Here is a page to keep track of options for creating data smoothing filters.

Here is the first page from Google: http://www.vanguardsw.com/DpHelp4/dph00109.htm

This has some interesting points about the most common smoothing formulas: http://www.numberwatch.co.uk/smoothing_of_data.htm

Here is an email from Bob:
I hear that you are creating a differentiator for CC Graph so we can
graph the velocity. In the past, we have had problems with noise if
all we did was to take first differences of real data. Go ahead a do
that, but if it seems noisy, there are two ways out. I developed a
very fast algorithm for fitting to 5 or 7 points around each point,
fitting them to a quadratic and just taking the coefficient for the
square term. This is fast and stable. Otherwise, you could do a
running average, which is simpler.