4 ms·

I skimmed through this, but their conclusions are far too simplistic to be of any use. A company repo is more popular than a personal one? Are they implying for

by biot 10y ago

I skimmed through this, but their conclusions are far too simplistic to be of any use. A company repo is more popular than a personal one? Are they implying forming a company around an unpopular repo will make it popular? It's likely that if there's a company behind it then it's being done as paid work, that it has specifications needed to interop with something else, with management behind it driving quality, perhaps to a deadline. Second, saying more contributors equates to greater success is a tautology. Or are they suggesting one can simply give strangers commit access and that alone will determine success? More likely, success and popularity attract contributors. And repos tend to get more stars after a release, so if one releases every hour they'll get the most stars, right? Fix a typo, it's version 79.0. Document a method, it's version 80.0. Or maybe it's announcements around a release that serve as marketing? Lots of correlations in their conclusion, but no causation. The paper could be shortened to "Have more visible activity in your repo".

mlinksva 10y ago
> Second, saying more contributors equates to greater success is a tautology. Or are they suggesting one can simply give strangers commit access and that alone will determine success? More likely, success and popularity attract contributors. They don't say this. They note that the measure of popularity they're using (stargazer count) is weakly correlated with commits and contributors. > And repos tend to get more stars after a release, so if one releases every hour they'll get the most stars, right? They certainly don't draw that conclusion.
- biot 10y ago
  Their conclusion states "... confirming the importance of a large base of contributors to the success of open source software". Also note that my questions are facetious and are my own. My apologies if that wasn't clear. Overall, it's a good study of properties that popular repositories have. In terms of factors that impact the popularity, I found it lacking.

by biot 10y ago

mlinksva 10y ago

> Second, saying more contributors equates to greater success is a tautology. Or are they suggesting one can simply give strangers commit access and that alone will determine success? More likely, success and popularity attract contributors. They don't say this. They note that the measure of popularity they're using (stargazer count) is weakly correlated with commits and contributors. > And repos tend to get more stars after a release, so if one releases every hour they'll get the most stars, right? They certainly don't draw that conclusion.

biot 10y ago
Their conclusion states "... confirming the importance of a large base of contributors to the success of open source software". Also note that my questions are facetious and are my own. My apologies if that wasn't clear. Overall, it's a good study of properties that popular repositories have. In terms of factors that impact the popularity, I found it lacking.