Re: [LIU Comp Sci] Need tutoring on Relational Calculus

On Sunday, December 21, 2014 11:56:45 PM UTC-8, ruben safir wrote:
> On 12/21/2014 09:37 PM, compdb wrote:
> > (∀ t)(F) doesn't say anything unless it has no free variables.
>
> what does that mean?

Go back to the beginning.

My message was addressed to Eric. One can't understand it without understanding the section on tuple calculus.

> A variable is a storage unit that can take the value of data that can
> satisfy a condition. This is what the point is , right?

No. A "(tuple) variable" is some text in a tuple calculus expression. You are not noticing when the text introduces terminology. In 6.6.1 it appears at the very beginning, it is in boldface to indicate that it is a newly introduced term, it is described as a certain part of an expression and an example uses the letter "t" as a variable. That's all a variable is in this context: text that appears in certain places in a tuple calculus expression.

However this textbook is sloppy about introducing the term "variable". They're also really sloppy about the term "assignment" which they don't even define. These meanings of "variable" and "assignment" are not the ones from programming. They are from logic and historically preceded the meanings from programming, which borrowed the words with related but modified meanings. Also "universe" and "tuple in the universe" are not defined. (There are unclear attempts at definitions earlier.)

Section 6.6 is poorly written. I can image some of your difficulties and frustrations. My suggestion is to read the presentations in Maier, Alice and/or Wikipedia. (At least to get those terms explained.)

HOWEVER You tolerate lack of clarity far to long. Also you miss things. So try to read more carefully so that when content is there you won't miss it and when it is obscured you can hunt around for what they meant, not what comes to your mind. (In my experience people have a tendency to frequently jump to conclusions and not actually notice the details of what was actually written even when their interpretation is thereafter continually contradicted by what is further written.)

I agree that there are a lot of confused notions and presentations in the database field, both in industry and academia. It is one of my personal themes. Eg: There is no point in learning either the tuple calculus or the domain calculus in that one should just learn predicate calculus and the equivalence of relation (algebra/operator) expressions to predicate expressions (aka wffs).

philip Received on Tue Dec 23 2014 - 03:34:54 CET