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

On 2014-12-21, Ruben Safir <mrbrklyn_at_panix.com> wrote:
> Eric <eric_at_deptj.eu> wrote:
>> On 2014-12-18, ruben safir <ruben_at_mrbrklyn.com> wrote:
>>> ...
8>< ----- many things snipped

A lot of compdb's responses are things I might have said, though probably in a different way.

> Despite that, thank you verymuch for taking the time to clarify some of
> these ideas. Looking at this chapter, the subject itself should be a
> semeters woth of work centered on Cobbs theories and model.

No problem, you hit one of the days when I feel like answering things.

BUT...
... what worries me the most now is that you wrote "Cobb" - twice. The man's name was Codd, so it does create a sort of "If he can't get that right..." impression.

8>< ----- snipping some more

> It then includes the entire universe? Star Dust and Dark matter
> included, if I understand the theory correctly.
>
>> It is free (note that word - "FREE") to take
>> any value at all.
>>
>> To paraphrase part of the "Formally" quote above, if a tuple variable
>> is not free...
>>
>> ...it is bound!
>>
>
> But it still has degrees of freedom. Everything is bound to something.
> Everything is free within its bounds of degrees of freedom.

Good point. Though the "universe" of a database is only those things that it is designed to store, not the star dust and dark matter. Free and bound here are merely relative to that "universe of discourse".

8>< ----- snipping some more

>> It matters because this sort of formula will end up representing a
>> relational query, whose result is a set of actual values of a tuple
>> variable. Before you construct the actual query (or formula) the tuple
>> variable is free to have any value - you don't know what the answer
>> will be. Afterwards, because of what you put in the formula, the tuple
>> variable will be bound - to the answer you need!
>>
>
> That is good! You should write the book.

Thank you!

8>< ----- snipping some more

> Let me think on this. I I think he is saying the tuple variable is
> bound to the quantiers definition.

1. In the formula F, the variable t is not bound, it could take any, all, or none of the values possible in the universe of discourse.
2. F' = (∃ t)(F) is a different formula and in this formula t is bound because "none" is no longer allowed.
3. The variable is bound by the presence of the quantifier in F', not to the definition of the quantifier.

> That would apear to be simple but why chose only these two quantifiers.
> I think Universal is really throwning me off.

What other quantifiers would you have? And what's wrong with universal? (∀ t)(F) just says that F is a fact, true for all values of the variable t.

8>< ----- snipping some more

Eric

-- 
ms fnd in a lbry