A couple of weeks ago, Kaitlin Thaney blogged about a discussion happening in one of Software Carpentry’s GitHub repos of what we should teach scientists about the “last ten yards” of science: writing the paper and getting it published. The web has opened up dozens of new ways to do this, but that proliferation of ideas can be confusing to newcomers — what should we recommend? And why?

Many people have offered a lot of interesting suggestions, and I’ve summarized some in a blog post on the Software Carpentry site. The key insight, perhaps, is that while there’s a wide gulf between what’s possible and what today’s scientific publishing systems can digest, we can and should show scientists a few key practices that can improve their lives right now. We’d welcome further comments on the discussion thread.