Gnome DVCS Survey

In December, the Gnome team did a survey of its contributors about which DVCS they might be interested in moving to in 2009. The GNOME DVCS Survey Results are now out and the raw data is available, which some of the community has started analyzing.

GNOME contributors with an SVN account who had an SSH key installed on their account were invited to fill in the survey. A total of 1083 account holders were invited, and 579 filled in the survey.

It had a ranked choice component, where you could choose your first, second, third etc choice for which VCS you preferred to use.

Elijah Newren of the Gnome project put up his Gnome DVCS Survey analysis, which is really interesting. Among the various graphs that he has is this one showing the average weighted rank of each VCS by different demographic groups of which VCSs they knew or used:

Elijah notes:

No real surprise here as far as the favorite goes — users who are familiar with or regularly use a certain system tend to prefer that system. However, git enjoys positive support in all cases and at least comes in second? I found that somewhat surprising. I thought it would get a average ranking lower than 3 by those familiar with or using bzr/hg — much as bzr, svn, and hg did among those familiar with or regularly using git.

What was interesting to me there is that of people who are familiar with or regularly use git, any other system seemed to be equivalent as a second choice – they would as soon stick with SVN as switch to bzr, and would slimly prefer either over using hg. That struck me as really strange – I like Git more than anything, but I certainly like Hg more than SVN. Even stranger to me is the bzr users, who rank bzr their first choice, then often Git a close second and then Hg last, again after staying with SVN.

Shawn McCance also did an interesting analysis of first picks by which systems users were familiar with – this doesn’t take the secondary votes into account, just which system got #1 for each demographic group:

Some interesting data:

• Of everyone who knows bzr and git, they choose git 2:1
• Of everyone who knows hg and git, they choose git nearly 2:1 – (1.8:1)
• Users who know all three choose git or bzr mostly evenly over hg
• Users who actively use git and hg divided their first choice basically evenly between them (though the sample size is pretty small)

Furthermore, even the less technical users seemed to prefer Git over the other systems. The only group that did not was the ‘documenter’s, but in that case the survey size was 4 (out of 579). This is interesting to me because of the nature of this group – it’s not the ruby/rails or linux kernel sort of group – I would have assumed Hg to have gotten a much better showing, but it seems to have done worse than sticking with SVN in most cases.

In any case, it’s interesting that Git is gaining ground pretty significantly in all corners of the open source world and that when developers use Git and compare it to the competition, Git is very often the favorite.