tecznotes

Michal Migurski's notebook, listening post, and soapbox. Subscribe to this blog. Check out the rest of my site as well.

Oct 10, 2008 5:49pm

dunbar's dungeon

Thought experiment for the day, spurred by jets flying overhead for fleet week. If something out of ordinary is going on, and you need to ask people around you, what do you do?

Imagine a chat system, similar to IRC or AOL - a big room where you log in and talk to lots of other people simultaneously. When you log in, your Fire Eagle account gets a ping, so that the server can know your geographic location to some level of precision. You are then dumped into a big room, with the 150 people physically closest to you in the real world. As long as there are fewer than 150 total people in the system, everyone gets to talk to one another. As more people join, you begin to see overlapping conversation bubbles. You might be in San Francisco, talking to someone in Kansas. That person in Kansas can talk to someone else in New York, but you can't. Your conversational circle is strictly limited to the nearest 150 people, some of whom drop off occasionally as they are bumped out by more proximate newcomers.

As the population of the system grows, everyone's personal horizons begin to shrink. With enough people, eventually you're talking to the people right in your neighborhood. To get a message to someone across the country, you might lie about your location, or ask that it be passed on, Milgram-style.

Would this feel like a natural way to interact with people around you, self-limited to a reasonable number of participants but always those around you? Would people lie about their location? Would it be like Usenet, dense with microscopic subcultures? Would certain people emerge as hubs, offering to post messages around between bubbles in the absence of other means of communication?

Comments (15)

  1. I think you might want to allow some way to encode asymmetrical/situational abilities to cope with information flow into the system. There are times when I can effectively deal with informational flow generated by the conversation of significantly more then 150 people (well rested, focused, hooked up to a fast connection, etc), but often I'm not acting at anywhere near this optimal state. Now if the idea is to re-create the villages of Dunbar's strained analogies, then having the membership of the community flux too quickly (if say you saw a late-stage Web 2.0 style usage spike) would be problematic. But yes, we need to start building tools around group forming as identity, rather then individual as identity.

    Posted by kellan on Friday, October 10 2008 6:40pm EDT

  2. There was a service along these lines: an IM platform with location awareness: http://www.meetro.com/ They've shut down. -d

    Posted by Daniel Howard on Friday, October 10 2008 7:04pm EDT

  3. I never did try Meetro, like a log ot things I guess the idea is a lot more interesting than the actual service. =) Daniel, did you ever use it? Was it worthwhile? I think I sort of filed it away with the other double-vowel web sites of 2006. Kellan, what do you do when your mental state changes and you can/can't handle large numbers of people? I feel like there's been all these suggestions of pause buttons for websites or services (notable the microbloggy ones - Twitter, FFeed...), which suggests that it should be possible to build in some form of attenuation to each system. A dial, a button, a switch that rapidly shrinks your horizon when you're feeling too put-out. Sort of the equivalent of a hot cup of tea and a couch in sweatpants. I'm liking how Fire Eagle actually turns your account off when you haven't been around for a while. It's been fun kicking around the implications of flows and Jabber (e.g. my previous post) even now, over a year after your XTech talk with Blaine. I keep feeling like the whole design approach is more suited to the behaviors we're clumsily modeling with Rails-type frameworks right now. Also liking Tom Insam's idea of making message queues cheap enough that they're a day-one feature of new personal projects, like databases.

    Posted by Michal Migurski on Friday, October 10 2008 8:24pm EDT

  4. A fantastic idea Michal. I've been pondering something similar for a while, but never considered the idea of overlapping circles of visibility. I'm tempted to try and build it...

    Posted by Tom Taylor on Saturday, October 11 2008 5:37am EDT

  5. Hi, I'm the creator of Graffitio, an iPhone app which offers much of what you are saying. One of the first iterations of Graffitio did close to what you are saying. I decided against that idea because of the overlapping circles of visibility. let's take persons A, B, and C each 50 miles apart along a straight line. Let's say to get to 150 people, each of them pulls in everybody within roughly 75 miles. That means B can be talking to A and C simultaneously. To person A, half of B's messages won't make any sense, as they are intended for person C. What would also be confusing is that person B would see messages from A and B, but would not realize that A and B can't see each other. With Graffitio I solved this problem with "Walls". Instead of associating users with a location, a Wall is associated with a location. Anybody can create a wall. When you start the application, you see all the walls in your area. In addition to these walls that are points, there are City, State, and Country walls, so you can always find people to interact with in a manner similar to what you've mentioned. Of course, like so many ideas, anything in this realm is heavily dependent on critical mass. I'm working on some ideas for Graffitio to try to make it still be valuable as the community builds up to that critical mass.

