LWN: Comments on "Who made Gentoo Linux, and when? A commit analysis"

Please use "long tail" correctly!

dthurston — Sun, 14 Oct 2007 17:31:43 +0000

You have the notion of "long tail" exactly backwards. A long tail in probability means that unlikely events have a large total probability; in your context, a "long tail" would mean that the developers with very few commits nevertheless contributed a substantial portion of the project. You haven't given us enough data to see if it happens, but it's certainly not the term you want to use.

Who made Gentoo Linux, and when? A commit analysis

eris23 — Sat, 13 Oct 2007 18:05:00 +0000

There's also the issue of Gentoo metadistributions. I, for instance, use SabayonLinux. Others can be found http://en.wikipedia.org/wiki/List_of_Linux_distributions#.... Plus the various Gentoo overlays. To check the health of Gentoo one should check the commits to all these.

Who made Gentoo Linux, and when? A commit analysis

dberkholz — Thu, 11 Oct 2007 23:00:47 +0000

That's an intriguing idea. Any suggestions for ways to figure out whether this is happening, besides waiting a year?

Who made Gentoo Linux, and when? A commit analysis

dberkholz — Thu, 11 Oct 2007 17:45:10 +0000

There will be a little weirdness on the very last data point, because it's Sept. 1-22 rather than the full month.

Who made Gentoo Linux, and when? A commit analysis

incase — Thu, 11 Oct 2007 15:56:48 +0000

Your comment would make sense if only applied to the developer lifetime charts. But if you take a look at the total number of commits, you also see a noticeable drop in the number of commits. Certainly not as noticeable as the developer lifetime charts seem to indicate, but still it is there.

cu, Sven

Who made Gentoo Linux, and when? A commit analysis

tialaramex — Thu, 11 Oct 2007 14:24:29 +0000

I now attribute this drop to Gentoo's developer population returning toward an equilibrium after an explosive, uncontrolled growth. The Gentoo structure and governance could not scale quickly enough to deal with all the new developers, but it took some time to normalize and continues to do so.

Um, maybe I'm not being very bright, but isn't the most likely explanation for this part of the chart simply that you don't have a list of future commits yet to happen?

The reason I thought this was because of another chart with the same characteristic shape I've seen. The Potaroo automated address exhaustion site includes a chart of IPv4 address allocation vs usage against time, and whenever you look at it, there appears to be a very low ratio of usage for the last few months, which would suggest a disaster in Internet governance. But in reality if you look at the same period on the chart in a year's time the anomaly vanishes, because by then those allocated addresses are in use, it just takes time after they're allocated for that to happen.

I suspect that the same thing is wrong with your chart. Some of the people who haven't committed in July, August or September will pop up in November and thus retrospectively make the chart's "lifetime" indication wrong.

Who made Gentoo Linux, and when? A commit analysis

ekj — Thu, 11 Oct 2007 13:38:53 +0000

Gentoos popularity seems to be following a similar curve, for example the searchvolume on Google for "gentoo" peaks around the same time as your curve for active developers.

My guess is that Gentoo -was- at the time seen as a new and interesting distribution, many new developers arrived. That is no longer the case. Today gentoo is, essentially, a niche-distribution. Ubuntu and family has taken over as the biggest "up and coming".