Alec Flett wrote: > Maybe I have a naive understanding of standard deviation then. My > understanding is that if you repeatedly run the same test, and the > standard deviation is "high" (by some human-chosen measure) then the > variability in the measurements is too high to consider the test > reliable. What samples are you using for standard deviation? I don't see > how a high standard deviation results in something we're interested in.
No, you have the right understanding. And I believe so far the std dev we are seeing is pretty small. But it could change, and if it does, we need to figure out what is causing the jitter and fix it. But my understanding of std dev is also that a change that is less than std dev is most likely noise in the measurement. > nope, pasted from tinderbox right now: Ah, that explains it :) We are discussing my proposal, not what is currently visible on Tbox. Please see my email in this thread on 9/30, especially the attachment. -- Heikki Toivonen
signature.asc
Description: OpenPGP digital signature
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Open Source Applications Foundation "Dev" mailing list http://lists.osafoundation.org/mailman/listinfo/dev
