Re: Expressing SQL in relational algebra

Steve Kass <skass_at_drew.edu> wrote in message news:<b51638$fps$1_at_slb3.atl.mindspring.net>...
> The question had nothing to do with what you think makes sense.
> So far as I know, in relational algebra, SQL, algebra, or just about
> anywhere else, T_Part is a valid identifier that works just as well
> as Part and exhibits the same behavior when neither identifier is
> bound to anything else in the scope.
>
> Just what about identifier names is supposed to or required to make
> sense, and why does T_person or T_Badour fail the test?

I added you to my ignore list so I don't usually see your posts. I am forced to use deja tonight, so the ignore list is inactive.

Prepending every relation name with a fixed prefix reduces brevity without any compensating benefit. The "T_" part of the name is truly meaningless, and if you are arguing that abstract, meaningless names are appropriate in simple examples, I would agree and suggest an abstract name that is wholly and clearly abstract. By informal convention, foo is often used for this purpose in computing examples, and single letter names are often used in mathematics.

I still see no sensible reason to prepend every relation name with "T_", and I observe that your objection makes no attempt to provide one. Received on Wed Apr 02 2003 - 08:50:55 CEST