Date: Fri, 11 Feb 2005 23:02:03 +0000 (GMT) From: Robert Watson <rwatson@FreeBSD.org> To: Ivan Voras <ivoras@fer.hr> Cc: hackers@freebsd.org Subject: Re: How many developers? Message-ID: <Pine.NEB.3.96L.1050210142952.26319k-100000@fledge.watson.org> In-Reply-To: <420B6747.7050807@fer.hr>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 10 Feb 2005, Ivan Voras wrote: > For statistical purposes, where can I get information such as the number > of developers (with commit bit?) active on the FreeBSD project? You might also find this interesting: http://people.FreeBSD.org/~peter/commits.html Assuming that commits were the only activity of interest, this might allow you define a threshold for "active" for various bits of the tree. The usual caveats apply: some committs commit a lot infrequently, other developers aren't committers and submit patches, etc. However, it's probably a useful starting point. I guess you would probably want to combine the results of all of these to get more reasonable total figures. I assume Peter generates them using a cron script of some sort based on the commit logs in CVSROOT-src/commitlogs, so you might look there for the mouth of the horse, so to speak. They're a little hard to parse because the commit message format has changed over time. I run occasional statistics on commit rate, etc, on parts of the tree, and you can find a couple of bits of output from that here: http://www.watson.org/~robert/freebsd/20050211-commits-sys.eps http://www.watson.org/~robert/freebsd/20050211-committers-sys.eps You'll notice that there's a gradual increase in developers over time, and that there are quite noticeable dips in commit rates (and slight dips in the number of committers) around releases, which happen pretty often so it's fairly jaggy. You'll notice that the jags cascade down branches due to the MFC process -- when we freeze for a release in a higher branch, the lower branches slow down, for example. This is limited to the src/sys subtree, so excludes all of user space, docs, ports, etc. The sampling interval is the first through last days of each month, and the "committer count" is the number of unique committers that month. Since some committers commit on cycles that aren't monthly, it isn't perfect, but it's interesting to look at anyway. :-) Robert N M Watson
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1050210142952.26319k-100000>