Learning from 10 yrs of Bugzilla data
January 18, 2010 § 8 Comments
Diederik van Liere is a postdoc at Toronto’s Rotman School of Business. His passion: finding out whether open source communities actually make software better, faster. With this in mind he’s taken an in depth look at 10 years of bugzilla.mozilla.org data to look for bug fixing patterns.
Here is a (very amateurish) video of a talk Diederik gave at the Mozilla Toronto office last Friday:
You can also view this in open video (ogg theora) and download the full slide deck (more slides than you see in the video) as a PDF.
Following the presentation on Friday, the Moz-TO crew gave Diederik all kinds of constructive feedback on how to make this research better — mostly around different ways to run the data. If you have similar suggestions, please post comments and I will pass them on.
Hey Mark, this was definitely really cool to see this information. There was something rather shocking about it, though–the strange coincidence that all the community numbers start to decrease right at the start of 2005.
The reason that this is so funny is that I found the exact same problem when I was researching the history of the Bugzilla community several weeks ago–that it peaked in 2005 and went down afterward. My methodology was totally different too, where I was counting code contributors to CVS over a certain period of time.
Yes, it’s true that we all had major releases around that time, but does an improved product, that’s being used by more people, really continuously correlate to fewer contributors over a really long period? That doesn’t make sense to me–it seems more like a hypothesis than something founded on actual evidence, and a hypothesis that doesn’t help us improve the community, at that.
I’d be interested to know if there was some significant change in the Mozilla community near the start of 2005–was there some person who no longer worked on community-building? Was there some policy change? It’s an interesting question, one that I’m not equipped to research, since I don’t have the data or know where to go to find out these things about the broader Mozilla community.
[…] people have recently linked to Diederik van Liere’s talk “Learning from 10 yrs of Bugzilla data“. If you’re interested in that kind of thing, there’s a whole sub-field in […]
Can I have your permission to put this video on YouTube?
Reynold: why do you want to do that? It’s available here already as open video and as non-open video.
I was hoping to watch this on my iPhone and couldn’t open it. Anyway your call. (I’ve already watched it yesterday.)
[…] Learning from 10 yrs of Bugzilla data « commonspace Diederik van Liere, a postdoctoral researcher, analyzes Bugzilla data (for Firefox and Thunderbird) for patterns concerning community size and bugfixing patterns. (tags: programming video opensource software bugs) […]
Great stuff, Mark. I’m sending this to all the Fedora QA folks.
[…] – bug report and trackingLooking at ten years of Bugzilla data, Diederik van Liere noted a couple interesting findings about that community: Retention of community members is key Get community members through the […]