Always a gamble (and a security risk)— Mastodon|Remi Rampin (@remram44) July 19, 2017
Plus squishier things like whether author seems aware of related packages and literature, author's reputation and other contributions.— James E. Pustejovsky (@jepusto) July 20, 2017
@seankross applies the punch in the face test -- if I could find them to punch them in the face, and I know them well enough, I'll use it— Elissa Redmiles (@eredmil1) July 20, 2017
With packages, more is better – more documentation, more examples, more unit tests, more downloads.
But as important as these things are, people will look at you and your package’s online presence to gauge whether they should use it. We do a similar thing when judging candidates for a position. We make a first impression based on their appearance, their institutional or company affiliation, and the people in their network. Academics cultivate a presence in their field by going to conferences, giving seminars, and editing in journals, so their work is more widely recognized; similarly one needs to cultivate an online presence in the open source community to give their packages credibility. This seems suboptimal.
As a final note, Karthik Ram is working on a paper about this topic, which I’m sure will be more comprehensive than my musings!