Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: Example of expression bias?

Re: Example of expression bias?

From: Tony D <tonyisyourpal_at_netscape.net>
Date: 20 Jun 2006 06:40:30 -0700
Message-ID: <1150810830.580876.203440@h76g2000cwa.googlegroups.com>


Cimode wrote:
> You defined relevance of FP by its usefulness.

No; you enquired what FP would be useful for, and started rambling about implementation.

> I simply responded that soundness is not determined by usefulness...
> I do not have time to do reading on FP.

Your loss.

> The little I have been exposed to was sufficient to
> consider it irrelevant to data management issue. All the nonsense you
> produce about data management concepts do not encourage me any more...
>

What nonsense would that be, precisely ? So far on data management concepts, we've managed to agree to differ on perspective on the definition of relvars and relations (I accept Date's position, with which you are not comfortable). You seem to think Date's definition can cause confusion between the definitions of attributes and variables; I don't see how this confusion can arise. We disagree on the definition of a data type (you hold domains to be a pool of possible values, distinct from types; I don't see this as a useful distinction and don't accept it). Please describe the nonsense therein, given the position that data types are orthogonal to relations.

> Good. All I wrote to describe what is a sound implementation model was
> totally useless.
> The predicate is
> Mathematical expression 1(is implemented as) Logical Model for data
> management1 (is implemented as) Implementation Computational
> Model1...For RM, it comes to
> Mathematical Relations(is implemented as) RM (is implemented as) TRM?
> (Last one not confirmed)...
>
> What is the occurence of predicate you state to give me any interest in
> FP.
>

You're bashing on about implementations again; as I have said on *several* occasions, implementation issues are of no interest to me.

However, we have (I think) agreed that user defined types are essential to a useful relational system; therefore, we need some way of defining those types and the associated operators. That's where I want to use functional programming. (In fact, I want to use an FP system which just happens to offer full support for relations.)

> Huh? What the hell are you talking about...?
>

I mean, you are being completely irrelevant.

> So what is your point? I will not get back to higher order functions.
> relations are sufficient to describe more simply the concepts refered
> to as higher order functions...
>

The point is, Erwin asked about where he could learn about higher order functions. Marshall gave him some references. You bashed Marshall in a completely irrelevant way, and cast ignorant aspersions on the soundness of the basis of functional programming (that is, the lambda calculus). I pointed this out to you, and advised you to read up on the lambda calculus before making any more ignorant pronouncements about it. You have responded in your usual delightful manner. Hopefully Erwin is reading about higher order functions in blissful ignorance of yet another trail of useless postings involving you being irrelevant and insulting with it.

> I am not shaking anything except your delluded brain...To encourage get
> significant education about RM which has been proven sufficient to
> abord the problem of formal representation of information in mechanized
> systems...
>

RM answers some problems. It answers those problems better than any other available model. But it doesn't answer *every* problem. So it will have to be joined with "something else" to answer other problems. Given the mathematical basis of RM, and the mathematical basis of the lambda calculus, wouldn't you agree that there is more chance of those resulting in a mathematically sound, reasonable union than say, RM and OO ?

> undecideability relies on a concept that is totally off scope for data
> management issues...

So, computability theory isn't relevant to data management issues ? If your data management involves pen and paper, maybe. As soon as a computer is involved, computability is more relevant than you care to think.

> Don't you get it?
>

Yes. But I don't think you do.

> As a lot of people, you probably perceive RM through misinformed SQL
> audiences who have limited the relevance of RM theory.
>

This is one of your own "prepackaged notions" that you fall back on with monotonous regularity. And, for your info, no, I don't.

> It is perfectly (fill the predicate proposition and you will understand
> hopefully the chain of reasoning)

The chain of reasoning (such as it was) was irrelevant to the discussion at hand. Received on Tue Jun 20 2006 - 08:40:30 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US