Re: foundations of relational theory?
Date: Thu, 23 Oct 2003 11:25:25 -0400
Message-ID: <3F97F2E5.5080202_at_quest.with.a.w.net>
Paul Vernon wrote:
> "Anthony W. Youngman" <thewolery_at_nospam.demon.co.uk> wrote in message
> news:T1ljmMCJ9ul$EwsP_at_thewolery.demon.co.uk...
>
>>EXCEPT this was an ACADEMIC study, not a MARKETING study, and as I >>understand it, the researchers were surprised by the result. >> >>They picked a bunch of large companies with the idea of calculating db >>spend as a proportion of turnover. No bias there. They plotted those two >>figures, and were surprised to get a double peak. No bias there, either. >> >>When they investigated this unusual phenomenon, they discovered the low- >>spending peak was almost entirely Pick-based dbs, the high-spending peak >>was relational.
>
>
> What did they expect? If Pick was as expensive to run as Oracle, then nobody
> in their right mind would run it.
>
> I.e. It's market niche is low cost and low functionality.
>
> Or are you saying that it is as *functional* as an Oracle or DB2, and the
> only reason that it has not taken over the world is that old war horse:
> 'poor marketing' ?!
Actually, Paul, yes. Just because it's 'an old war horse' does not mean it cannot be true. And your comment about Pick being 'low functional' flies in the face of the sum of the reason and experience of us in the Pick list, who have used it extensively for thousands of complex (and modern) apps.
Pick was popular before Larry Ellison expertly trumpeted the virtues of Relational, and 15 years ago Pick apps could (as has been endlessly bruited here) run rings around anything else on *both* a functional *and* an economic level. The latter advantage has somewhat faded due to the exponential drop in hardware cost, tho' Pick apps are still probably an order of magnitude simpler to create and maintain. You might dismiss this as 1) anecdotal (enough anecdotes and you'd see a trend,) and 2) not relevant to the theoretically pure, mythical Relational dbms (as it compares Pick to Oracle/SQL etc.) but that's what our bias is-reality, not theory.
In other words, you can't have it both ways. Either we compare theories, in which case your side has all the weapons, or we compare reality, where we have the edge. It's not fair to dismiss a comparison you don't like by saying "don't use SQL/Oracle as an example, that's not the same as Relational."
Chandru Murthi
>
>
> Regards
> Paul Vernon
> Business Intelligence, IBM Global Services
>
>
Received on Thu Oct 23 2003 - 17:25:25 CEST