Tom Ivar Helbekkmo" <tih+nr_at_eunetnorge.no> wrote in message news:86acqzkkn5.fsf_at_athene.hamartun.priv.no...
> pc <toledobythesea_at_oohay.moc> writes:
>
>> in that case, what would be the point of the model?
>
> If I understood Dawn correctly, exactly the same as before such a
> change. The only difference would be that a one-to-one relationship
> (from a referencing row attribute to a referenced row in another
> table) could also be a one-to-many relationship, without introducing
> an intermediate, two column, relationship table, simply by letting the
> original referencing attribute hold multiple values.
>
> Off-hand, I can't see why it wouldn't have worked.

Yes, you did understand -- thanks!
Perhaps the new variations on relational theory that permit relation-valued attributes allow for this more flexible model. However, it seems that those working with today's RDBMS's, even if they support RVAs or multivalued attributes, don't seem to model data that way. They still use what they learned as first-normal form (on which all other normal forms claim to be dependent).

When modeling for many of the existing non-RDBMS's professionals have been able to make multiplicity changes of this nature without significant changes to the model for years. This has been going on since data were first modeled for computer implementations. It is starting to "feel" like the industry is moving forward (and back) to this approach, but a bit of a shame that it took XML to get us back to our more flexible roots.

cheers! --dawn

>
> -tih
> --
> Tom Ivar Helbekkmo, Senior System Administrator, EUnet Norway Hosting
> www.eunet.no T +47-22092958 M +47-93013940 F +47-22092901 FWD 484145
Received on Wed Jan 26 2005 - 15:39:01 CET