FRANCISCO JAVIER LOPEZ FRAGUAS <fraguas_at_sip.ucm.es> wrote:
> Is it related to the property
> that C[e ? e'] should be equivalent to C[e] ? C[e'],
> for any context C and expressions e, e' ?
>
> I remark that the property is certainly true
> for CRWL programs. An interesting point is that the
> proof of it is easy if one uses semantic-based reasonings.
> Thus, an operational procedure can use the above equivalence
> while preserving the semantics.
Since generators typically introduce alternatives recursively,
this equivalence has to be used with utmost care operationally ---
otherwise, applying this equivalence may just happily replicate redexes in C,
without making progress.
This could be alleviated using second-order sharing for C[_]
in C[e] ? C[e'], similar to optimal reduction in lambda calculi ---
is anybody aware of any work in this direction for FLP?
Wolfram
_______________________________________________
curry mailing list
curry_at_lists.RWTH-Aachen.DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/curry
Received on Do Nov 08 2007 - 09:17:25 CET