From: <HIDDEN>
Newsgroups: comp.soft-sys.matlab
Subject: Re: How to use Looping in a numerical computation?
Date: Sun, 19 Sep 2010 04:15:05 +0000 (UTC)
Organization: The MathWorks, Inc.
Lines: 6
Message-ID: <i742o9$mb7$>
References: <i73mrj$4j7$> <i73ums$br4$> <i73vu7$rln$>
Reply-To: <HIDDEN>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
X-Trace: 1284869705 22887 (19 Sep 2010 04:15:05 GMT)
NNTP-Posting-Date: Sun, 19 Sep 2010 04:15:05 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 1187260
Xref: comp.soft-sys.matlab:671202

"Ryan " <> wrote in message <i73vu7$rln$>...
> You're right to a degree. While the "theoretical" result should remain constant at 1.1. The "computational" result is different. 
- - - - - - - -
  Any deviations from a constant 1.1 are due solely to round-off errors.  The computer cannot represent 10.1 and 10.01 exactly.  At each step in an iteration, past round-off errors become multiplied by a factor of 10.1, so after 20 iterations these errors are magnified by something like 10^20.  In my opinion such results are totally meaningless.  It just means your iteration algorithm is highly unstable with respect to round-off errors.

Roger Stafford