    Posted by anoopr on Sunday, October 12 2008 1:32am EDT

  6. Hmm… Reading the post, I thought right away about the association of my identity with my location. This is "dangerous", except if you can go into it pseudonymously. The idea of discussing with local persons is interesting and is happening for example in Yahoo! Chatrooms. There are groups by cities for example with meetups. I have been to those in Montréal. As to be automatic (FireEagle) to be thought. I have a tendency has a user to prefer the opt-in as opposed to the opt-out. Still pondering. About location, you need granularity on the lie ;) Big lie (another state, country) to micro lie (another street). When there was the trend of geolocalizing your blog with meta name, I was giving the central point of the city (in Montreal, it was the top of Mont-Royal).

    Posted by karl on Sunday, October 12 2008 8:05am EDT

  7. sounds cool but it could be limiting.

    Posted by billy on Sunday, October 12 2008 9:59am EDT

  8. Karl, I agree that all the usual caveats around location and privacy apply here, even in this silly thought experiment. Actually have you had a chance to look at Fire Eagle's way of handling this? There are some interesting choices they've made. When you opt to share your location with some 3rd party application, you can do so at a variety of granularities, from country all the way down to precise position. What I find interesting is that they've massaged their place hierarchy somewhat, so that your choice of granularity for any given spot in the world has a standard level of population-based fuzziness associated with it, so that say "neighborhood" generally means that you shouldn't be easily findable whether you're in a small town or a big city. Also it's clear that I need to make sure this site is served as UTF-8.

    Posted by Michal Migurski on Sunday, October 12 2008 1:28pm EDT

  9. @anoopr: I quite like hearing one side of the conversation on Twitter, as it effectively gives me one more degree of sensitivity to a topic of interest, and often leads me to find people who I wouldn't have otherwise spotted.

    Posted by Tom Taylor on Sunday, October 12 2008 4:18pm EDT

  10. Okay, keep going while I ask Tom Coates to get in here and talk about the thoughts behind <a href="http://www.plasticbag.org/files/misc/etech.ppt">UpMyStreet Conversations</a> [PPT]...

    Posted by Matt Webb on Sunday, October 12 2008 6:47pm EDT

  11. Oops. Right. Er. So my big worry with the model you're proposing here is that I think it might be focused on easing one aspect of sociality but inadvertently cause some friction in others. So, yes, social stuff is more interesting if the people have geographical proximity, but if you're personal experience of the process is of a cycling of people through your local community, with you occasionally popping into a different community, and people around you doing something similar, then you're also losing the continuity and ability to build up relationships with people which I think makes the thing less appealing. I wonder if it wouldn't be like someone deciding to optimise your friends every day with people who were more like you and who you'd find more interesting and fun, but with the consequence that you could only be friends with them for a couple of days. We took a different approach with Conversations, although I don't know how much better it was. It was very much a message board, although it was one in which the initial posts in a thread were geocoded rather than the people themselves. You could place a post in any place in the world, and see threads that had been posted near that place (Stef Magdalinski made the really interesting argument that smoothness of proximity was a big deal here - ie. show me the twenty nearest ongoing threads close to this street address). The good thing about it was that the individual got to travel around independently, could go and visit those other places that were important to them, etc. One of the big problems was keeping it agile and helping people spot what conversations were ongoing (both things that contemporary mapping tech could probably solve really easily). It was also quite difficult to deal with the fact that when you start digging into local conversations you start exposing a whole range of unpleasant things. UpMyStreet Conversations had a number of threads ongoing in far-off parts of the UK that were essentially far-right tracts. Plus the simple axis of timeliness (which threads had been updated most recently) tended to get swamped by the proximity stuff, making it hard to spot activity in your threads. What I'd probably suggest now is that any service like the one you're suggesting didn't think of itself as entirely organised around location, but used location as one way of cutting an endless set of conversational threads. If you go back to asynchronous communications, and give people the local view alongside the subject-based view or the view of threads that their friends are active in, then you might find that you support all the various was in which people might want to chat, while bringing in the geo aspect effectively. I'm sort of going on a bit now, so I'll stop.

