Seb's Open Research
Pointers and thoughts on the evolution of knowledge sharing
and social software, collected by Sébastien Paquet

Webfeed (?)
email me

My keywords
My popular pieces
Stories and articles
2002 weekly archives
2003 weekly archives
2004 weekly archives
Neighborhood tour
Technorati cosmos
Blogstreet profile
Today's referers
Seb's home

My other weblogs:
Many-to-Many: Social Software groupblog
My public mailbox
My 'Quantum Bits' blog
En français SVP!

Montreal, QC
Social Software
Augmented Social Net
Emergent Democracy
New webloggers

Research Blogs
Minciu Sodas
First Monday
AmSci E-Prints
Weblog Kitchen
Knowledge Board
Weblogs at Harvard
EduBlogging Network
Reputations Research
Know-How Wiki

Paysagement M&M
JC Perreault
Smart Mobs
Blue Oxen
Web Tools- Learning

with a weblog

Spike Hall
Chris Dent
John Baez
Bill Tozier
Erik Duval
Clay Shirky
Jill Walker
Jim McGee
David Tosh
danah boyd
Sylvie Noël
John Taylor

Ton Zijlstra
Joseph Hart
Ed Bilodeau
Peter Suber
David Deutsch
David Brake
Steve Cayzer
Lilia Efimova
Mark Hemphill
Alex Halavais
Mike Axelrod
Paul Resnick
Cosma Shalizi
Andrew Odlyzko
Lance Fortnow
Tom Munnecke
Henk Ellermann
Mark Bernstein
Jeremy Hiebert
Jacques Distler
Michael Nielsen
Thomas N. Burg
Hassan Masum
Ian Glendinning
Marc Eisenstadt
George Siemens
Howard Rheingold
Stephen Downes
John Bethencourt
Sebastian Fiedler
Kevin Schofield
José Luis Orihuela
Martin Terre Blanche
Elizabeth Lane Lawley
Paul Cox
Jon Udell
Don Park
*Alf Eaton
Lion Kimbro
Phil Wolff
Jay Cross
Julian Elvé
Matt Webb
Adina Levin
*Marc Canter
Matt Mower
Kevin Kelly
Dina Mehta
Greg Searle
Ross Dawson
Al Delgado
Rajesh Jain
Lee Bryant
Jesse Hirsh
David Sifry
Jeff Bridges
Stowe Boyd
Walter Chaw
Piers Young
Barbara Ray
Dave Pollard
Ian McKellen
Josep Cavallé
Hylton Jolliffe
Lucas Gonze
Jerry Michalski
Chris Corrigan
Boris Anthony
Michael Fagan
Mary Messall
Denham Grey
*Ross Mayfield
*Phillip Pearson
Whiskey River
David Gurteen
Tom Portante
Chris Wenham
Pierre Omidyar
Stuart Henshall
Greg Costikyan
David Gammel
Renee Hopkins

Peter Van Dijk
Peter Lindberg
Michael Balzary
Steven Johnson
Robert Paterson
Eugene Eric Kim
Jason Lefkowitz
*Flemming Funch
Bernie DeKoven
Edward De Bono
Maciej Ceglowski
Charles Cameron
Christopher Allen
*Philippe Beaudoin
Richard MacManus
The Homeless Guy
Ward Cunningham
Hossein Derakhshan
Stewart Butterfield
Stefano Mazzocchi
Evan Henshaw-Plath
Gary Lawrence Murphy
Karl Dubost
*Dolores Tam
Norbert Viau
Patrick Plante
Daniel Lemay
Sylvain Carle
Bertrand Paquet - Hydro-Québec
Michel Dumais
Mario Asselin
Robert Grégoire
Roberto Gauvin
Clément Laberge
Stéphane Allaire
Gilles Beauchamp
Jean-Luc Raymond
without a weblog
Steve Lawrence
Simon B. Shum
Stevan Harnad
Brian Martin
John Suler
Christopher Alexander
Johanne Saint-Charles
Douglas Hofstadter
John Seely Brown
Murray Gell-Mann
Steve Newcomb
Howard Gardner
Anthony Judge
Patrick Lambe
Donald Knuth
Phil Agre
Jim Pitman
Chris Kimble
Peter Russell
Roger Schank
Howard Bloom
John McCarthy
John C. Thomas
Doug Engelbart
Seymour Papert
Hossein Arsham
W. Brian Arthur
N. David Mermin
Tommaso Toffoli
Brian Eno
Will Wright
Jean Leloup
Daniel Boucher
Daniel Bélanger
Laurence J. Peter
Plume Latraverse
George Pólya
Thomas Kuhn
Edsger Dijkstra
Hermann Hesse
Abraham Maslow
Benjamin Franklin
Shiyali Ranganathan
Andrey Kolmogorov
Jiddu Krishnamurti
Georges Brassens
Bertrand Russell
Astor Piazzolla
Kurt Cobain

Google Search
Fagan Finder Blogs

Google Glossary

NEC ResearchIndex e-prints
SEP Bibliography
citebase search

Complexity Digest
Principia Cybernetica

All Consuming
gnod musicmap
Logical Fallacies
W3C Link Checker
Wayback Machine
RemindMe Service

Music streams:
Radio Tango Argentino
Boombastic Radio
Limbik Frequencies
Radio Paradise
shoutcast streams
electronic streams index


Subscribe with Bloglines

Click here to send an email to the editor of this weblog.



Tuesday, April 25, 2006
CHI 2006: Games & Performance session

Alone Together? Exploring the Social Dynamics of Massively Multiplayer Online Games is the title of a very interesting study of World of Warcraft (WoW) social dynamics by Nicolas Ducheneaut and collaborators. The talk was given in an overcrowded room. I was standing and couldn't take notes.

Basically they used WoW's open API to scrape data about hundreds of thousands of players on five servers, registering who was present where and when and what character they were playing. The paper is surely worth a look if you're interested in online social interaction patterns from data. If you dislike PDFs, a few of the findings seem to be in this blog post. Big kudos to the team for doing their research openly on the playon blog!

Next I listened to a talk about Interweaving Mobile Games With Everyday Life, which dealt with an urban multiplayer PocketPC game called Feed the Yoshi. It gave me some game ideas for the Ile Sans Fil community wireless project. Something to discuss with Michael... Transcript follows.

Presenter: We had two foci in this work:

  1. Weaving ubiquitous computing into 'the fabric of everyday life". What happens if you use a system over a long time?
  2. Seams and seamful design. Seams are gaps and 'losses in translation' in digital media. We actually argue for seamful design.

WiFi varies in position, range and access controls. There are gaps, overlaps, passwords, fees for commercial hotspots, legal constraints. Those are Seams.

We made a game that exploits seamful design. It's implemented on PDAs (HP iPAQs) so you can play everywhere. The game is called Feeding Yoshi. A Yoshi is a critter that eats fruit. Players feed them for points. You can also sow seeds at empty plantations. Fruits grow in them, and you can pick them up and put them on your basket. Yoshis and plantations are scattered across the city. Players carry fruit to the Yoshis. Yoshis and plantations are 802.11 access points. Yoshis are actually locatedat secure access points and plantations are at open points.

Here's a map of Glasgow, we did some wardriving and found 483 points. You get more points for feeding Yoshis multiple fruits. You can swap fruits with other players.

Points are submitted to the game web site via 'codes'. The site shows a leaderboard. The game uses p2p ad-hoc networks.

To study Yoshi, we had 4 teams of 4 in Derby, Nottingham and Glasgowplay it over 7 days. Players had various backgrounds. Our data comes fromdiaries, interviews, logs.

(Shows a video diary of a player beginning his day, leaving for work, finding a strawberry tree.)

Overall the players found the game fun to play, engaging, worth taking time out for.

Patterns of play reflected life/work styles. There was a large spread in the length of play sessions. Journeys were often good for play. Players built an understanding of where the Yoshis and plantations were on their routes to work. One player came up with the concept of the 'Drive-by Yoshi', wherein a friend would drive himaround and he could rack up points efficiently.

Work was actually both a resource and a constraint. You could use work's WiFi. People's jobs sometimes allowed them to take breaks, be late, slip out. Work habits were not a predictor of success.

The coupling with location led to awareness of urban character and conversation with other players. On crowded streets players would "run into folks". There were distinctive patterns of movement - shuttling back and forth, etc. People felt uneasy just hanging around suburban homes, and uncomfortable in industrial and business districts. Some areas were felt "too dangerous to play in".

The social setting affects coordination and collaboration, both with other players and with non-players. One player's movement patterns annoyed his girfriend greatly. Play eventually bridged teams, even though members of the different teams didn't know one another.

Reflections. Should players be forced to move out of rich areas? It might encourage mobility. Were some locations 'too good'? Should we support play 'at speed'?

Future work involves seams in software and eHealth.

This was the first study of a long-term mobile game. The quality of play was flexible with everyday life. Players augmented existing routines and established new ones.

If you have any questions I'm sure my coauthors will be really happy to answer them. :)

Q. You didn't actually require transmission over the hotspots? So you could have tied this to arbitrary locations.

A. We wanted people to actually learn something about wireless networks.

Q. This work is a great opportunity for exploration. Glasgow gets a lot of rain. Have you thought about extra rewards for people who go out and feed the Yoshis in the pouring rain? Have you thought about introducing a predator?

A. They're cool ideas. The first I'm sure we could implement using weather reports. If the game had more characters, some of which are unveiled over time, it might make the game more interesting.

Q. Tell us about how much people modified their normal routine?

A. At the end most people told us they actually altered their routine.

Q. How long does it take for plantations to grow?

A. It's immediate.

Q. Have you considered making them more tamagochi-like, where you can have your yoshis grow and evolve

A. We did. (Said some more things.)

Q. What are the system requirements?

A. You can download the game for any PocketPC from the website. We'vegot thousands of downloads, it's getting quite popular.

Q. (Google guy). This looks like a great way of providing incentives for people to go to certain locations, open their wifi networks, etc.

