Ubicomp 2008 is over

Ubicomp 2008 in Seoul is over now, and it was quite an interesting experience. Not only because of the papers, many of which were very inspiring, but also because of the setting: Seoul is quite metropolitan, but as a Westerner, you can neither understand anything nor read any sign, and there is this interesting 'abstraction' of other cultures - we had lunch in a Bräuhhaus, which was serving sushi and pizza. Add to that a good helping of jet-lag, and suddenly lost in translation seems to resonate very strongly with one's experience.

Anyway, what still lingers in my mind is the closing panel, which discussed (to some degree) what Ubicomp is/should be and where the conference is going as a whole. Gregory Abowd made a similar point he made last year about the need to look into problems that really matter to real people, which I strongly agree with. In a way, this could be one defining factor of what Ubicomp is about: bringing together results from various disciplines and integrating them in a thoughtful way to make a real difference in people's life. I think this year the number of papers that followed this general idea has increased compared to last year.

Another clear trend is definitely energy efficiency - I very much enjoyed the talk about 'the potential for location-aware power management' by R.K Harle and A. Hopper. Combining this trend with the above, might lead to a generic approach to Ubicomp, where the underlying question is 'what can improve a particular situation for the people in it, and how can we realise that with the minimal amount of resources (technology-wise and energy-wise)?'

The other issue discussed at the panel was the future of the conference. I guess everyone agreed that it would be good to have more industry-involvement and to have a stronger exhibition/demo section. I'm not sure whether I agree with the idea of merging all ubicomp related conferences into one big event. Comparing the CHI experience to mobile HCI and Ubicomp, I think I prefer the smaller, single-track event. And for Ubicomp in particular, becoming too big might be quite counterproductive in terms of mixing up researchers from different disciplines (which in my opinion is one of the key strengths of Ubicomp).