Re: Days and hours
Re: Days and hours
- Subject: Re: Days and hours
- From: has <email@hidden>
- Date: Sat, 22 Oct 2005 21:29:49 +0100
deivy petrescu wrote:
>>>Yes, that's cool. However you get a coarser granularity than with chrono.
>>>OM(BN)M, Automator says 0.002, while chrono says 0.0002.
>>
>>Which hardly matters if the margin of error is significantly larger than either measurement.
>
>I am sorry, but your analysis is wrong. The significancy of your measurement is determined by the instrument you use to measure.
Nope, the instrument in turn is only as accurate as the amount of background noise affecting it relative to the duration of the test.[1][2] The shorter the duration, the less accurate the measurement. You might have a super-accurate timer that measures to the bazillionth place, but there's no guarantee that the garbage collector won't kick in or other processes suddenly grab more (or less) processor cycles during the test and throw its result completely for a loop. Run a short-duration algorithm multiple times and [hopefully] such unpredictable variations are averaged out till they're no longer significant. Run it only once, and there's every chance of that particular measurement being wildly out due to factors outside its control. Just because computers are supposed to be deterministic doesn't mean they're not out to get you. :p
has
[1] Unless of course the timer's resolution is even lower (i.e. crap), which would be true for something like 'current time' but for the OPs' examples seems unlikely.
[2] ±0.1 sec was suggested only a worst-case example to illustrate the point, btw. I'd expect it to be quite a bit better than that in practice, but you'd really have to run a pile of tests to determine the "exact" level of precision you can expect.
--
http://freespace.virgin.net/hamish.sanderson/
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Applescript-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden