# RE: Re: Bug in stats? - commit-treshold is doubled compared to the actual commits

From: Carsten Johnsen <cjo_at_kirktelecom.com>
Date: 2006-02-15 09:43:23 CET

>> What I did:
>> Analyzed commits: 1959.
>> User1 has committed 377 revisions of those = 19.24% Now setting
commit
>> threshold to:
>> 19%: User1 not moved to "others" group as expected.
>> 20%: User1 not moved to "others" groud - this is NOT expected (or is
>> it?)
>> 38%: User1 not moved to "others" group.
>> 39%: User1 moved to "others" group.
>
> If we would move User1 to the "others" group at 20%, then it would
maybe
> be mathematically correct. But it wouldn't really be useful:
>
> Imagine a project where user1 does 95% of all commits (it's his
project),
> but the other 5% are done by testers.
>
> You only could change between 0 and 5% so see any changes in the
graph.
> After 5%, you would only see *one* bar/line. That's not interesting at
all.
>
> So: it's done so that even with 100%, you will see at least two bars
> (if there *are* more than one committer). One bar/line isn't
interesting,
> so why waste a whole 50% of the slider for that?

Ahhh - well that makes a lot of sense, but I still think the percentage
on the slider should reflect the commit-percentage that would move the
user to the "others" group. This could be done in different ways.
The easiest way would be to just set the slider to have max at 50% and
then use 0.5% steps.
Another way could be to set the max of the slider to whatever percentage
the second most active author is at. - But I guess this could be a
difficult calculation to make - esspecially calculating the step size -
and in some cases make for a wierd slider.

Anyway - these are just my thougths and certainly not something that
will haunt me in the future.

Keep up the good work Stefan.

Regards,
Carsten

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tortoisesvn.tigris.org
Received on Wed Feb 15 09:42:49 2006

This is an archived mail posted to the TortoiseSVN Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.