François Proulx is the coordinator and lead developer at MIT's iFind project. iFIND is MIT's new location-based application for "friendspotting" and among other things, serves to bring Web 2.0 social networking into the academic realm. To find out more about this upcoming talent, turn your attention to his blog or his flickr page.

How did iFind get started?

Exactly a year ago, the SENSEable City Lab at MIT launched a project called iSPOTS. iSPOTS studied the usage of the Wi-Fi network on the MIT campus. A year ago, I heard about the iSPOTS launch and I was very interested because I am an active developer in one of the largest Wi-Fi community groups in Canada. So I got in touch with the SENSEable City Lab and after our discussions, they offered me an internship.This idea has been growing inside SENSEable City Lab for the past 12 months and they had one or two prototypes, but none of them was ever released. I started working in the lab in September of ‘06 and in October we started building iFIND.We first started by developing the “location algorithm” on a Pocket PC.

What are some of the major problems your team faced with iFind and how did you address them? What do you envision as some of the major obstacles to iFind’s future successes?

From the start, we really wanted to address the privacy issue. This is one of the most significantly different features of iFIND when you compare it to other locations-based social software today. It took us some time to figure out a secure peer-to-peer scheme. We focused on a distributed data approach and so data was client-driven as opposed to server-driven (central server). As for major obstacles we envision, it is to see a similar system grow outside of a closed-campus. Because in order for that to happen, we need to enhance 802.11 to include location data. A very interesting document is RFC 3825 which talks about adding location data to DHCP responses.

The biggest advantage I can see with iFind is it solves the privacy issue of data location sensitivity, with a distributed data approach. How does that tie in with this secure peer-to-peer scheme you mention?

You can always choose to whom you will disclose your location, but what is really important is to ensure you are authenticating the receiving end.So public key cryptography is essential. As a result, we built our own public key infrastructure for iFIND. iFIND generates a unique certificate for each new user.We still need a central server to store sessions and exchange IP addresses between friends but after that, each user is “free” to talk in their “own” peer-to-peer environment. We make it a point to only keep the last session and no location data whatsoever.

To realize MIT’s Information Services & Technology’s goal of creating a fully wireless campus, for any other coordinators behind campus or city projects that are trying to do the same thing, what skillset do you recommend for the team to have or what needs to be in place?

Some of the tools to build this software are really cutting edge so we needed a team of people who not only are constantly looking for the latest developments and improvements but more importantly possess the skillset we need in using those tools efficiently and creatively. When I say tools, I mean libraries. Having a team of people proficient in tools like Sun's SwingX-WS project was critical to this project’s success.