> Interestingly enough, contrary to popular belief, the goal with DB design > (normalisation), more tables is better than less tables - if you do proper > data normalisation.
That is more of a side effect than a goal. Take a simple set of data you want to store consisting of a name and an age. 'bob', '23' 'sam', '27' You could store this in one table with two columns (name:varchar(X), age:int). Of course you could have more than one bob, so you'll probably have a third column (myIdentity: int, identity(1)) Now, this is simple and it perfectly models the data we want to store. Splitting it into multiple tables, say one with myIdentity and name, the other table with the foreign key reference to myIdentity and an age, doesn't improve our design. In fact, it's worse because now we are storing a 4th column that we didn't need (the repeat of myIdentity). Having said that, in general, yes, more tables generally means a better normalization, but "more tables" is a side effect of normalization, not the goal itself.
