The use of attendee backchannels at conferences, a a favorite theme here, is part of a larger trend, towards ad hoc organization, or even ad hoc creation of value.
You can see the context backchannels are happening in by looking at the Users create the schedule process for this weekend’s 2004 Planetwork conference.
Anyone can propose a topic, anyone can create a login to rate a topic, and the half-hour speaking slots are given to the top ranked topics. To get such a slot, a talk needs to be both highly and broadly rated. (In subsequent passes at this method of selection, organizers will have to work against gaming-the-system options, of course, but the current style is fine for now.)
Interestingly, the Planetwork folks have handed out the first 3 half-hour slots, and are going to do 3 more on June 2nd, and 3 more on June 3rd, meaning that the conference emerges over time. It also might let voters optimize the slots over time, as they see unaddressed topics and vote related proposals up.
I say might, because it’s not clear how coordinated the voting can get in this framework. One class of risk in this system is ‘slashdot risk’, named after the reflexive stance on slashdot in favor of Linux, making even well-meaning criticism of that OS much less popular than even the most vapid pro-Linux boosterism. Groups have a hard time selecting topics or speakers who violate their cherished assumptions, so the interface could in certain groups amplify existing prejudices.
The emergence of new classes of risk, however, is inevitable (as with ‘clique risk’ that happens in backchannels) because the weakness of the current conference form is so great that new ways of handing power to the users, however beset with problems, will be preferred by the users themselves.
The social dilemmas of a conference are many, but most of them can be grouped under one heading: social loss. At a large, topic-specific conference, there are several obvious forms of loss
- the conference schedule not matching the interests of the attendees (which Planetwoirk is trying to address)
- speakers and panelists not being asked to address hard questions (“So, tell me Bob, just how good is your proprietary product?”)
- members of the audience can have more knowledge than the speakers (as with Alan Kay being lectured to about object oriented programming)
- members of the audience preferring to speak with one another, in groups, alongside or instead of listening to the speaker (In-room chat as a social tool.)
Conference organizers will object that these new styles of arranging and participating in conferences will do more harm than good, and in many cases that will be true, but it won’t matter, because the real change here is not that technology is allowing new forms of participation, but rather that it is allowing new forms of creation — a conference has heretofore been an artifact, crafted by a small group for a large group, and as usual, the small group has found many ways to justify its existence (and I say this as a veteran of conference planning.)
The ace in the hole, though, was capability — the small group model is required because the coordination cost for the involvement of a large group is simply too high. Whatever arguments there might be for involving attendees directly run aground on the difficulties of actually doing anything about it.
Until now. Because of its plasticity, because of the tech-savvy nature of the road warrior clan who make up the core of its attendees, and because the “money for value” equation is quite direct, the conference form is an early warning of the pressures other social forms, better but not perfectly insulated, are going to undergo as social software continues to blow back through existing institutions.
1. Yoz on June 1, 2004 9:44 PM writes...
"One class of risk in this system is ‘slashdot risk’, named after the reflexive stance on slashdot in favor of Linux, making even well-meaning criticism of that OS much less popular than even the most vapid pro-Linux boosterism."
This isn't nearly as true as you'd think. In fact, I'd say just the opposite. I've seen complaints about predictability of *anti*-Linux postings getting modded up. And I'd agree: in my experience, anti-Linux/pro-Windows posts that aren't obvious flamebait do bubble up quicker than the opposite.
This is doubly intriguing given that the odds of being a moderator are increased by your own karma, meta-mod judgements made on you and, most importantly, whether you made use of mod points last time you had them.
What this says to me is:
1: Comments that go against the grain in a reasonably intelligent, non-flaming way are inherently more interesting - either because they provide a new and useful insight (which is usually the case in the Slashdot example: "Hey, Windows is actually better than Linux on this one feature because A, B and C") or because people just like being riled up (like the guilty pleasure of reading columnists who you violently disagree with). (Because of the latter, even flamewars are interesting in some communities... but not Slashdot, where it's a constant background noise.)
2: The moderator probability system is obviously somewhat self-reinforcing, yet it does actually seem to work, in that the highlighting of the more interesting posts happens pretty quickly and effectively. (As long as those posts are made in the first couple of hours, of course - there is a negative reinforcement as well due to anyone who regularly reads Slashdot setting their preferences to sort comments by mod score, and after a couple of hours few moderators will travel down far enough to see new unscored comments) That this works in such a rabidly opinionated community raises the further question: Is it *despite* or *because*?
This effect - the 'slashdot anti-risk' if you will - is probably the main reason I bother to wade through most Slashdot threads; I know that if there are decent contrary opinions there, I'll probably see them. If it wasn't for that, I doubt I'd bother.
Permalink to Comment2. evan on June 2, 2004 1:07 PM writes...
I like a lot the ideas Planetwork is trying to play with it's open scheduling of part of the conference. The software is basic but works. The rating system is also quite simple, but it's combined with some human intervention in terms of scheduling.
One flaw i found was that there was no way to comment and the rating was a very shallow indicator of value. Both of these get to the point of trying to address WHY something is highly rated or rated poorly.
For example, the top rated proposal right now is a non-technical discussion about what is free and open source technology. Do we really need another introduction discussion? No. I bet most of the people who rated it highly don't even want to sit through another introductory discussion. But, they do want to be discussing free software, so they rank it highly. Not having richer comments or feedback prevents us from knowing WHY people rated it hightly.
The second serious flaw is it's non-interactivity. If we are going to develop something like this then it should be integrated in to a wiki page. Every entry should get a wiki page about it created automatically for people to expand on and learn more about the topic. There is nothing to stop people from adding their own links, it's only that it's unlikly to happen if it's not encouraged.
The next step would be providing a tool which tried to find a best fit between the sessions that are requested, and who wants to go to what. I know it's a traveling salesman problem but an interesting one non-the-less.
Permalink to Comment