Path: utzoo!attcan!uunet!mcsun!unido!fauern!ira.uka.de!sol.ctr.columbia.edu!caen!sdd.hp.com!spool.mu.edu!news.cs.indiana.edu!msi.umn.edu!noc.MR.NET!gacvx2.gac.edu!hhdist
From: NORM%IONAACAD.BITNET@CUNYVM.CUNY.EDU (Norman Walsh)
Newsgroups: comp.sys.handhelds
Subject: RE: HP48: GCF and LCD programs
Message-ID: <5E61BBE10000251E@gacvx2.gac.edu>
Date: 1 Mar 91 20:51:00 GMT
Lines: 46
Return-path: <@CUNYVM.CUNY.EDU:NORM@IONAACAD.BITNET>
In-reply-to: Your message of Fri 1 Mar 1991 13:32 CST
To: handhelds@gac.edu, eagle@catt.ncsu.edu

>I'm having trouble translating a program to RPN.  It's a very useful
>program that I used in my Casio fx-4000P.  It calculates the greatest
>common factor between two numbers.
> (text deleted)
>the GCF program into RPN, here it is in some pseudocode form:
>
>input a;a=abs(a);
>input b;b=abs(b);
>if (a<b) then
>  swap(a,b);
>repeat
>{
>  c=-(int(a/b)*b-a);
>  a=b;
>  b=c;
>} until c=0;
>print b;

Daniel,
Here is my translation of your psuedocode.  It is not particularly
efficient because I have endeavored to duplicate the psuedocode above
as closely as possible.  'a' and 'b' are taken from the stack, 'a' is
returned (not 'b' as the code above would have us believe):

<< 0 -> a b c
   <<
       IF a b <
       THEN a b 'a'
STO 'b' STO
       END
       DO a b / IP b *
a - NEG DUP 'c' STO
b 'a' STO 'b' STO
       UNTIL c 0 ==
       END a
   >>
>>

I have taken the slight liberty of optimizing the storage of b into
a and c into b by duping the result of calculating the new 'c' on the
stack and storing that into b rather than recalling it from 'c' and
storing it in 'b'.

If there is any part of the above code you would like explained,
just let me know.  'though I'll be away next week so be patient...

                                                    ndw