Re: O'Reilly interview with Date

Paul wrote:
> dawn wrote:
> > 2. The way that these alternatives are tossed aside is by marrying the
> > models to hierarchical and network and then dismissing those as having
> > failed already. I realze you cannot put everything in an interview, but
> > I have read quite a bit of Date's writings and haven't seem much more
> > rationale than what is given here.
> >
> > Is there a mathematical argument that shows why these approaches should
> > be tossed aside as having nothing to offer?
>
> I think it comes from practical experience of these models being
> complicated to use rather than any theoretical argument. I've never used
> hierarchical or networks database but I understand people who did found
> relational database much easier to use and develop.
>
> The old-style databases must have been theoretically correct at some
> level otherwise no-one would have used them if they were costantly
> giving wrong results.

I made a pitch in 1986 to do a pilot project with db2 instead of IMS. In retrospect, the person I disussed it with was right to dissuade me and not jump into db2 at that point. My rationale was that db2 was relational, so it would have less complex schema and be more flexible, working only with relations etc. In fact, my rationale sounded a whole lot like many on this list today. How far I've wandered, eh? But IMS is a far cry from MUMPS, for example (I say "PICK" too often) and both might be considered hierarchical. Just because one implementation that appears relational isn't agile, doesn't mean anoter isn't.

I like theory based on practice, but that is not the same thing as proving that other theories are not as good. I haven't seen such a proof from any relational theory camp, just bad-mouthing of tree or di-graph models, dismissing them as "hierarchical" and "network". It seems like buzz and not logic.

thanks. --dawn Received on Thu Aug 04 2005 - 05:05:01 CEST