Newsgroups: comp.lang.apl
Path: watmath!ljdickey
From: ljdickey@math.uwaterloo.ca (L.J. Dickey)
Subject: Re: apl*plus to IBM/apl2 conversion problem
Message-ID: <BuzCvD.65r@math.uwaterloo.ca>
Organization: University of Waterloo
References: <5835@blue.cis.pitt.edu.UUCP> <1992Sep17.163252.6015@unibi.uni-bielefeld.de>
Date: Tue, 22 Sep 1992 12:47:37 GMT

In article <1992Sep17.163252.6015@unibi.uni-bielefeld.de>
        (bstrohm@erasmus.hrz.uni-bielefeld.de) writes:

>I+m sure everbody who has ever tried to convert from IBM-APL to APL*PLUS or
>from APL*PLUS to IBM-APL will agree on the following points.
>
>1. All functions, who perform file-manipulating actions must be rewritten.
>2. All functions, who perform graphics must be rewritten.
>3. Most of the APL*PLUS QUAD-functions and -system-variables must be  
>   rewritten.
>4. Conversion performance is unbearable slow.


So, should one do it by hand?
This is not possible for some of us.  

Any person who works in an environment with multiple
platforms and/or interpreters soon learns to write their code
with cover functions for the things you mention in 1, 2, and
3 and for other system dependent features. 

I don't know any solution to 4.  The most constructive thing
I can say is that just about any solution is better than doing
it by hand.  Do you have any good speed-ups?


-- 
Prof. Leroy J. Dickey, Faculty of Mathematics, U of Waterloo, Canada  N2L 3G1
   Internet:      ljdickey@math.UWaterloo.ca
                  ljdickey@math.waterloo.edu
   UUCP:          ljdickey@watmath.UUCP
