Review for Power of the Few vs. Wisdom of the Crowd: Wikipedia and the Rise of the Bourgeoisie

The paper is an interesting account of how wikipedia has changed in the past few years from being created and amended by very active users and administrators to being adjusted more frequently by less-active users. The main findings are interesting, however, unsurprising.

It was intriguing to see the amount of data that the authors have analysed and many of the graphs are very interesting, in particular in combination. Still, I had the sense that it was mainly pointing in the direction of one not-so-interesting or unsurprising finding: That Wikipedia has become more well-known from 2002 to 2006 and therefore more commonly used among broader segments of the population. This might be what the authors aim to contribute with, but the structure of the paper takes us on a sometimes tedious journey to explain all the data based indications, sometimes in a slow fashion. The first part of the analysis 'Rise and Fall of Admin's influence' for example completely ignores the changing number of admins, and it is obvious that there is a need to know this number in order to understand the graphs. Not until figure 7, do the authors look at population growth in the different user grades, however, not including admins. Most of the figures support each other, but before the paper looks as the type of edits, the conclusions are quite simple: a rise in users generally. It is interesting to see that the low-level users are deleting more than they are contributing with, and such finding could easily be explored further.

The topic of evolution of online networks like this is very interesting and the authors are on the right track of presenting their findings. However, my main concern with the paper is its vague and almost obvious finding, the rise of the common people. With so much data available, the paper could easily bring out more detailed conclusions and look at the special cases, such as the spikes and exceptions in many of the graphs in 2002 (was there a case of vandalism here?). The comparison to del.icio.us is interesting and support the conclusion, but again, could be fleshed out more.

The discussion look at how the patterns found correspond to other processes of technology adoption, but there is no talk about possible consequences or future directions for such systems. Even the conclusion asks 'who writes wikipedia?', however, this is not attempted to be answered even though you have data to explore this. By the end I was still curious about this, since the bourgeoisie writes few words and delete more, is it the admins and high level users who write wikipedia?

All in all I think it is an interesting topic and there is a lot of material. However, as the paper stand now the contribution it not interesting enough, but I look forward to seeing a revised published paper later.

One small thing: the method section is very vague, it is not described for example that the data dump used has record of edits, throughout the years (is every little tweak recorded in the data dump?) , this only becomes apparent when reading the analysis part.