[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
RE: [eclipse-dev] Bug Tossing Graphs
|
Sung Kim,
could you please specify and alternate public return channel for
follow-up questions and discussions. Not all thousands of mailing
list subscribers will be interested in this (and it's a moderated
mailing list), so you should have a channel where interested
parties can "opt-in".
This could be a wiki page of yours, a different newsgroup or mailing
list, or a bugzilla bug.
The number 1 item I'm missing from your "research" is exact data
such as the time window from which you took the bug tossing events.
>From a superficial look, some of the developers you have on the
charts have retired, and many of your tossing events might actually
be those retired developers giving up on all of their bugs - you
might, for instance, take a closer look on whether many of the
bug transitions from "sluiman" and "douglas.pollock" happened to
be on the same day, which would indicate a single day getting rid
of all their assigned ones.
Cheers,
--
Martin Oberhuber, Senior Member of Technical Staff, Wind River
Target Management Project Lead, DSDP PMC Member
http://www.eclipse.org/dsdp/tm
> -----Original Message-----
> From: eclipse-dev-bounces@xxxxxxxxxxx
> [mailto:eclipse-dev-bounces@xxxxxxxxxxx] On Behalf Of Sung Kim
> Sent: Sonntag, 01. März 2009 10:54
> To: eclipse-dev@xxxxxxxxxxx
> Cc: Gaeul Jeong; Tom Zimmermann
> Subject: [eclipse-dev] Bug Tossing Graphs
>
> Dear Eclipse developers,
>
> This is Sung Kim, a software engineering researcher at Hong Kong
> University of Science and Technology. Recently we analyzed bug
> assignment process by mining Eclipse bug reports. We found that many
> bug tossing events, bug reports are assigned developers and the
> developers toss the bug reports to other developers. Based on the bug
> tossing information, we develop a Markov model which captures bug
> tossing history to help bug assignment process. Please find the
> attached sample bug tossing graph. (Perhaps you need to zoom-in to see
> the details.)
>
> Each node in the graph indicates a developer and edges indicate
> tossing and fixing relationship. For example, among all bugs assigned
> to Mddumn, 34% of them are tossed to and fixed by Jerome.Bozier (if
> they are not fixed by Mddumn). The size of nodes indicates the sum of
> incoming edges. The bigger nodes get more bugs from other developers.
> For example, Darin Wright gets many bugs from others.
>
> We believe this graph can help (1) discover developer structures, and
> (2) bug assignment process. For example, the clusters in the graph
> show developer groups. This graph might be used as a cheat sheet to
> assign or toss bugs to other developers. For example, if you are
> considering tossing a bug to Chanskw, you might want to CC the bug to
> Darin Wright as well.
>
> We'd like to get some feedback from you, since you are actively
> assigning, tossing, and fixing bugs. Please let us know what you think
> of this graph. What do you like/hate about this graph? Do you think it
> might be useful? Are there any other possible usages of this graph?
> Any comments are welcome. Based on your comments we will adjust our
> research direction.
>
> Best Regards,
> Sung Kim
>
