BFAV statistics

Marcus Denker denker at iam.unibe.ch
Sun Oct 10 10:25:13 UTC 2004


Hi Lex,

Yes, grouping the stuff in a sensible way is extremely important. This 
is one
of the great features of Mantis, and I already moved some posts (e.g. 
Bug
reports and Fixes/Enh for external packages) there. This already helped
a lot.

Another categorization, especially for [ENH], would be a "needs more 
work".
I'd guess that at least a third of all posting would be of that 
category.

     Marcus


Am 10.10.2004 um 01:17 schrieb <lex at cc.gatech.edu>:

> Marcus Denker <denker at iam.unibe.ch> wrote:
>> The question is: What do we do with those postings that require more
>> work?
>>
>> Just holding all those postings in BFAV forever makes no sense.
>>
>
> As food for thought, I notice that with the Debian tools it actually
> does make sense to keep them around.  The ancient bugs don't seem to 
> get
> in the way.  The prevent people from endlessly re-filing the same
> complaint.  The attached discussion thread provides a pointer to anyone
> who wants to work on the problem in the future, and it provides
> workarounds for people to use in the meantime.  In short, it's useful
> information, and it is nice to keep it around if possible.
>
> One of my packages, audiooss, has a bug against it that has been there
> for well over a year.  I want it to stay there, and somehow, it does 
> not
> get in my way nor other people's way when we use the bug tracking
> system.
>
> I don't know fully how those tools make such ancient bugs stay out of
> the way, while ours in Squeak do  not, but I think one part of it is
> that the bugs are associated with packages.  You would only see this
> particular bug if you are actively trying to look at bugs for package
> audiooss.  This heuristic seems quite effective: if you aren't
> interested in audiooss, then you definitely aren't interested in this
> bug.  If you are interested in audiooss, then there are few enough
> audiooss bugs floating around that theres' a good chance this is the 
> one
> you are looking for.  To be fair, this heuristic does seem to break 
> down
> a little for large, popular packages with many dozens of bug reports.
> The mozilla package, for example, tends to get gazillions of repeated
> bug reports onthe same underlying bug, probably because people can't
> bring themselves to read through the entire list of existing mozilla
> bugs before posting a new on.  (Not to pick on mozilla -- huge, popular
> programs inherently have more potential for bugs.)
>
> Lex
>




More information about the Squeak-dev mailing list