September 17-20, 2001
ECSCW 2001 is an easy to follow single track conference. It comes in alternance with the CSCW conference in the US. The only previous CSCW conference I had attended to before was CSCW 1998 in Seattle. I didn't participate to the workshop day nor the tutorial day, so this report is about the main conference. My contribution to the conference was a poster presentation.
There were 249 registered participants, 30 of whom could not come mainly because of the US past events. For instance, I could not meet Wendy Kellogg and Thomas Erickson from IBM Watson research center, who were supposed to present three posters about their textual chat/conversation management tools based on the concept of social proxies.
In my opinion, nothing really new was presented at this conference. The only surprise is the apparition of wearable computing as a CSCW research issue (one keynote and the only panel). The only computer science and hci papers were Prashun Dewan's paper about software architecture and session 9 papers about delays and interruptions in groupware applications. Most of the other papers were fieldwork studies. However I am still surprised (I had the same feeling in 1998) that there is still no common language or framework for presenting fieldwork studies. The resulting impression is that these studies are very much subjective. My idea is that CSCW is beeing invested more and more by social and communication scientists.
To my surprise I could not see anybody from W3C at the conference. No presentation even mentionned the web standardisation process (data modeling on the web or the semantic web for instance), thus I am wondering if CSCW is not missing something. I have suggested a couple of people (Allan MacLean and Paul Dourish) to organize a workshop around these topics, but it was late and after the social diner :) so I do not know if this idea will spread.
My last impression is that the W in CSCW (for Work) can be relaxed as more and more projects are addressing leisure time (the article about music sharing, on-line drama in 3D worlds, the wearable computing panel).
Bonn is a convenient place to organize conferences: good public transportation system, green, and not the least, acceptable wine production area :)
The city spreads along the Rhine, North-Sud. My hotel was located south, 108 DM a night, 3 subway stations from the conference center. I was happy to learn that 3 ministers from Ghana, India and Zimbabway had already choosen that hotel !
I was about 12 subway stations from the central train station and the city center, but subways are running late until 0.55 pm.
Polical representatives made the usual pro-science discourse. However, I was amused that they insisted on the fact that collaborative technology and all its promises would be useless without unions and people making their own choice of what's good for them.
David May's keynote on Why wear a computer? seemed to have been prepared during the flight from Bristol to Kolhn. It just contained incantations about all the wonderful information we would get from all these wonderful wearable sensors, displays and GPS systems.
The presentation contained only text, no picture, and David May didn't even wore any computer... He could not really propose credible applications of the wearable computing technology. He has just cited some experiments made in his group, some of them about soundscapes. As I understood it, it is the retrieving of sound left by others based on your location in the campus. This idea has been published in a previous CHI conferences by a MIT Media-lab researcher he didn't mentionned.
Design patterns applied to classification of fieldwork studies (named after Christopher Alexander's book).
Carla Simone's design patterns are called classification schemes and have been worked out ten years ago. It was surpising nobody asked any question.
The first article was about Petri-net modelization of access control in shared workspaces. Lot of impressive diagrams. The authors pretend the model is easy to understand because of the spatial metaphor, which is more intelligible than access control lists.
Then, a CASE tool for Extreme Programming which displays awareness information into the code browser and gives some control of each programmer's mode of collaboration. A mode of collaboration is the readiness of a user for interruptions and feedthrough (it makes me think of the way you control your accessibility with ICQ).
Extreme Programming is based on the specification of simple programming tasks on cards (for instance write a function to convert euro in DM). The cards are then shared and assigned to different programmers. They can serve for tracking the progresssion of work. Complex programs are made by assembling cards.
Somebody asked an interesting question about the CASE tool: is awareness a property of the tool or a general mode ?
In his opinion, we work in different collaborative modes associated with each tool we use. For instance when I work on an article with a word processor I may accept that the others know that I am available for a phone call, but not when I am writing some code.
Prashun Dewan presented an article about distributed architectures and toolkits for building groupware applications. That time Prashun has put together all his previous toolkits into one: you can now choose layer by layer the replication architecture (centralized of replicated) of your code. He made some experiments to compare performances according to the architecture/processor speed.
My problem with Prashun's work is that he has always considered that users want to collaborate with synchronous shared data. I think this only corresponds to few situations. In my PhD, I have proposed some other collaboration styles based on mobile instead of replicated data.
Heath's article is another contribution to the observation of embodiement and accountability of interaction rules and social norms into physical artefacts. They observed a shift from face to face collaboration towards object focused collaboration in different experiments (remote assistance through a video link, etc.).
