Re: What databases have taught me

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Wed, 28 Jun 2006 19:05:22 GMT
Message-ID: <SpAog.3387$pu3.82117_at_ursa-nb00s0.nbnet.nb.ca>


erk wrote:

> Bob Badour wrote:
>

>>I don't have time to read this paper right now. Recalling the last paper
>>you cited referencing Alloy, I didn't see anything in it that I thought
>>competed with predicate calculus. Are you suggesting it has some
>>advantage for constraints over a WFF ?

>
>
> Alloy uses relations as its sole data structure, with predicate WFFs to
> describe constraints (facts), define the effects of operations (really
> just a use of constraints with pre- and post- states), and to test
> whether certain predicates hold given the model (e.g. a target
> predicate for which to find a counterexample).
>
> What I like most about Alloy is its use of relations to acquire all of
> the desirable properties often touted by O-O advocates: subtyping,
> polymorphism, etc. It doesn't depart from predicates, set theory, etc.
> - in fact, its treatment of atoms as singleton sets, and sets as unary
> relations, gives it clean syntax and typing semantics that doesn't
> decompose into unfettered "dynamism" as advocates of "duck typing" and
> such would have it.
>
> If you saw something in Alloy that runs counter to RM (even though
> Alloy is not intended for data management), I'd like know, since I must
> have missed it.

Perhaps it was not Alloy that I was looking at. Or perhaps it was just my unfamiliarity with the syntax. Received on Wed Jun 28 2006 - 21:05:22 CEST

Original text of this message