It would be interesting if they would share stats though, maybe I am wrong.

Quite easy: we used the exact same metrics. Aside from the more reliable metrics we have, we also can comb through our server logs like Fedora does and extract the IP addresses. As a matter of fact, that is what we used to do, see the link to http://en.opensuse.org/openSUSE:Statistics in the article.

So, there is no mapping needed, it is a straight comparison.

Cheers,
Jos

Can you describe how you mapped Fedora’s unique IP metric into your update usage metric or vice versa?

The article sort of brushes over that bit and I’d like a human language description of what the mapping was meant to do from the perspective of the people who did the implementation, before I go in and attempt to validate that the code does what is expected. I mean I know how I would do it, but I want to presuppose my preferred solution is what was used here.

-jef