Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site cadovax.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!whuxlm!harpo!decvax!ittvax!dcdwest!sdcsvax!sdcrdcf!trwrb!trwrba!cadovax!keithd
From: keithd@cadovax.UUCP (Keith Doyle)
Newsgroups: net.math
Subject: Re: multiple-precision arithmetic
Message-ID: <323@cadovax.UUCP>
Date: Wed, 5-Dec-84 20:48:17 EST
Article-I.D.: cadovax.323
Posted: Wed Dec  5 20:48:17 1984
Date-Received: Sat, 8-Dec-84 05:46:31 EST
References: <163@faron.UUCP> <732@reed.UUCP>, <6213@mcvax.UUCP>
Organization: Contel Cado, Torrance, CA
Lines: 24


Speaking of PI and multiple precision arithmetic, after once upon a time
looking at the infinite series for computing PI, it seemed possible to 
write a routine that could run forever, computing digits for PI.  No
need for umpteen digit precision arithmetic, as it can be determined
when a digit will no longer be affected by further computations, thus
allowing the digit to be produced, and the 'remainder' of the PI
computation to be scaled up in such a way as no precision is lost,
and only a reasonable integer precision is necessary for any of the
intermediate results.

I've always wanted to do this, and produce a subroutine that would
return the 'next' digit of PI, but never got around to it.  Has anyone
out there done this?  Thought it would be fun to see how it performs
as a random number generator, etc..


Keith Doyle
{ucbvax,ihnp4,decvax}!trwrb!cadovax!keithd
"You'll PAY to know what you REALLY think!"

P.S.  Who knows, maybe I'll finally get around to working this one out
      myself.  Hardly high priority.