We’re working on some awesome new visualizations and statistics here at GitHub. While they’re not quite ready to roll out, there is one chart I wanted to share.
These are the top 10 languages stored on GitHub, based on some simple heuristics. Forks are not counted – only unique code.
wonder if all that javascript is from duplicate libraries in application trees
or all the php from that you have to write so much of it to do anything
@sprsquish We ignore any path with these terms in them:
vendor jquery bundle cache prototype.js effects.js controls.js dragdrop.js@defunkt can you say in a little more detail how you came up with these numbers? Are you counting lines of code or just files? Mattly brings up a good point about the succinctness of different languages.
It’s lines of code per language on the site.
Keep in mind that Ruby, too, is grossly inflated due to all the unit tests and specs :)
It would be interesting to see how many projects/repositories use which language in any way, regardless of number of files or loc.
@defunkt: Take Edward Tufte’s advice, don’t use pie charts for visualizing information. A bar chart is vastly superior when comparing values against each other.
you shouldn’t count js lines that look like this :)
}Clarification: We count file size, not lines of code.
@uggedal Seems to get across the point just fine to me.
@nnc Agreed! Coming soon.
Both file size and lines of code are unfair to terse languages. That said, I don’t know how else to measure. ;)
WRT pie charts vs bar charts, notice that PHP, Ruby, and C have pretty similar slices. Hard to tell by eyeballing whether they are the same or different. With a bar chart slight differences, if there are any, are immediately apparent.
I’m waiting for the new extended search with some kind of tags for language and status( alpha, beta, release, murdered etc ). For example, it’s quite hard to search projects written in Io, because of language name.