Re: Is Inheritance in RM possible?

From: Roy Hann <specially_at_processed.almost.meat>
Date: Wed, 14 Dec 2005 17:00:03 -0000
Message-ID: <Gv2dnZ7GDZ920j3eRVnygA_at_pipex.net>


"JOG" <jog_at_cs.nott.ac.uk> wrote in message news:1134575938.366961.78280_at_f14g2000cwb.googlegroups.com...
> Does inheritance currently exist in RMDBMS anyhwere? I'm not referring
> here to OO or O-R databases, but rather the addition of the
> functionality in an RM system. Is it even possible or at odds to the
> underlying principles of the model?
>
> I have heard mention of the INHERITS keyword in Oracle and PostreSQL
> but having only very limited experience with the respective products, I
> have no idea of the extent or relevance of its functionality.
>
> I've also encountered Darwen's proposal for an inheritance model via a
> google seach, and a trawl to thethirdmanifesto.com, but outside that I
> am need of illumination. Any experiences of those in this group are
> much appreciated.

No doubt it depends which school of RM you mean--but perhaps it doesn't. AFAIK Codd's RM doesn't give a fig about inheritance. Inheritance is a completely orthogonal concept. RM is concerned only with predicate logic (c. 2000 years old) and set theory (c. 300 years old). Those are both pretty mature technologies and no one has bolted inheritance onto them in all that time. I assume they haven't seen the need.

Date/Darwen and other of course do incidentally talk about type systems, and inheritance is on that particular agenda. I don't know that RM sheds any light on those problems, nor is it enriched by the results.

Roy Received on Wed Dec 14 2005 - 18:00:03 CET

Original text of this message