The next talk reported the use of a 2D 1/2 window manager (you can tilt 2D windows onto a 3D projection view of the workspace, as with Xerox's Perspective Wall). I cannot remember what was specifically collaborative, maybe these windows were used to create workspaces linked with different tasks, and all the group member's workspaces were visible on the shared desktop. Just a remark about this presentation: the speaker used to display the current line in his slides in a bigger font, which became smaller when the next line of text appeared.
The last article reported on a virtual world where multiple users create dramas. They were insisting on a magic hand pointer that is manipulated by specialized users to help other ones. I really wonder of the generality of the results of that study.
A paper about Napster usages followed by the description of the design of the Music Buddy, a web interface for sharing music. In the paper I have learned that the first music piracy act was the fact of Mozart copying Allegri's Miserere in 1769 (by reproducing it by memory after one audition).
The second paper presented an interesting work about recommender systems that provide users of a web site with collaborative filtering. The authors have designed a system to recommend movies. This could be integrated for instance in a video shop web site. The system goes beyond some limitations of Movie Lens, a previous system where no social interaction really occured between users. The new system uses some of the irc concepts (group creation, group operators, etc.) but applied to an asynchronous web based BBS. I have liked the discussion of rating functions which illustrates different ways to combine votes into usefull recommendations.
Finaly the last paper is a study of teenagers' use of Text Messaging on mobile phones in the UK. They have logged messages during a few weeks. The result is an average of 3 messages received a day, and the observation of new behaviours. For instance, a common practice is to use Text Messaging as a control channel to organize a phone conversation. There is a discussion of price impact on media use too.
This was a fake panel where each participant has made a short presentation of his views or of someone else's views. The technology behind are Reactive Rooms (or roomware) and Wearable Computers. Functionalities are encapsulation (as I understand it: beeing part of a whole) and augmentation (having more information and more processing power at hand).
Lucy Suchmann presented Steve Man's provocative research on wearable computing applied to personal video capture and diffusion. Steve Man has now moved from MIT to Toronto for a House of the Future project (or maybe that was the project at MIT I cannot remember). From the ethical viewpoint, Steve Man's vision is that in an environment where we are monitored by objects and by other through video, privacy can be traded for utility.
One question asked was: is research in that field was more than incantations seeking for new applications for the underlying technologies? I can try to reformulate like this: now that every one has got his PC, how to make people buy more than one PC ? I still do not know the answer...
Papers in this section were explicit field studies: two in a medical setting and one at a mechatronic devices constructor.
For the second presentation, the speaker, Yan Xiao, was joined by phone in the US. He made a distance presentation, asking a student volunteer to change slides. Curiously he was loudly applaused, as if his performanced was enhanced by the distance.
All of these studies always evolve around the same concepts, but still without common vocabulary/methodology. Key concepts are:
I could not really visit the poster and demo presentations as I had to present my poster. During the 2 hours I made an intensive demonstration to 5 or 6 people (got some of their cards).
It seems that my ideas have been welcomed even if nearly anyone could describe me a work he had done in the past or others had done with similar concerns. The most original part of my work seems to be the specification of documents with contracts attached to fragment instances (and not based on a document class). That's what I call a bottom-up approach. These discussions gave me ideas I should write about in another place. Paul Dourish also asked me if I had published anything, which makes me think I should submit something to the next CSCW conference as soon as my document farm prototype is ready. Now, I remember I have forgotten to talk to him about the CIDE paper, I will send it to him...
A five stars restaurant in a state residence. We were dispatched by groups of 8 to 10 around round tables. Wine and food quickly warm up the ambiance. There was an after party at the conference center where the bus dropped us.
Missed that one (and that's a shame because it may have been the more computer science and HCI oriented session in the conference).
Larry Prusak from the IBM institute for Knowledge Management, Cambridge, must be touring the world with the same talk for the last 30 years, without changing a single word either he talks to managers of SME in Tucson Arizona, to CEO of big corporations in Davos, or to CSCW researchers. Just a few extracts (rephrased with my words):
And unfortunately no source was cited so that we could learn more. All we could do was to rely on the great experience of the speaker in that domain.
Kaplan and Seebeck were not here to present their paper.
The second paper was a fieldwork study of a group of snow vehicle drivers in an airport after the introduction of a new display system in their truck. It seems that this visual aid was not really adopted by the drivers who prefer to coordinate through verbal communication.
To present the last paper, Geraldine Fitzpatrick read the slide commentaries sent by email by one of the authors. It was a study of the perception of complex data sets shared by molecular bologists. Yet another study of shared objects. The originality of that study is to be based on interviews (and not video recording of joint activity as it is often the case).
Forthcoming events: