On Tue, 26 Feb 2008 14:46:38 -0800, Ask Bjørn Hansen
<[EMAIL PROTECTED]> wrote:

>No, the code should be converting it to an absolute offset.
>
>"-2 * $offset_abs + 1" is, currently, the calculation that kicks in  
>when the offset is more than 0.075 seconds.

Hmm..  All I can offer in rejoinder is some data from the .csv file:


>1204056117,"2008-02-26 20:01:57",-0.149893641471863,0.700212717056274,10.1
>1204054368,"2008-02-26 19:32:48",-0.150670409202576,0.698659181594849,9.9
>1204052746,"2008-02-26 19:05:46",-0.151970148086548,0.696059703826904,9.7
>1204051246,"2008-02-26 18:40:46",-0.146223902702332,0.707552194595337,9.5
>1204049500,"2008-02-26 18:11:40",-0.154271364212036,0.691457271575928,9.2
>1204047755,"2008-02-26 17:42:35",0.0667705535888672,1,9
>1204046107,"2008-02-26 17:15:07",0.316627502441406,0.366744995117188,8.4
>1204044368,"2008-02-26 16:46:08",0.321823477745056,0.356353044509888,8.5
>1204042533,"2008-02-26 16:15:33",0.314275026321411,0.371449947357178,8.6


Absolute value of offset is > .1, yet score goes up.

>                                             Yes, maybe it should be  
>more aggressive.

That seems like a good threshold to me, but there are signs that the
code doesn't quite do what it should.

_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to