On Saturday, 2011-10-01, Joshua Blocher wrote: > I think we are acting like it all has to be done manually which is > simply not true. Why are we tackling bug triage as something that only > a human can do?
Because it potentially requires interpretation of natural language text, understanding of relations between concepts and ideally the ability to combine those to reproduce the problem. > Computers are good a repetitive tasks. Indeed they are. They are also very bad in doing things outside a very strict set of rules and are almost certainly not able to adapt those rules. > A little bit of > intelligent use of technology would reduce the "burden" on all of our > developers. KDE has semantic technology, use it. Or if that isn't good > enough there has to be another route. We can find possible duplicates > by matching debug text ranked by percent of similarity. 50% and below > most likely not a duplicate. 50% - 75% possibly a dup, if 75% and up > probably a dup. Then auto assign it to the correct original bug. How would this decide which of the reports is the main one and thus has the best description of the problem? Or how would it detect additional information present in a duplicate and copy that into the main one? IMHO a software like that should not automatically decide to do things, at most it could be tool for triagers to help them discover candidates. Cheers, Kevin -- Kevin Krammer, KDE developer, xdg-utils developer KDE user support, developer mentoring
signature.asc
Description: This is a digitally signed message part.
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<