GitHub classifies LOTS of projects as JavaScript when they are actually more correctly categorized as C#, Ruby, or another language. My guess is that this occurs because smaller web projects that include a number of JavaScript libraries end up having more JS than their other language. (even though the librares are just being used and not the focus of the project)
I always wonder why nobody brings up this point when github stats are discussed. Most of my projects show nearly 50% js. Typically I don't need to write all that much js, but all the includes add up.
It'd be like having to include the source of the linux kernel in my project b/c it's a dependency and then saying that most github projects are mostly written in C.
I've long kept an eye on github as a potential source for http://langpop.com stats, but it still seems way too tilted towards languages that, by all other measures, are not that popular.
(And that's no knock on Ruby - lanpgop.com is built with Ruby!)
The languages we use here is for the code in GitHub. So it's a lot more robust than simply counting the dominating language of a repository. For example if there were only two repositories on GitHub, both were classified "Ruby" but they were 51% ruby / 49% js — we'd report ruby as 51%, while his method would report 100%.
You're correct - there are a few popular OSS .NET libraries on GitHub (RestSharp immediately comes to mind) but the vast majority of .NET projects live on CodePlex.
And since almost all C# devs are on windows, mercurial is the source control of choice as opposed to git as hg "feels" better on windows that git. So no github.
