Re: database market share 2003

Paul Vernon apparently said,on my timestamp of 17/06/2004 10:56 PM:

> That's arguable. In some ways Codd's greatest achievement was simply showing
> that theory could be usefully applied to databases. Most pre-relational
> databases did not have much theory behind them, certainly nothing that could
> be regarded as a complete model. A better characterisation would be that the
> implementations and/or specifications came first, and only later did the
> more theoretically minded try to elicit abstract models from the mess of
> detail.

Dunno. Codasyl seemed to be pretty well established. Note: the point I'm making is not if relational is better. We all know it is. The point is if it was the _only one_ with a sound theoretical base. IMHO, relational had one big advantage over the others: a sound mathematical foundation. But theory behind them, they all had in heaps.

> a.. First of all, Codd realized that to compare the very concrete CODASYL
> specifications and the much more abstract relational model would be an
> apples-and-oranges comparison and would involve numerous distracting
> irrelevancies.

Agreed.

> a.. Hence, it would be necessary first to define an abstract "network
> model." The comparison could then be done on a level playing field, as it
> were, in a fair and sensible manner.
>
> a.. Codd therefore proceeded to define an abstraction of the CODASYL
> specifications that might reasonably be regarded as such a model.

and in the process took some "liberties" that were never defined in the Codasyl standard... Nothing wrong with that: he was just trying to explain why relational was so much better. Which is true.

> Thus, Codd has some claim to being the first person to attempt to give an
> abstract definition, not just of the relational model (of course), but also
> of a network model! Certainly none of the CODASYL documents ever attempted
> any such thing

Beg to disagree. Codd gave it one _interpretation_. But the Codasyl model was as well defined as it could get, outside of formal maths.

> BTW if you want to continue this arm of the thread, I suggest you start a
> new thread in comp.databases.theory.

Narh, thanks. I'd rather talk to you folks than have to endure that incompetent abortion called Celko. Or the "OO-expert-du-jour", an even worse specimen if that is possible.

-- 
Cheers
Nuno Souto
wizofoz2k_at_yahoo.com.au.nospam