I wrote: > The more I hear of your antics to avoid roundoff errors, the > more I feel you are on a doomed quest for Maxwell's Demon or such :-)
This criticism doesn't necessarily apply to the check sum (described in "The Math" page). Since the chance of one roundoff error in an iteration must be v.small, the chance of two compensating errors is negligible. So the checksum (if economical enough to be performed on every iteration) would effectively reduce the error rate to zero. _________________________________________________________________ Be one of the first to try Windows Live Mail. http://ideas.live.com/programpage.aspx?versionId=5d21c51a-b161-4314-9b0e-4911fb2b2e6d _______________________________________________ Prime mailing list [email protected] http://hogranch.com/mailman/listinfo/prime
