By Alexander Serebrenik
Developers are people (citation needed) is how Robert Jongeling started his report. In a way, this is a pity: we always talk about the product, rather than about who create software. Based on some heuristics, they developed the “genderComputer” to guess the gender of persons, e.g. on location, names etc. From behavioral economics it turns out that in general women are less effective than men in competitive environments [src]. If you know the gender, one could ask: does diversity improve software? Diversity can be bad, for example because people prefer working with others similar to them, and people within the same group tend to treat themselves better. At the same time, it can be good: think of having different perspectives in problem solving. So, they looked into a plethora of elements and their influence on productivity. It turns out that gender and tenure diversity. both have a positive effect.
Story 1: codes of conduct
Unfortunately, there are still many biases, especially on diversity. Codes of conduct can help to support the welfare of the main stakeholders and rights of all constituents. So, what do major codes of conducts stipulate? They found 5 topics: honourable behavior, unacceptable behavior, enforcement, scope and purpose. And, how are these used? They interviewed several people who developed these codes. It is a balance between rules and values: too many rules and one focuses more on the rules than on the principles underneath it, and the other way around validating on values only is difficult.
Story 2: how do they communicate?
For R, questions on StackOverflow are answered faster than on R-help. Why? For example reputation, gamification and google outperforms on StackOverflow rather than on R-help. How do people react on sentiment? What is the emotion people have in software development?
To summarize: the story of computing is the story of humanity, filed wit drama, hubris, passion, betrayal and serendipity (Grady Booch).