Re: Codd and many-valued logics

Op zondag 12 juni 2016 10:24:17 UTC+2 schreef com..._at_hotmail.com:
> On Sunday, June 5, 2016 at 4:03:38 AM UTC-7, Nicola wrote:
> > On 2016-06-05 02:44:56 +0000, com..._at_hotmail.com said:
>
> >> On Monday, May 30, 2016 at 6:48:00 AM UTC-7, Nicola wrote:
>
> >>> It seems to me that Codd never referenced the literature on many-valued logics in his papers (I'd be glad to be proven wrong). Do you have any idea why?
>
> >> without logical inference it's a bit much to claim that one has a logic let alone the "full power" of one with it.
>
> > But at least there is potential for adding inferential power, witness Datalog.
>
> Datalog adds inferential power to Codd-style DBMSs. But it is entirely 2VL. It relies on what is not derivable being taken as false.
>
> >> But the little of logic that he used hardly needs references.
>
> > Sure, Codd's treatment is self-contained. But since the role of classical predicate calculus is (rightly) emphasized so much, why not emphasize connections with existing many-valued formalisms?
>
> Because the connections are trivial (my last quoted sentence) and Codd's system is not useful (my next quoted sentence).
>
> > Maybe, because the foundations are not as strong?
>
> Don't understand.
>
> >> There isn't much else that's "logical" about his use of MVLs. I don't know what you would expect to find of use in a reference.
>
> > The "shoulders of giants" upon which everyone builds upon.
>
> I don't understand what you are trying to say, could you clarify? No standing on shoulders was done by Codd re applying higher level logics to the relational model.
>
> philip

As for "It relies on what is not derivable being taken as false." :

But is that enough ? I mean enough for a user to understand how the system is behaving ? I mean "It relies on what is not derivable being taken as false." is exactly what SQL does also. But in order to understand the system's behaviour you also need to know which precise question it was that got its "unknown" answer treated as "false".

Take the question "am I going to accept this update or not". SQL treats "unknown" answers as "false" and accepts updates if the answer to the constraint violations question is unknown. Therefore, the question was "is any constraint violated" and not "are all constraints satisfied".

Under 2VL, these two are equivalent. Under "pure" 3VL they'd be as well, but that doesn't matter because in this particular case, SQL must require the question/the answer to it to be 2VL. Received on Mon Jul 25 2016 - 21:01:35 CEST