Re: Date's First Great Blunder

From: Neo <>
Date: 21 Apr 2004 11:50:27 -0700
Message-ID: <>

> > > What sort of problem are you talking about where RDM becomes impractical?
> >
> > Try implementing with RDM.
> I would have no trouble representing that data with RDM.

In general, models "match" reality within a scope. The more specialized a model's scope, the more efficient it is within that scope but conversely the more miserably it fails outside of that scope. I agree that RDM has a scope which covers many common situations. The above example is designed to push RDM beyond its practical scope. XDb1 (a partial implementation of TDM) has a broader scope than RDM. Thus XDb1 is less efficient within RDM's scope.

> Writing the program to find common ancestors would take a little longer.

This problem has been intermittently posted for over a year now and no one has yet presented an "equivalent" RDM solution yet, so I am willing to wait.

> It would involve writing recursive procedural code.
> Presumably your Xdb1 program does something similar under the covers?

Yes, recursion is very important. The degree to which a data model allows recursion is related to its scope. NULLs hinder/kill recursion. XDb1 is highly recursive and has no NULLs. Received on Wed Apr 21 2004 - 20:50:27 CEST

Original text of this message