Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!clyde!uunet!mcsun!unido!infbs!tubsibr!hafer
From: hafer@tubsibr.uucp (Udo Hafermann)
Newsgroups: comp.lang.apl
Subject: Re: Syntax for extended matrix operations
Keywords: APL primitives, beyond quad-divide
Message-ID: <1989Dec5.123747.24958@tubsibr.uucp>
Date: 5 Dec 89 12:37:47 GMT
References: <539@s5.Morgan.COM> <6458@tank.uchicago.edu> <558@s5.Morgan.COM> <6510@tank.uchicago.edu> <566@s5.Morgan.COM>
Sender: hafer@tubsibr.uucp (Udo Hafermann)
Organization: TU Braunschweig, Informatik (Bueltenweg), Germany
Lines: 10

How about discussing a "standard" way of having user-defined functions
written in a "foreign" language, say, C?  (I myself found C to be a
fairly convenient environment when implementing auxiliary processors
for APL.68000, as a good deal of interface details can be "hidden" in
macros or functions, and as C is rather good at accessing "foreign"
data structures.)

Or do the internal representations of APL arrays vary too much across
implementations to allow even thinking of this kind of thing?