Re: Multiple keys and transition constraints

Brian Selzer wrote:
> Given a relation schema R {A, B, C}, where A and B are
> each candidate keys.
>
> If the current extension is
>
> r {{A:1, B:9, C:3}
> {A:2, B:8, C:4}}
>
> and the proposed extension is
>
> r' {{A:1, B:8, C:4}
> {A:2, B:9, C:3}}
>
> is only A different? Or are both B and C different? From
> one perspective, both B and C remain constant but A is
> different. From another perspective, A remains constant
> but both B and C are different.

What is A? What is B? What is C? When you say "is only A different", what if I respond "no, A is A"? Is A a symbol denoting a set? If so then A denotes {1,2} in extension r and A denotes {1,2} in extension r'. Thus that which A denotes remains unchanged in r and r'; and in that sense "A is not different". And neither are B nor C different. Or does A denote a domain {0,1,2,...} for example? If so then perhaps A, B, and C never become "different".

Or look at it this way. A, B, and C do not become different any more than 1, 2, 3, 4, 8, and 9 become different. Has 1 changed? Or is only 4 different?

What is different are the two extensions r and r'; what is different are the tuples that compose r and r'; what is not different are the ever constant /values/ that compose those tuples.

This is the semantic gulf in which many of your arguments have drowned. When you say "is only A different?", it seems that you have in mind some implicit thing, other than A, that you hold /is/ A. Some physical storage perhaps?

It is the view you have, that A /is/ a thing other than a symbol denoting either a domain, extension, or property that separates you from many against whom you have butted heads recently. Because your notion seems rather physical. Ie that by A you mean some storage space or memory location. If it is not physical, then it's up to you to provide a clear (and concise) explanation of what you mean by A.

• Keith -- Fraud 6