Recently I have been on holiday and I have used this opportunity to catch up on some of my KDE number crunching. Also whilst on holiday I took the opportunity to catch up on listening to one of my favourite radio shows from the UK: The Infinite Monkey Cage. The Infinite Monkey Cage is a panel show blending science and comedy where, each week, they pick apart a particular topic. Add it to your feed reader now!
Now, this week on TIMC they were talking about degree of separation; you know, the idea that anyone can be linked to anyone else on the Planet in 6 steps. For example I can get to Barack Obama, thus: Paul Adams (has met) Prof. Brian Fitzgerald (has met) Bono (has met) President George W. Bush (has met) President Barack Obama.
Now, it is possible for us to use this as some form of (extremely weak) quality metric; the closer you are to some known high-quality source, the better. Within the mathematics community they have done this for a while. Mathematicians have a whimsical measure known as the Erdos number. Paul Erdos was a prolific contributor to mathematical literature, co-authoring papers with many people. He himself, has an Erdos number of 0. Anyone who wrote a paper with him has an Erdos number of 1. Anyone who wrote a paper with someone who wrote a paper with Erdos has a number of 2…. and so on.
For completeness, I have an Erdos number of 5, with the following path to the great man himself: Paul Adams (co-authored with) Diomidis Spinellis (co-authored with) Bart Preneel (co-authored with) Peter Landrock (co-authored with) Carl Pomerance (co-authored with) Paul Erdos.
For any finite network we can, of course, work out who is the most connected person in that community. By that I mean that we can find the one person who on average requires the fewest steps to get to any other person in the community. This principle was popularised by the Oracle of Bacon. The Oracle of Bacon is a online tool which links actors whom have acted with each other in order for you to find out how long their path to Kevin Bacon is. Kevin Bacon was chosen because at the time the tool was first developed he was, mathematically, the center of the Hollywood community.
I do not have a Bacon number.
In the same spirit, therefore, I would like to introduce the Oracle of ERVIN.
The Oracle of ERVIN is a Python script I wrote to determine who is the most connected person in the KDE community. KDE contributors are linked on the basis of “has worked with”, by which I mean that they have both committed in the same file in the VCS during the period represented by the input log. The first “real” input I used with this script was the log of KDELIBS in 2010. For that input, ERVIN was the most connected committer in that community, requiring an average of 1.6 steps to get to any other KDELIBS community member. So the tool was named after him.
I have now run the tool against the entire KDE SVN log for 2010. So, here’s a question: Who do you think was the center of the KDE community last year (and why)? I will report the answer sometime later in the coming week.