(disclaimer: i'm the author) it's important to keep two things in mind when considering any ranking: first, and most obviously, no language ranking will be perfect for all readers, because the metrics for ranking languages will vary on an individual basis. second, the intent of the ranking. for our purposes at redmonk, this is an important consideration, because we neither intend to nor claim to produce rankings that are representative of language use broadly. if the rankings were representative of all use, languages such as COBOL would have a substantial presence on the list. we are rather interested in communities that we believe to be more predictive in terms of future use, of which github and stackoverflow are two obvious examples. the sustained strength of javascript on both properties has been one example of their ability to identify trending languages.
as for the criticisms regarding the usage of stackoverflow above, this is why we correlate the stackoverflow rankings with github. one represents discussion and research about a language, the other is manifestation of activity within a language. what's interesting is that the correlation between these properties has historically been strong and appears to be getting stronger over time.
again, no ranking is perfect - ours included - but we feel that measuring programming language interest and traction via these properties is at the very least an interesting datapoint.
as for the criticisms regarding the usage of stackoverflow above, this is why we correlate the stackoverflow rankings with github. one represents discussion and research about a language, the other is manifestation of activity within a language. what's interesting is that the correlation between these properties has historically been strong and appears to be getting stronger over time.
again, no ranking is perfect - ours included - but we feel that measuring programming language interest and traction via these properties is at the very least an interesting datapoint.