On Wednesday, October 07, 1998 3:44 PM, John Payson
[SMTP:supercat@circad.com] wrote:
> From my reading, he reads sixteen input samples for each output sample
> he produces, rewriting the entire array.  For some applications this is
> fine (esp. if you can read the sensors much faster than you need to show
> the data).  For other applications, the data 'delay' would be too much.

I had not thought of that, but it makes sense if the app can handle the
over-sampling rate.

> BTW, I was pondering... how well would it work if one were to read one
> item at a time, and alternately overwrite the minimum and maximum values
> in the array?  It would seem that wouldn't be as good as doing a true
> FIFO, but might be simpler.

This is an interesting idea!  Sounds like it might be OK.  Probably worth
modeling before actually using it (or in my incredible amount of spare time
<G>).

Bob.