[racket] Does Redex have support for multi-hole evaluation contexts?

From: Ryan Newton (rrnewton at gmail.com)
Date: Sun Jul 8 19:56:22 EDT 2012

Ah, thanks again!  This is for a paper submission (and we'll post the draft
soon) so there isn't time to change our current model too much at the
moment, but I'm very interested in exploring this parameterized-mark
proposal after that.

For one thing I hope we'll then get to dig a bit more into redex
performance issues -- i.e. compare this version vs. our define-judgement
formulation.  We would like to maxmize the number of paths/sec we can

I think you could similarly parameterize the `mark' by an oracle that tells
> you which redexes to select.  Quick = all, full = any.  You could randomly
> test the CR property with a random oracle selecting two sets then doing all
> subsequent reductions looking for a match.


P.S. Oops I misspoke before -- not that it's important but our current
model isn't multistep like Takahashi.  Rather it's still single step but
allows multiple individual reductions to happen simultaneously on disjoint
parts of the expression.  That makes it more directly comparable to the
proposal by David.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20120708/84b52ad3/attachment.html>

Posted on the users mailing list.