Re: Codd's Information Principle
Date: Fri, 6 Nov 2009 10:54:38 -0800 (PST)
On 6 nov, 01:19, com..._at_hotmail.com wrote: I thank you for your response and the effort you have put into detailing what is an internal predicate or what is an external predicate is to the the relational model but I believe you are entirely missing the point I have raised to paul c. See my comments below
> > I would be glad to hear how we establish a valid quantifier in
> > relational algebra using only internal predicates. The lack of
> > clarification of the external predicate, while being symptomatic
> > limitation of traditional RM relational theorists gladly recognize,
> > does not bother them much when it comes to operate relations
> > algebrically using only the internal predicate.
> Please read Bob's recent postings and especially my posting of Oct 28http://groups.google.ca/group/comp.databases.theory/browse_thread/thr...
Will do. I don't spend as much time in cdt as I used to.
> A query evaluates the extension of (ie tuples that satisfy)
> a predicate expression (the one corresponding to the query
> relation expression) built from external (ie base relation variable)
> An internal predicate is just a necessary but not necessarily
> sufficient constraint on the tuples that can appear in a variable
> (evaluated to avoid (some) erroneous inputs).
> It has nothing to do with querying.
Who mentionned querying?
> (I don't even find the notion of internal predicates helpful.
> It's the overall database constraint that's important.)
That's because domain constraint analysis has been left out of relational algebric definition since it was prior to relational model definition.
> > The concept that a relational *operation* (projection) involving a
> > relation R1 would also serve as a quantifier for the same relation is
> > a concept I am having difficulties with.
> Use of a relation operation in a relation expression corresponds
> to use of a connective or quantifier in the corresponding predicate
That is precisely the concept I feel unconfortable with. In traditional algebra, valid quantifiers are values not operations. I do not see why ra should have the privilege to define its own rules on that perspective.
> Read my referenced message.
> Try an example.
> It's all so straightforward.
Only if you accept as a premice that an operation can also serve as a valid quantifier. There are other functions or intervals quantifiers that can increase the expressive power of algebra but they do require digging into domain analysis and combinatory analysis.
I do not see for instance how can such premice allow to develop a computing model for effectively representing data independence in the context of relation manipulation and operation. I tend to think from current and past research that such premice leads to confusion and limits the expressive ability and opportunity to logically represent relational operations as a part of a turing complete machine.
But of course that is just my opinion.
Regards... Received on Fri Nov 06 2009 - 19:54:38 CET