The Open Source Census rolls forward, but I'm not sure how far it has gone as yet. In the summary, it shows just 789 machines scanned (as of the time that I read it). That's not a bad start, but it is just a start. As such, it's hard to read much into the data.
To be more representative, it will need to get more responses from those employed by larger companies. With just 22 percent of respondents employed by a company with more than 1,000 people, it's clear that the Census skews toward SMBs (small and midsize businesses, with an emphasis on the "S").
It will also need a more representative geographic spread. For example, France, which always shows up as second or third, in terms of open-source adoption in every open-source survey I've seen, apparently doesn't even scrape 2 percent of participants. The United Kingdom, by contrast, is third, behind Canada, despite its dismal commercial open-source penetration.
So the data appears to be highly imperfect, but it will get better as more participate.
The data on Ubuntu's amazing adoption, however, is nigh impossible to dispute, looking at the data.… Read more