A Mixed Bag of Filesharing, WiFi, and Me Talking about Wikinews

And we're in the first Association of Internet Researchers conference session for Saturday - unfortunately I couldn't blog the first presenter as she was running her Powerpoint off my laptop. Sunyi Lee from Northwestern University presented on possible business and licencing models for p2p filesharing, and ended with a useful point on the change of the conceptualisation of music, from music as product (selling CDs, DVDs, etc.) to music as service - where users may pay for access rather than distinct units of merchandise.

Sorin Matei: Mapping WiFi and Encryption in Lexington

The second speaker is Sorin Matei from Purdue University, presenting on the process of diffusion in wireless networks. Can there be a predictive model for the diffusion and encryption standards in wireless networking technologies (focussing here on WiFi, 802.11 standards)? What is interesting about WiFi is that at least in the beginning it was a replacement techniology for ethernet LANs, but was soon sold as a technology of freedom (from wires) in the residential market, creating always-on, personal connectivity. Further, WiFi can also be seen as a 'realm of dissent' in which the 'community network' movement can reinvent itself.

The main weakness of WiFi is that its radio transmissions can travel outside of the intended area (the home or business), which raises issues around abuse and privacy. The way to work around this is encryption, of course - originally WEP, with other protocols now also becoming available. Encryption, while possible, is however rarely used.

So, at what pace do wireless networks diffuse, and how prevalent is encryption? What factors explain the diffusion of encryption practices? Sorin studied this through observation, taking advantage of the easy detectability of wireless networks - essentially he conducted a kind of wardriving in Lexington, Kentucky using the Network Stumbler application to identify wireless access points and their encryption practices - however, in difference to wardriving his study was less random and based on selected sampling points related to the existing population density (more sampling in more densely populated areas).

To develop a large enough sample of locations Sorin also developed a sense of the shortest path between locations, and so overall about one quarter of all city streets in Lexington were covered, with one sample point per thousand residents at which WiFi availability was sampled, and the closest route between each point travelled and mapped as well. This was done twice, in August 2003 and May 2004, and it identified some 2652 unique WiFi access points; each was hypothetically identified respectively as business, residential, and public, using the Lexington zoning map. Using this methodology, the distribution was around 66% residential, 21% business, 13% public in 2003, and 74%/16%/10% in 2004.

Encryption use increased from 2003 to 2004: 18% residential, 23% business, 17% public points used encryption in 2003, which changed to 28%/32%/26% in 2004. Sorin now asks whether social capital explains an increase in encryption in the residential areas, and he mapped his findings against socioeconomic data in Lexington (as an indicator of social capital).  Areas where more encryption took place were correlated with areas where more individuals are likely to be members of community organisations, and wealthier, Sorin found, and encryption therefore seems to be explained by the presence of structural social networks, rather than interpersonal networks. But this is still to be further studied.

Axel Bruns: The Problem(s) of Wikinews

Finally to my paper - of course I couldn't blog it, but please see the notes I've already posted here. I also recorded the session and will post the audio here as soon as I get around to editing it.