Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!ut-emx!walt.cc.utexas.edu!lihan
From: lihan@walt.cc.utexas.edu (Bruce Bostwick)
Newsgroups: comp.lang.postscript
Subject: Re: PostScript Language
Message-ID: <25084@ut-emx.UUCP>
Date: 26 Feb 90 01:07:05 GMT
References: <9447@imagen.UUCP> <38910@apple.Apple.COM> <147@heaven.woodside.ca.us> <18027@rpp386.cactus.org>
Sender: news@ut-emx.UUCP
Reply-To: lihan@walt.cc.utexas.edu (Bruce Bostwick)
Organization: UT-Austin, Texas
Lines: 25

I must disagree about 'thinking backwards.'  I grew up programming in
algebraic-notation languages too, and originally had the same problem with
FORTH, PostScript, and my HP-41CX.

Then I discovered that RPN is actually SIMPLER than algebraic notation.
Sounds weird, but think about it:  _EVERY_ operation is simply a matter of
doing something to the numbers on the stack.  They can all be precisely
defined very easily, which makes proving your code much easier.  Variables
are a necessary evil which introduce more complication into your definitions.

It all depends, I guess, upon one's point of view.  I prefer RPN now, mainly
because I use it a lot and find algorithms easier to formulate in it.  You
may not -- but remember one is not necessarily 'forward' or 'backward' ...
or even 'right' or 'wrong' as some say it.

:-)


===============================================================================
Internet:  lihan@vondrake.cc.utexas.edu   :-) aka BB/CIV (-:
BellNet:   (512)346-2744
SlowNet:   varies chaotically, e-mail for current value

nothing cute to say or disclaim -- sorry...
===============================================================================