Re: Which normal form is this violating?
Date: Mon, 6 May 2002 11:33:53 +0200
On Mon, 6 May 2002, Andrew Hamm wrote:
> Jan Hidders wrote ...
> >Andrew Hamm wrote ...
> >> I think he probably meant "distinct" sets... Don't let the argument get
> >> lost in pedantry.
> >I think we know what he meant (Joe Celko uses that phrase a lot) and I also
> >know what you meant. ;-)
> I couldn't hold my tongue when the argument got nudged off-course by a
> mis-interpretation. Some people seem to latch onto hiccups in the argument
> and then consider the argument wrong.
True. But my point was that the phrase "a schema is a set of tables" is in some sense true, but also a bit ambiguous and can lead to misunderstandings as you saw. It is indeed true that a schema should not describe two tables that are logically the same, i.e., represent the same predicate, because then one table would be redundant. So if you use the term "table" synonymous with "predicate" then it is correct, but if you define the term as "a set of tuples" then it is not. So if you are going to say that "a schema is a set of tables" and "a table is a set of tuples" in the same paragraph you have to make clear that the term "table" has two different meanings here.
So since exactness of speach was the point I felt that a little joke about the ambiguity of "the number of sets of integers" was not inappopriate :-) even if it concerns Joe Celko, for whose database expertise I have nothing but the highest respect, by the way.
- Jan Hidders