Newsgroups: comp.lang.apl
Path: watmath!watserv2.uwaterloo.ca!mach1!torn!howland.reston.ans.net!usc!wupost!csus.edu!sfsuvax1.sfsu.edu!vpcsc4
From: vpcsc4@sfsuvax1.sfsu.edu (Emmett McLean)
Subject: Memory faults, translators, and amps
Message-ID: <1993May28.014844.19016@csus.edu>
Sender: news@csus.edu
Organization: San Francisco State University
Date: Fri, 28 May 1993 01:48:44 GMT
Lines: 37



1. Suppose one would like to create an A struct z
   corresponding to z=i. 100 100 .

   He might use the command :

   z=iota(v2(100,100));

   Or he might allocate the memory for z first as in

   z=jma(INT,(100*100),2);
   z=iota(v2(100,100));

   The method using only iota seems cleaner, but is there any
   possibility one could get a memory fault if they
   didn't use jma or GA?

2. At first glance it appears to me (at least) that it should be easy to
   write a J to C translator which translated only tacitly defined
   programs, and that something which translated explicit definition
   would have to be much more sophisticated. Could one use yacc
   and lex, say, to create a translator, say for starters, which only
   translated tacitly defined verbs? Or would some other set of tools
   be better?
   
   
3. I notice that little-smalltalk uses a struct for it's objects
   similar to J's A struct.  However, l-s implements arbitrary
   precision integers while J does not.  Is there a scheme or
   a reference which explains how amps could be implemented?
   (or couldn't be?)


Thanks,

Emmett
