Re: Extending my question. Was: The relational model and relational

From: Bob Badour <>
Date: Sat, 22 Feb 2003 14:22:56 -0500
Message-ID: <EuQ5a.219$>

"Steve Kass" <> wrote in message news:b372u8$c5q$
> Are you familiar with the Millikan oil drop experiment? I'm
> not aware of a more famous example of finding the cardinailty
> of a collection of indistinguishable objects.

You are forgetting two things: 1. Millikan first had to identify individual drops of oil and 2. His experiment did not count the number of charged particles (ie. the cardinality of electrons and protons) in any drop of oil.

As such, the experiment successfully showed that charge comes in discrete quanta but did not establish the cardinality of charged particles. Granted, the discrete unit of charge has been used to estimate Avogadro's Number, which of course does involve cardinality. It is, however, only an estimate.

Let's ignore the deficiencies of your example for a moment and let's assume that, by first identifying discretely quantized properties, we can sometimes devise methods to derive cardinalities without identifying individuals. What are the discretely quantized properties in a multiset logical model that allow a dbms to derive the cardinality of elements lacking logical identity?

If multisets are useful to you for database management, how exactly are they useful? How does this utility sufficiently compensate for the lack of logical identity and for the lack of physical independence? Can you offer anything specific that has direct relevance to any theory of database management? Received on Sat Feb 22 2003 - 20:22:56 CET

Original text of this message