Re: Very simple question to relational theorists.

Evgeniy Grigoriev wrote:

> My questions again.
>
> Am I right that I can can use any new names c1, c2 if they are distinct?
>
> Can I use some mnemonic rules (only mnemonic!) to form the names c... of
> attributes of C relation from the names of attributes of A and B
> relations? For example name á1 is string "a1.b1"?
>
> Philip, where is difference?

Philip starts with the predicate and you never mention it.

The names in the relation header have to match the names in the predicate otherwise you are (literally) uttering nonsense.

Change the names in both to anything you like, for any reason you like (or none at all), but change them in BOTH you must. The only requirement for a name is that it be comparable (so you know when you're looking at the same name). Use pictograms, Cyrillic, Zapf-Chancery dingbats, Latin or Greek alphabets--no one and nothing cares, as long as the names used in the predicate and the relation header are the the same.

Of course if you choose names with some mnemonic value it will be easier to recall the predicate and relation you are working with, but that is a purely psychological issue. (I doubt anyone on c.d.t. has anything to say about psychology except that we see some amazing exhibits here from time to time!)

-- 
Roy