Nicolas Barcet Free zone of nijaba

I think that was done specifically to mimic the 2009 eclipse survey. Why eclipse would want to lump them together in a survey however is beyond me. Might as well lump all the Debian and Ubuntu derivatives together too if you are going to do that.

Anyways, general problems with accounting for skew in opt-in usage information aside (you'd be amazed at what political pollsters do to for correction factors when doing sample polling).

Two main questions I have about this graph:

1. Does this graph include both the no-cost community edition product and the for-pay Enterprise products? Can the OS breakdown be generated for the community and enterprise products separately?
2. Does this include the pre-built appliance images (both community and enterprise)? The blog specifically mentions the Suse based vm image and the pre-built ec2 AMI (is that an Ubuntu based image?) for GWOS. If those pre-built images are included in the stats how are those accounted for? Are the 8% listed as Suse the prebuilt vm image? Is that 14% listed as Ubuntu a GWOS prebuilt EC2 AMI? Looking at the available literature at the groundworks website the Community Edition vmware vm is based on CentOS. So is the 40% CentOS the community edition virtual machine running under vmware? prebuilt appliance images really complicate how you account for OS usage. I would have liked to see virtual machines broken out to get an idea of how many people are running appliances under vmware as a windows or linux host, versus xen based EC2.

• jef