    Posted by Tom Coates on Monday, October 13 2008 1:51am EDT

  12. People will definitely lie. There's an academic paper somewhere about a Swedish project called GeoNotes which allowed people to post virtual notes attached to physical places. One of the physical places used was "floor 41, VIP room" or somesuch, that was attached to a seven story building :) Also, FireEagle might be useful for privacy reasons, but otherwise you might also use Skyhook Wireless (or Geode) for immediate gratification.

    Posted by Tom Carden on Monday, October 13 2008 1:57am EDT

  13. (Can someone please get Tom Armitage in here for a Complete British Tom-In?) So this is interesting, you guys are really picking apart the geographic aspect of this thing, I think it might have been a bit of a red herring on my part. I'm actually most of all interested in the shifting horizon, and the focus-dependent group memberships. Tom Taylor, you mention the one-side-of-a-conversation aspect of Twitter, which I have also found myself enjoying very much. It's interesting to use that a mechanism for finding people or at least confirming their identities based on conversational context, like being able to see around corners with other people holding mirrors for you. Actually this kind of gets into the continuity thing that Tom Coates brings up - a lot of the times I see a new name referenced on Twitter, I figure out that they're someone I know in meatspace but don't take the step of making the relationship explicit on Twitter. Having already dealt once with the RSS feed explosion that ate up a ton of time, I'm wary of doing the same thing with a Twitter stream run-up. How often do you see people join a new service, go totally bonkers adding everyone, and then quitting in disgust because it's too much? I did it in January of '07 or so, and had to leave Twitter for a few months to regain some comfort with the sudden flush of intimacy. One place that I'm really enjoying at the moment is http://fivethirtyeight.com, vibrant and very new community almost entirely organized around the "Win Percentage" pie chart (upper left corner) and how nice it is watching that red slice shrink. There's an absolute mob commenting on every thread, there for the frequent updates and a bit of camaraderie. Once November 4th has come & gone, I'm sure all those people and their various in-jokes and sub-communities will disband back to whatever worlds they've come from. I'm sure that for a red-slicer, a lot of 538 looks like one of your UpMyStreet right-wing nutfarms. The internet is like a very potent indignation amplifier for these temporary conversation pits. Tom Carden I'm going to pester you at work this week about invented places standing in for geography. The 41st floor thing is some serious Narnia wardrobe business.

    Posted by Michal Migurski on Monday, October 13 2008 3:06am EDT

  14. This is a fantastic idea. While I can't speak to the technical part of what you've described, I think it could be extremely useful as a way to connect various communities (and not just by geography). Have you thought about applying for a Knight News Challenge grant? http://www.newschallenge.org/

    Posted by Chrys Wu on Wednesday, October 15 2008 1:18am EDT

  15. "What I'd probably suggest now is that any service like the one you're suggesting didn't think of itself as entirely organised around location, but used location as one way of cutting an endless set of conversational threads. If you go back to asynchronous communications, and give people the local view alongside the subject-based view or the view of threads that their friends are active in, then you might find that you support all the various was in which people might want to chat, while bringing in the geo aspect effectively." I liked that idea from Coates. To push it to an extreme - People, Place, Subject: pick two. There's something interesting about designing around a kind of restriction, and you could potentially get around some of the issues by making the locational coincidence something you choose to replace a different axis of data in your view. You're always pivoting n-dimensional data into a view of it with less than n dimensions; what becomes interesting is not just the information, but the way it changes as you pivot it. Information emerges from the act of pivoting. (Hello! Full house!)

    Posted by Tom Armitage on Tuesday, October 28 2008 5:29pm EDT

Sorry, no new comments on old posts.

October 2017
Su M Tu W Th F Sa
    

Recent Entries

  1. planscore: a project to score gerrymandered district plans
  2. blog all dog-eared pages: human transit
  3. the levity of serverlessness
  4. three open data projects: openstreetmap, openaddresses, and who’s on first
  5. building up redistricting data for North Carolina
  6. district plans by the hundredweight
  7. baby steps towards measuring the efficiency gap
  8. things I’ve recently learned about legislative redistricting
  9. oh no
  10. landsat satellite imagery is easy to use
  11. openstreetmap: robots, crisis, and craft mappers
  12. quoted in the news
  13. dockering address data
  14. blog all dog-eared pages: the best and the brightest
  15. five-minute geocoder for openaddresses
  16. notes on debian packaging for ubuntu
  17. guyana trip report
  18. openaddresses population comparison
  19. blog all oft-played tracks VII
  20. week 1,984: back to the map

Archives