This year at CHI 2010, there was a very interesting paper+panel about infrastructures in HCI research called “The Infrastructure Problem in HCI“. The paper identifies different types of user experience difficulties caused by underlying infrastructures. The central conclusion of the paper was a call for the HCI community to become more involve in creating technical infrastructures, which implies a “substantial expansion to the methodological toolbox of HCI”.

This paper is particularly relevant for my own work, as we are trying to build a technical infrastructure called the P2P collaborative grid, and explore different aspects of HCI on top of it. As you might guess, I am working on the HCI part of the project, but I count with quite some experience in building technical systems and infrastructures. My engineering degree thesis alone focus on creating a software framework for supporting multi player games on ad-hoc networks.

When I joined the project, the infrastructure was well into development, and there were clear blueprints of where the technology was going to. That means that I was facing the fact mentioned by Keith Edwards in the above mentioned article: I was doing HCI research on top of a pre-existing technology that pretty much limited anything I could do. The real consequences of this limitation became clear to me in the months to come. As a good HCI graduate student I was given a setting on which to apply ethnographic methods: a molecular biology research group. I was suppose to dig out the information about the researchers and the way they use technologies, so that we could study the way they use the P2P collaborative grid. As I went through my observations I realized the biologists would never use the P2P grid for a number of reasons. First, because the grid was attached to a software suite that even thought they all have license to, only very few use it. Second, because the grid didn’t support the kind of algorithms they did care about. And third, because there already exists plenty of other tools that could do the same as the P2P collaborative grid, with less effort for the biologists.

I was shock, the very straight forward research path I was given when I started suddenly ended in a dead end, even before I did anything. However, we decided to work around this problem and came up with the notion of Infrastructure Awareness – and that’s where I plan to do my dissertation. We will talk more about this in later posts.

However, the paper at CHI 2010 brought everything back into focus. The infrastructure not only limits the kind of user interactions that can be built, but also the kind of research questions we can ask. Keith Edwards proposes to CHI community to get dirty and build infrastructures and accept them as contributions. And I somehow agree, as long as they conform to certain standards (this is a long discussion here). My gut feeling here is that there is a lot of grease in making infrastructures, and that even though making them should be part of our focus as a community, there are other ways for the community to engage with them. For instance, I will propose to examine “infrastructure critiques” as a possibility. Going back to my own example, there are many things I could see after fieldwork were wrong about the way the technology had been designed: the fact that it run attached to an unused software, the dependence on contributors, etc. I would so much like to write a piece of paper where I take my ethnographic fieldwork and challenge the given technology with it, to finalize with a section called “Implications for Redesign”.