Mark Miller writes:
>(define (accept-product (calc-factory :deep-frozen))
>     (let (diode 0) 
>         (define (diode-write (val :deep-frozen))  (setq diode val))
>         (define (diode-read)                      (diode))
>         (define q    (calc-factory ... diode-write ...))
>         (define bond (calc-factory ... diode-read ...))
>         ...))

I want you to fill in the ellipses so that after translating it
into Haskell, I have something that will compile and run.

My initial impulse was to print out of copy of "Regained" and
study it till I knew enough to fill in the ellipses myself. 

After all, your papers are well worth studying and you're an
important computer scientist.  (Though I think I understand the
reasons for your career choice, I find it quite ironic that you
are a grad student right now.)

But then it occured to me that when the communication between you
and I sticks as much as possible to _formal language_, the
collective cognitive burden of communicating is likely to be much
lower.

I.e., I believe formal language--at least the kinds of formal
languages (Scheme, E, Haskell) we are using here--makes a great
interface between two computer scientists who wish to swap/share
knowledge of their respective subspecialties.

(If only our formal languages were protean enough to encompass
_more_ of the topics we are interested in! I say.)

For my part, when I finish my coding and testing in Haskell, I'll
translate my results back into pseudo-Scheme, "novice"-E or
whatever other formal language you prefer.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to