Re: Codd's Information Principle
Date: Sat, 7 Nov 2009 05:25:41 -0800 (PST)
Message-ID: <44c8e6da-4bcd-42b7-92ae-95be0f509bbe_at_37g2000yqm.googlegroups.com>
On 7 nov, 04:11, Tegiri Nenashi <tegirinena..._at_gmail.com> wrote:
> On Nov 6, 5:19 pm, com..._at_hotmail.com wrote:
>
> > Anyway, the quantifiers are not in the relational algebra,
> > they are in the corresponding predicate expression.
>
> The main objective of Algebraic Logic is eliminating the concept of
> quantifier.
True but in the context of RM, algebra is a tool not an end motive.
> The two success stories are Boolean Algebra (aka
> Propositional Calculus in algebraic form) and (Binary) Relation
> Algebra (corresponding to some fragment of Predicate Calculus?)
>
> Arguably, there is no ubiquitous algebraic system for Predicate
> Calculus despite Tarski, Halmos, and many others exerted quite an
> effort. (There is an inspirational essay by Halmos that I posted link
> on sci.logic a while ago -- cant find the reference!)
I would be glad to put my hands on it.
> Codd's
> relational algebra can be considered the first genuine algebraization
> of predicate calculus...
I agree. But my belief is Codd's refining process (using domains)
that led to RM final formulation is even more significant because it
allows to add combinatory analysis and probabilism in RM toolset (as
OO crowd would say *by inheritance*) .
Regards. Received on Sat Nov 07 2009 - 14:25:41 CET