Re: A different definition of MINUS, Part 3
Date: Sat, 20 Dec 2008 02:52:25 -0800 (PST)
On 19 déc, 20:54, paul c <toledobythe..._at_oohay.ac> wrote:
> Cimode wrote:
> > On 19 déc, 19:37, paul c <toledobythe..._at_oohay.ac> wrote:
> >> Cimode wrote:
> >>> [Snipped example]
> >>> <<As soon as they started
> >>> making premises, they lost me because it doesn't look like their
> >>> method is anything more than case analysis that have any chances to
> >>> be
> >>> scaled up to practical problems. >>
> >>> Precisely. I would go further as to say that hoping that using solely
> >>> algebra would be sufficient to achieve such resultis nothing more than
> >>> the holy graal of RL. As you mentionned in an earlier post, the
> >>> example of lack of quantifiersi is one major obstacle most current
> >>> DandD work seem to ignore without consequence or awareness of the
> >>> price to pay.
> >>> Regards.. .
> >> I think around 1972 Codd wrote a proof that the algebra was logically
> >> equivalent to FOPC, later others (I forget their names) corrected a few
> >> errors and proved the equivalence. This is why I'm happy to try to show
> >> things with the algebra, even though it can be tedious trying to see the
> >> forest for the trees.
> > To my knowledge, Codd never mentionned that FOPC would be sufficient
> > to clarify RL. Choosing to work *solely* on that angle is a matter of
> > personal choice.
> > Regards...
> I don't think that's what I said. I think Codd meant that his
> relational model obeyed (ie., is based on) both of FOPC and his algebra
Codd never, to my knowledge, closed the door in claiming that FOPC or algebra would be a sufficient tool to clarify all aspects or RL. By binding algebra to FOPC, he simply opened the door to an entire spectrum of investigations that could involve not only algebrical concepts but also non-algebrical concepts that can be expressed at some point through algebra. I explain Codd's choice of algebra to clarify FOPC as a choice for the expressive power of algebra as opposed to other mathematical tools.
> The personal choice is whether one uses algebra or fopc not only to
> understand his model but to define it.
or to redicover it and seeing the limits of the initial formulation (any great theory does have its limits). RL existed much before being expressed and clarified by Codd's algebra. Choosing to close that possibility is not what Codd had in mind when he opened the door. The fact that RL was immediately bound todatabase theory was purely a circumstancial fact. It is not hard to imagine that had Codd being a network engineer working at IBM that did not have to deal with IMS and MVS datasets problems, the relational model would now be predominant in solving traffic problems.
Above this minor point, I find it very frustrating that relational algebra as expressed by D&D is somehow coming to the same pitfalls, it tried once to get around. Crumbling under its own weight of ineffective and redundant semantics and rules, it does not serve the primary purpose of the RM anymore: the usefulness of solving computing problems by setting a more effective conceptual model that avoids the pitfalls of the hierarchical an dnetwork model.
That is a fact, ra theorist will have hard time accepting in subjecting their certainties to scrutiny of math .
This is one of these moments where I wish Codd would be here to answer so many questions.
Regards... Received on Sat Dec 20 2008 - 11:52:25 CET