OpenSourceWatershed is a project aimed at understanding the relationship between distributions (downstream) and the individual software components (upstream). It is the basis for a larger study of distributions and their evolution. It is distrology. In the future, more distro oriented statistics will be available. More details are below. For now search in the top right for your favorite package to see how up to date the different distributions. Or look at the right to see what new releases happened in the last 24 hours.

The aggregate analysis is done over twenty packages including firefox, gcc and openssh. The full package list is in the OSCON slides. In the future, users will be able to set custom groups. The three forms of analysis are percent obsolete, the average number of newer releases per package and the time since the oldest new release. In other words the lag is the amount of time a distro had to move to a newer package.

There are errors in the database which you can help fix. Just email me if you find one. In the future, you will be able to fix it yourself. For more information about the process behind this analysis please read my senior thesis or email me.

Rank	Distro	Codename	% Obsolete	Avg # New Rels	Avg Lag
1	fedora	12	40.0%	2.45	10w
2	arch		50.0%	1.65	4w
3	freebsd	8	60.0%	5.93	37w
4	sabayon	5	70.0%	5.95	22w
5	slackware	13.0	82.35%	11.59	34w
6	ubuntu	karmic	85.0%	5.10	22w
7	gentoo		85.0%	7.05	22w
8	opensuse	11.2	85.0%	5.15	27w
9	funtoo		85.0%	7.05	22w
10	debian	lenny	100.0%	26.00	76w

Rank	Distro	Codename	% Obsolete	Avg # New Rels	Avg Lag
1	fedora	13	30.0%	0.55	1w
2	arch		35.0%	1.40	4w
3	ubuntu	lucid	40.0%	1.70	9w
4	gentoo		45.0%	2.35	10w
5	funtoo		45.0%	2.35	10w
6	slackware	current	47.05%	2.24	5w
7	freebsd	9	57.14%	5.07	15w
8	debian	squeeze	65.0%	5.30	13w
9	opensuse	factory	65.0%	1.35	7w
10	sabayon	5	70.0%	6.05	18w

Much more work will happen to OSWatershed in the future. The most urgent addition is distro pages which will feature data on multiple branches. User accounts are also coming in the future. Users will be able to add more data into the database and configure their own package groups to use for analysis. Ideally, spreading the work over a larger number of users (crowd-sourcing) will make the data scope more manageable. To get Scott working on these new feature email him and let him know you are waiting!