Finally, there was a workshop organized by MobileLifeCentre which I considered really important: how to conduct ubicomp in the large considering all these app stores and mobile devices...
First the participants had 3-5 minutes to briefly report about their work and experience with ubicomp in the large. Examples ranged from social soccer fan apps [1] to architecture planning [2] and public displays [3]. Myself I briefly introduced Appaware but quickly realized that most of the people already knew about it when they showed it to me on their phones.
During the second part of the workshop we had a lively discussion about
- how to get users?
- how to keep control of the quality of the results?
- which research questions suitable to be investigated in the large?
Most importantly, the apps to be distributed have to create some value for the users, be it a game or a productivity app. Hitting the blogosphere is great which is probably best achieved by a well-designed app.
Keeping control of the data quality probably is a tricky one as we do not know who the users actually are: a bunch of people with nicknames. This might overcome by linking facebook profiles to the app, but this requires some added-value functionality of the app justifying this link. Throughout the discussions we found out that with every question we add to an embedded questionaire we limit and design the target group of users: only a subsets likes to answer - pure self-selection. Giving rewards and payment might be an incentive, but this again also incentivizes another special subset of users. There is probably nothing to be done than to try to capture as much information as possible from the users which has to be done in a transparent way with opt-out.
Finally, we struggled about why doing research in the large at all?
Reason were: because it matters more, gets closer to reality, finds unexpected uses, promotes the own lab, creates longer lasting results, gives something back to the public, yields more justified statistics...
In order to distinguish from pure app developments there should be a research question beyond simply finding out whether the crowd would accept a certain app.
I'm looking forward to the special issue following up on this topics soon to appear in IJMHCI.
Thanks to Henriette Cramer, Nicolas Belloni, Mattias Rost, Frank Bentley and Didier Chincholle for this great workshop!
