Hello Ross: So good to hear from you again. I should have known that you would be the one to come up with mathematical definitions of "precise" and "accurate". It is obvious that you have thought about this before, in connection with your own sundials. Yes, your definitions and explanations make a lot of sense, as I still remember a little bit from college statistics. Thought it very instructive to see how your definition could be applied to anything that is measurable, even darts! Seems that the dictionary closely agrees with the mathematical definitions, in much simpler terms, of course.
Thank you once again for another one of your thoughtful, in-depth answers. John Carmichael new website: http:/www.azstarnet.com/~pappas p.s. After seeing your amazing Disney sundial, I can't wait to see any new projects that you have in the works. Keep us informed, I'm sure everybody in the group would be interested. >>From my recollection of the principles of experimentation, >accuracy has to do with the magnitude of the spread of a series >of repeated measurements of the same quantity around the mean or >average measured value, representing something like the >reproducibility of the measurement. The greater the accuracy the >narrower the spread. > >Precision has to do with how far the mean measured value is from >the "correct" value. Of course there is a logical difficulty >here. If you knew the "correct" or "true" value, why make any >measurements? > >In the context of sundials, however, where presumably one can >calculate or otherwise determine the correct time with great >accuracy and precision, a sundial's indication of time can have >both a random spread of values around the mean of a series of >measurements of the same thing, and a fixed, error of that mean >from the independently know true value. > >This also brings up a paradox. How do you perform a number of >repeated measurements of 1:23:00 PM? If you cannot do this >(because time is always changing on you), then perhaps we could >compare our measurements only with the moving correct time, >independently determined. In this case I suppose we could speak >of the precision of each measurement being its departure from the >correct time. If we repeat this measurement often over a period >of time and discover that the time differences do not add to >zero, then the non-zero amount would be an indication of the lack >of precision of the measuremenet. The standard deviation of the >time differences could be an indication of the accuracy. > >Does this make sense? > >-- >Ross McCluney, Ph.D. Principal Research Scientist >Florida Solar Energy Center, 1679 Clearlake Rd., Cocoa, FL >32922-5703 >Voice: 407-638-1414 Fax: 407-638-1439 e-mail: >[EMAIL PROTECTED] >Florida Solar Energy Center: http://www.fsec.ucf.edu >Sundials: http://www.sunpath-designs.com >Introduction to Radiometry and Photometry: >http://www.artech-house.com >-------------------------------------------------------------- > >
