Observations
2354
Comment:
|
2404
asked question
|
Deletions are marked like this. | Additions are marked like this. |
Line 10: | Line 10: |
* how do you search for these? Towsonu2003 |
To address the problems we face it's important to find out, what goes wrong.
Some observations:
Unaddressed
we seem to be gaining ~100 open bug per week (9500 2 weeks ago, 9600 last week, 9727 tonight). We are fighting a losing battle it seems
- over 2/3 of these open bugs are "unconfirmed". I think we should be doing better job at this. A lot of these bugs could probably be easily "confirmed"; often it is obvious from the comments that multiple people are seeing these bugs. Confirming a bug is a good way to show an unexperienced bug reporter that his/her bug report is valued. It's a lot better to do this than having a bug report that goes totally unanswered for months if not ever. So maybe we should put little bit more effort in confirming new and old bugs.
- In Debian's area of LP, there are nearly 3000 open bugs. The large majority of these bugs are actually closed from a long time ago; when bugzilla bugs were imported, the ubuntu task kept its bugzilla.u.c status, but if it was linked to debbugs it got an open debian task. These false-open-debian-upstreams bug show up in the list of subscribed bug reports of Ubuntu developpers. It would be nice to clean these up to remove the number of useless bug reports floating around. But it's a very long, boring, and with not a lot of rewards task.
- how do you search for these? Towsonu2003
regarding the unconfirmed bugs... i saw many unconfirmed bugs in the past too where there already was some discussion and more people having that problem. it seems people have some kind of fear to set a bug on confirmed
- we should encourage people to search for already existing bugs and confirm them instead of filing another duplicate one
- [:Kaleo]
I agree with Jeremie. Out of all the bugs ever reported, 15% were duplicates. Killing duplicates could potentially save us 2000 currently open bugs. Avoiding duplicates should be a priority. A way to do so without requiring extra work from reporters or bug squashers would be an automatic suggestion tool which would suggest similar bugs. I would like to try to develop that kind of tool but is there any way to access malone data without going through the web interface (XML-RPC...) ?
Addressed
Back to [:BugSquad].BRBR [:CategoryBugSquad]
BugSquad/Observations (last edited 2008-08-06 16:19:30 by localhost)