Re: [lisp-59] Event details changed: The New York Functional Programmers November Meetup

On 2007-11-16T14:56:20-0500, Anton van Straaten wrote:
> A point about the presentation itself: I know there'll be a number of 
> people attending who've read the paper being presented.  If anyone has any 
> particular points they'd like to see discussed, please let me know, either 
> by posting to this list or emailing me directly.  If things work out, we 
> may try to incorporate a bit of real discussion into the presentation, as 
> opposed to just a Q&A session at the end.

I am especially interested in the discussion of memoisation in Section
5.4 in the paper.  The more I work on embedded DSLs of various sorts
the more I see this problem, and I agree with the authors that it is
unsatisfactory to make eval a memo function.  I've been having good
experiences to make eval an *identity* function, so all the computation
is done (or at least declared) in what are ostensibly the combinators
(that is, constructor functions) of contracts (or whatever data type
there is in the DSL).  I would love to hear others' thoughts and
experiences.

-- 
Edit this signature at http://www.digita...
The old name for the Giraffe was Cameleopard, literally a Camel-Leopard.