ON "BEYOND REASON"

From: William Sisson

Date: 12 Nov 2005

I read your recent article Beyond Reason with much interest as I have had some long and heated discussions with someone developing applications for PDAs about whether or not he should use a DBMS. It is difficult because his background is programming application specific devices with embedded operating systems and he has almost no DBMS experience. I can therefore forgive him some of his misconceptions while at the same time trying to persuade him.

What makes it much harder is the amount of material around of a similar nature to that put forward by Margo Seltzer.

Now it seems to me that most of the promised mobile services could not possibly be implemented reliably without the use of an RDBMS. As you correctly point out, where the data sits physically is irrelevant. If what we are talking about is a distributed database then it is difficult to see how one could possibly synchronise the data correctly without adopting the principles of the relational model. If the question is whether we need better embedded DBMSs and better implementations of distributed DBMSs then I say a resounding yes, but as you say, this is probably not what Seltzer is talking about.

Now, I do not have a PhD from a prestigious university and my academic background is neither mathematical nor in computer science but these things seem obvious to me, especially now I have had the opportunity to learn from the excellent material offered on your web site. So I struggle to comprehend how Seltzer can come to the conclusions she does.

I can, in a way, understand the reasons for this particular PDA developer's resistance to using a DBMS on the handheld device (though I do not agree with them), but it is disheartening that people like Seltzer are there to confirm some of his prejudices.

Thank you again for setting matters straight in the article. I only hope enough people are listening.

By the way, I know I said previously that I did not want anything I communicated to you published under my name. I have changed my opinion on this now and no longer have a problem with publication under my name rather than anonymously. Of course it is entirely up to you to decide whether you think I have anything to say that is worth publishing.

From: Fabian Pascal

Beware of academics in business. They were not real academics in the first place.

If there is any distribution of data, the ONLY technology that can handle it is relational. In fact, that was one of Codd's declared intentions, among others of course.

Those who do listen, like you, already know it, and do not have much influence.

Good for you.

Posted 2/3/06