Re: more on delete from join

From: Marshall <>
Date: Tue, 1 Sep 2009 22:18:44 -0700 (PDT)
Message-ID: <>

On Sep 1, 8:12 am, Kevin Kirkpatrick <> wrote:


> Yes - the "sound" approach would seem to be: the DBMS should only
> accept propositions which exactly conform to the predicates of the
> underlying base relvars.  If there is a need within the user community
> to record propositions which do not exactly fit the existing base
> relvars, then the data model does not properly reflect the Universe Of
> Discourse, and needs to be modified accordingly.  If there is a desire
> within the user communit to record propositions which do exactly fit
> existing base relvars, then the end-user (or the application they are
> using) should do so with precise assignments to the appropriate
> relvars.

Do you have anything to justify these requirements? They sound like plain dogma to me: I can discern neither practical advantage nor theoretical basis for these assertions.

Eschew dogma. Embrace what works.

Marshall Received on Wed Sep 02 2009 - 07:18:44 CEST

Original text of this message