Email list hosting service & mailing list manager


Re: Alternative formulations of keywords John Cowan 12 Apr 2006 03:27 UTC

Eli Barzilay scripsit:

> (I'll be using the syntax from our library, no time to read Marc's
> syntax, sorry.)
>
> (define (assoc x l #!key test)
>   ... assoc using `test' to find an item ...)
>
> ;; searches for an x association in foo's table, same keywords as `assoc'
> (define (search x foo . r)
>   (apply assoc x (foo-table foo) r))

Ah.  But this does not require computed keywords as such; it simply
requires that whatever keywords are mapped into is a proper Scheme
object.  Under my #1, the keywords become a list of values and
#!missing's in a fixed order; under my #2, they become an a-list.
In either case, your SEARCH works correctly.

Computed keywords proper require a use case in which, given

	(define (foo x &key test) ...)

the call (foo bar baz zam) works iff the value of baz is the symbol TEST.
This might be a Good Thing, but I'd still like a use case.

--
John Cowan  xxxxxx@ccil.org  http://www.ccil.org/~cowan
Thor Heyerdahl recounts his attempt to prove Rudyard Kipling's theory
that the mongoose first came to India on a raft from Polynesia.
        --blurb for Rikki-Kon-Tiki-Tavi