A. Well, there's probably a business in there. Thank you!

What do you think? []  links to this post    2:35:10 PM  
CHI 2006: Dogear, Social Bookmarking for the Enterprise

(Notes from a talk about IBM's Dogear.)

Social bookmarking is a central store, where you can put keywords (tags), discover via "pivot" browsing, and subscribe to link feeds.

Going to the enterprise: (1) authentication (no anonymity - promotes more responsible work), (2) internet and intranet bookmarks, (3) support for both shared and private bookmarks (this was a subject of debate among us), (4) designed for remixing (REST and dogear api)

Interface: select text, right-click and pick "Dogear this" or click the dogear bookmarklet. Window pops up, with Title - Tags - Description - URL - Private? checkbox (default is public so there is a little extra cost in going private). Below, recommended, popular, and recent tags for this URL, and a visual indicator of how popular the link is

All Users' Bookmarks display page has a different styling than personal lists.

Tweak in the enterprise seach engine (w3): show Dogear results before the w3 results (looks like corporate intranet search is not so good and they wanted to let the poor searchers help one another).

REST style: substitute the /html for /atom or /js in URLs.

Reuse: in the Dogear developer's blog, through a few lines of Javascript, his latest Dogears are listed.

Field trial results. Friendly trial began in March. In July: IBM Technology Adoption Program (TAP) launch, dogear included. Out of 686 visitors, 185 created bookmarks, 350 clicked on a link in the first 8 weeks. Sustained growth over 30 weeks.

Content makeup: 56% is shared internet, 38% is shared intranet, 2% private intranet, 3% private intranet bookmarks.

Early survey results show benefits.

Good buzz in the IBM intranet blogosphere: 94 unique posters mentioning dogear. When a new feature or mashup shows up there is renewed buzz. There's a whole new ecosystem around adoption of tools in the networked enterprise.

Tags reveal groups of people interested in similar topics. There are still around 10% new tags every week.

Next steps are to investigate social navigation through pivot browsing, folksonomy development, role-based portals, integration with other software.

Q. Did you make an effort to reduce number of tags by providing recommendations? A. There is still debate over whether this would be a good thing.

Q. Did users continue to use their own browser bookmarks? Do you have data? A. Some folks are interested in putting their dogear bookmarks out on the internet. I'm one of those who stopped using browser bookmarks, I'm sure there are others.

Q. Sharing? A. There were some comments about the benefits of using other people's bookmarks, and also about the . It's really low-cost sharing for the organization.

Q. My team has used a common tag that was used by all on the team. Other ideas? A. Once you see tagging existing both on bookmarking systems, blogs, etc., you start asking about a tag repository in general.

(Had a nice chat with Dogear programmer Jonathan Feinberg after the session, where we talked about the next step of enabling ridiculously easy group-forming in Dogear, among other things.)


What do you think? []  links to this post    8:31:42 AM  
CHI 2006 course: Jared Spool, "Designing for the Scent of Information"

(Jared Spool has been working on usability for eons. It's the second time I've run into him in Montreal. Entertaining speaker. I wrote down the points that stood out for me.)

is what users are looking for. Scent should be reinforced as you get closer to what you're looking for.  The three-click rule - "everything must be reachable in three clicks" is bunk. On most ecommerce sites you need four clicks to do anything. Users tolerate clicking more if they get reinforcement along the way.

Iceberg effect: The user assumes that what is above the fold is representative of what is below, so won't scroll if they don't find what they're looking for above.

Banner blindness: Based on experience, we are trained to ignore the top 60 pixels of a page.

People would rather click than search. Amazon is an exception, they have trained us not to expect scent on the home page. Amazon's nav panel is scentless.

Content created from the provider's perspective is a scent-killer. If your links refer to your specialized vocabulary but your users do not master it, they won't get where they need to go.

Trigger words lead users to click when they recognize them. If they don't see a trigger word, they will type it into your search engine. To find trigger words you're missing on your site, you turn on your search logs and you look at what users input.

Short links don't emit scent. The best links are 7-12 words. Too few words, you probably won't hit a trigger word. Too many, the user just can't see it in all that text.

"Keep your pages short, because you don't want users to scroll". Untrue. There's only so much that you can put on one page. The most successful pages are very long, about 10 screens long. NYT has just gone to a 3,000 pixel high home page. 

There are designs that stop users from scrolling. A horizontal rule; tiny print; side-by-side columns of text that end exactly at the same point.

Sometimes the site map actually gives off more scent than the home page. Unfortunately the link to the site map is itself scentless. If your logs tell you a lot of people are going to your site map, try making it the home page for a day. Chances are you'll get rave compliments on your redesign!

Some scents throw you off, violating your expectations, actually sending you away from where you're going.

There are three types of graphics on web sites: navigation graphics, content graphics, ornamental graphics. Graphics can sometimes communicate scent, too.

What do you think? []  links to this post    8:16:11 AM  

Click here to visit the Radio UserLand website. Copyleft 2006 Sebastien Paquet.
Last update: 7/3/2006; 10:25:36 AM.
This theme is based on the SoundWaves (blue) Manila theme.

April 2006
Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
Jan   Jul

Syndicated content: