Re: Multiple specification of constraints

"Bob Badour" <bbadour_at_golden.net> wrote in message news:Mvqdndiu5esUZc3d4p2dnA_at_golden.net...
> "Mikito Harakiri" <mikharakiri_at_iahu.com> wrote in message
> news:7T24c.31$iI2.46_at_news.oracle.com...
> >
> > "Bob Badour" <bbadour_at_golden.net> wrote in message
> > news:gZCdnelwN-P8Lc3dRVn-jg_at_golden.net...
> > > We need to have logical identifiers. Once we have those, I fail to see
a
> > > problem.
> >
> > This contradicts to your earler statement (which I agree) that for every
> > value there are multiple physical representations. Solving identity
> problem
> > among different symbolic representations is by no means a triivial task.
> > What is "logical identifier" then?
>
> The representation is irrelevant to the value.

Aha, here is what I found when Googling "logical identifier":

"In general you don't allow users to update the logical identifier of an entity" -- Costin Cozianu

Therefore, since "entity" is meaningless concept from relational prespective, then "logical identifier" must be too!

Anyway, I should steer away of these discussions, as they go in rounds without any evidence of the progress (and with no entertainment too!). My apologies. Received on Fri Mar 12 2004 - 01:19:26 CET