So today I’m sorting out the list of users according to Bradford’s Law of distribution, which mainly consists of dividing the lists into three groups, where each group is responsible for 33% of posts.
The idea is that the first group, made up of those who post the most, will be small. Then the second will be a medium sized group, and the third will be the largest of them all.
Diving all users into these three groups should give us a core/periphery structure, albeit a three-layered one rather than a two layer dichotomy.
The three layer is of interest to me because I want to study the middle group, which is where — allegedly — most of the innovation comes from.
You see, the people in the core (those that are heavily involved in a community) are far too set in their old ways, and are therefore less likely to think outside the box. The people in the periphery (those who are less involved in the community) do have some interesting ideas, but they are often ignored by others in the community. Under certain circumstance this is not a problem, since the core and the periphery rely on each other to “get things done”, forming a defacto division of labour.
A division of labour operates within this core/periphery structure, where the core members and the periphery members each have a specific job to do in the community. For example, the core members are like the town elders who are responsible for making sure things work the way they should. The periphery members are like the new kids in town, with fresh ideas and ideologies. The new kids can tell the old people about the world out there and about how things have changed, but ultimately it is still up to the elders to decide whether those changes will be adopted or not.
This means that the core depends on the periphery for innovative ideas, and the periphery depends on the core to get the ideas adopted and implemented.
However, sometimes the difference between the two groups is so big that new ideas never get implemented.
In terms of open source developer communities, this is exactly what happens. A stranger can come in with an amazing idea for a feature or a fix, but the core developers may ignore it because explicit or implicit community rules on participation were not followed. Sometimes, the core member might look at a message, look at the user who posted it, and dismiss the entire message with one word — “n00b”.
This is why you need the guys in the middle.
The guys in the middle are community members have been part of the community long enough to follow customs and build a status, but remain sufficiently detached so as to be exposed to different ideas and philosophies. Therefore, the guys in the middle not only have access to the outside world and innovative ideas, but they are also better at getting the core to adopt these new ideas.
These are the guys I want to study.
This leaves me with the task of diving a list of users according to the number of contributions (or posts) they made to an Open Source project, so that I end up with the three core/middle/periphery groups.
Anyway, the next post will deal with how to do this using R.