3 ms·
I've never liked Redmonk's choice of Github only. It's just one OSS host, albeit an important one. Why not use Ohloh's numbers (which they do reference), which
by ternaryoperator 13y ago
I've never liked Redmonk's choice of Github only. It's just one OSS host, albeit an important one. Why not use Ohloh's numbers (which they do reference), which cover a larger swath of the OSS community? By choosing to use GitHub only, projects that have been around a long time are not included--such as many of the active projects on SourceForge, Google Code, etc. Not to mention other currently popular hosts like BitBucket.
- _random_ 13y agoOmission of BitBucket is a big miss. Not sure if both publish stats of private repo languages.
- dsberkholz 13y agoNot for public repos. Scroll on down to the "hosting provider" section of some research I did on forges to get a better idea (http://redmonk.com/dberkholz/2013/04/22/the-size-of-open-source-communities-and-its-impact-upon-activity-licensing-and-hosting/ http://redmonk.com/dberkholz/2013/04/22/the-size-of-open-sou...).
- Laremere 13y agoI think they just need more than two sources since there's more to program language popularity than they can two sources can show. Stack overflow naturally biases towards languages with worse documentation and those which cater to newer programmers. For example, it favors Visual Basic and disfavors Golang, resulting in VB having a rank of 19, with Golang ranking at 22 despite golang being far more popular on github. Factors which this survey doesn't account for include what languages new programmers are being taught in, how much code is actually running in production, what type of developers companies are wanting to hire, and what those companies are writing in behind closed doors.