From: Mark Holm 
To: atm{at}shore.net, imasloth{at}bigpond.com
Reply-To: Mark Holm 


Sixtests screen shots for John's  Foucault test results as of August 4,
2003 (the data with 4 zones and 3 measurement sets) can be seen at:

http://users.telerama.com/~mdholm/atm/murray/murray1.html

I don't claim to be a figuring guru, so I will let others suggest the best 
techniques for improving this figure.

I do know, from hard experience, that it is possible to improve a mirror
like this, but it may not be very straightforward, at least not for the
inexperienced.

You can see from the first plot one of the main reasons I suggest that
testing in more zones is probably a good idea.   The residual errors,
after fitting a parabola, have two maxima and two minima.  This is the
theoretical limit for four data points.  There is no redundency in the data
at all, and, to a certain extent, Sixtests is just guessing how high those
maxima and minima are.  These are good guesses based on the best
mathematics available, but they still have a significant element of
guesswork in them.   A few more data points would add a comforting
degree of certitude to the estimates of the maxima and minima.

Mark Holm
mdholm{at}telerama.com

--- BBBS/NT v4.01 Flag-5