On 9/7/07, Raphael Collet <[EMAIL PROTECTED]> wrote:
> Dear Enrico,
>
> It looks to me that message passing is the paradigm of choice, since
> processing one lookup might involve non-trivial state updates.  Message

I agree. But I'll try to circumscribe the stateful part to the store only

> By the way, I don't really understand the second requirement you stated.
>   What are the feedback and the equilibrium?

Another name for the object could be nondeterministic feature-feature
dictionary, because when a feature retrieves a value, the value itself
is fed into the object as a  feature. I'm still doing research on the
function that will stop feeding features; for example it could stop
when a cyclic situation is found: a->b + b->a, or a->b + b->c + c->a.

Enrico

>
> Cheers,
> raph
>
> Enrico Santoemma wrote:
> > Hi,
> >
> > the question is about which paradigm is better in implementing this
> > object: declarative concurrency or message passing concurrency, or
> > perhaps some other way.
> >
> > These are some requirements:
> > - should map keys to values
> > - a key lookup should find values which are then fed into the
> > dictionary _ a kind of feedback  _ until an equilibrium is reached
> > - a key lookup also modifies the object's internal store
> >
> > I guess message passing is needed, because this design has intrinsic
> > non-determinism and race conditions are welcome. I wonder if there are
> > other paradigms which are better, before start coding.
> >
> > Thanks,
> > Enrico
> > _________________________________________________________________________________
> > mozart-users mailing list                               
> > [email protected]
> > http://www.mozart-oz.org/mailman/listinfo/mozart-users
>
> _________________________________________________________________________________
> mozart-users mailing list                               
> [email protected]
> http://www.mozart-oz.org/mailman/listinfo/mozart-users
>
_________________________________________________________________________________
mozart-users mailing list                               
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users

Reply via email to