Skip to content

Rate of discussion analysis on centos-devel

I was curious how the discussion rate on centos-devel compared to previous time periods. I want to know if our work on growing the project participation at the contributor level is working, and while examples such as the increase in the SIG activity are a good indication, one simple one is to see if there is more discussion in the contributor communication channels. Based on what I see, the trend looks very good.

To make this chart, I simply grabbed all the sizes of the mail archives from the centos-devel archive page and dropped them in to a spreadsheet.

Chart showing discussion rate on centos-devel over 10 year period

Discussion level on centos-devel over last decade.

My analysis is pretty simple.

  • There have been other periods of time where the discussion level was this high (above 50 KB in archive size), but they appear more to be spikes than sustained discussion, with the exception of November 2010 through July 2011. If you were around the Project during that time, you know that is a reflection of work and associated noise around the CentOS 6 release. Although the sustained discussion levels are similar, I think the tone of the discussions is quite different, so I weigh the current trend as “good” by comparison because it reflects growing participation rather than concern about the timing of the CentOS 6 release.
  • 50 KB seems like a good level to judge against in that it seems most months reach at least 25 KB, but going above 50 KB is less common. In the nine years the archives track, the size has gone above 50 KB about 20 times out of 112 archived months, or about 18% of the time. (By comparison, the size has gone above 25 KB 64 times, or 57% of the time.)
  • The one largest-spike-of-all-time is January 2014, which is easily attributable to the announcement about joining forces with Red Hat and the subsequent discussions. Again, the tone of those emails was quite good, as compared to the previously largest spike of February 2011. That spike was to be expected since the news caught many people by surprise, so I’m generally ignoring it as an outlying data point in terms of having any more meaning than that.