I've appreciated the option to just tick "this bug affects me" when I find a bug report that describes the problem I am experiencing and I don't have anything significant to add except "me too" to help show the relative extent of the affect and that I can also reproduce it. I think the machine learning approach to de-duplication sounds very intriguing. Maybe the algorithm could send a report to the designated bug triage team listing all the matches above a certain confidence threshold for their consideration. Getting a hint at which bugs are likely duplicates could save time for the triage team as well as the developers: either the duplication is real and a bug is closed quickly by marking it so, or developer resources of sufficient understanding are concentrated on the thorniest issues. I think the supervised learning is important to help it adapt to different subject areas--X, kernel, sound, filesystem, E-mail client, etc.--although there may be some predictors that cross the boundaries. The idea of "Bug Q&A" sounds very promising, from the standpoint of a bug reporter I like the idea of knowing what questions to answer for the triage team and developers, from the standpoint of triage it would be nice to be able to search for all the "New" bug reports that include the needed information in particular standard fields, from the standpoint of the developer it would be nice to have a set of standard fields to look at for the basic information rather than reading through a sometimes very long discussion in the bug report. So +1 for the current features supporting crowd-sourcing, +1 for an automated de-duplication probable match report, and +1 for the "Bug Q&A". Interested, Richard
This is the launchpad-users mailing list archive — see also the general help for Launchpad.net mailing lists.
(Formatted by MHonArc.)