Last Build Date: Thu, 26 Oct 2006 15:09:48 GMTCopyright: Copyright 2006 Sebastien Paquet
Thu, 26 Oct 2006 14:54:38 GMTI gave an interview to Loïc Le Meur right after Webcom Montreal. I know it had been a long day, but still it doesn't explain why my eyeballs are largely missing from that video podcast. We talked about my experience as part of a distributed company. Listen and see how well you can handle spoken evil dead Canadian French!
Thu, 26 Oct 2006 14:31:58 GMTZero-Knowledge Systems is a Montreal-based company that went through the dotcom boom and bust, and rebounded as Radialpoint. ZKS cofounder Austin Hill (who just started blogging - welcome, Austin!!) gave a 15-minute presentation at BarCamp Montreal on Saturday. This is my impressionistic transcript of his talk.This is normally a much longer talk. I was one of the cofounders of ZKS. I want to talk about entrepreneurialism. On my title's meaning. Not "chasing while clueless". It's that you just don't know how it will end up. It's really about finding jobs you love.People chase their dreams for various reasons. The guy who invented microfinance did not do it to get rich. I've been doing startups since I was 13, (quickly enumerates) total.net.... Around 1997 I got into privacy. I had a passion for cryptography, info security, the internet was becoming very popular for normal people. My brother, father and I started this enterprise, ZKS. The vision was empowering people to control their information. It wasn't about attacking cookies. We had a vision for remaking the internet. In the Wayback machine you can see our old mission statements. Funny, I don't remember writing this. We wanted to make it absolutely easy for people to go online and keep their privacy. There was a lot of very complicated stuff behind the scenes. Now. You have a big dream. You may think VCs are living in treehouses, throwing money out the window. I have learned how VC works. I researched everything, and sent out all these faxes.An important thing is to set your sights well. We were thinking very very big. Spreading our vision, with a lot of passion and evangelism, which helped drive our early success. Fundraising. 1M seed from angels, 5M from California VCs in '98, including the cofounder of inktomi. Series A 25M Yorkton in '99, Series B (30?)M in '01.The thing is, we did not need to raise money. But money was falling from the trees. People were desperately trying to give us money. We came up with a structure which was a coupon that said, you can give this to us now, you get money when we go IPO.You want to stage your investments. You don't want to raise your money at a high valuation with a dumb investor early on. We had an order for about 100M US, we only needed 10M. If you get money, you now have pressure to grow. We grew from 50 to 200 people. Has anybody heard of our hiring tactics. (audience member): I remember a ZKS-branded "We're Hiring" truck that would go and park right in front of other software companies' offices.There was a CEO in town, she tried to block email to ZKS. But her employees emailed us that night from home. Then we raised 22M as a Series C. All told we raised 80M$. US dollars.Thankfully we didn't put it all into stupid things. We diversified. We had this grand vision that we were to protect all the world's privacy.There are lots of types of money men. The believers. They're enthusiastic. Finding the right ones is critical. They can bring a lot to your company. The people who bring underwriting. A company went public too quick. They should have raised the 1M$ privately. A team of good VCs can be an ally. Then the gangster types, who want to take control of your company. Now. Rule #1 of startups. Bad Shit Happens. For us that was, the VC industry changed. [... ] You end up having to analyze your company to figure out what works. When you're big you can't iterate quickly. You talk about turning off entire business units. There's the edifice complex. When you walk into a new building, a lot of people feel like they've made it. There's the ghost effect when downsizing. For employees, this is bad for morale - the person they made come in all of a sudden has lost their jobs. Check your motives. If you're in this for the money, then go find another industry. When the shit hits the fan, there are easier ways. Starting a company is because there's no other way to chase your dream. Startups are a team sport. This is not swimming. Pick your team members very carefully. What really saved our company [...]
Mon, 02 Oct 2006 02:50:08 GMTTurning thirty. Time to get serious!
Mon, 03 Jul 2006 17:25:18 GMT(So maybe my 2001-2002 dissertation work was going in a promising direction after all...)
Thu, 27 Apr 2006 21:14:16 GMTDan Cosley (I think): Here are the takeaways I hope you will get from my talk: Artifacts of lasting value matter.Designing for contribution is interesting.Modeling value over time is useful.Intelligent task routing works. Communities like IMDB, RateYourMusic (35,000 people), and Wikipedia are building collective artifacts of lasting value. But not all communities succeed. Nupedia, how many smaller groups have faltered. Grouplens is kinda failing because everyone is using a different format. Often, one enthusiastic guy does all this work of setting up the system, inviting users, maintaining, etc. You know, like Chad adds all the movies in MovieLens. But you ideally want many people to do work. The virtues of the many: Scale (SlashDot), Speed (Wikipedia repairs < 3min), Robustness. But... users say, "I don't want to add movies, I want them to be there for me." CommunityLab is researchers from CMU, UMich and UMinn. We contribute theory insights and design ideas. I want to talk about creating value in a specific community, the MovieLens community. When we started this project we had 8800 movies x 8 fields = 70,000 fields to fill, 23,000 of which were empty. How to do it? Part I of my talk is about contribution reviews. We know that editing improves quality. Very often, in an editing process, you don't see the internals. You don't see how the sausage is made - it's pre-publication review. Wiki-like processes let people publish right away. We wondered which of those models (let people provide information and add a reviewing step, or not) would help the fields fill up faster. We tested it empirically. As it turns out, wiki-like beats pre-publication review short term, and in both cases contributions taper over time. In the wiki model, quality hits an equilibrium, a steady state where the good elements and bad elements balance each other. Our model says that in the long run, it actually doesn't matter if review is pre- or post-publication - you reach the same state. Part II of my talk is about intelligent task routing. How do people find tasks to perform in the system? Randomly (e.g. Slashdot metamoderation), chronologically, alphabetically? These kinda suck. You want to help people find work they want to do. People often work on their interests - match people with tasks they'll like. Karay & Williams' collective effort model deals with social loafing. They posit that people decide whether to contribute according to how much it benefits them, the group, and whether the contribution matters to the group. We assigned people to four groups; in each group people were given tasks according to different algorithms. We found that four times as many contributions were made into MovieLens if we asked them to fill in information about movies they were among the few to have seen. "This Needs Work" did not work so well as a motivator in terms of number of movies, but if you count the fields filled it's pretty good. Based on this work I made SuggestBot, which suggests Wikipedia articles you might like to edit, and is intended to optimize participant contributions. Q. (Alain Désilets, NRC) You asked people to edit movies they had not seen? A. Yes, 3 of the 4 groups didn't depend on whether they had seen the movie.Q. (me) So your model predicted that eventually the good slows down and the bad balances it out - does that mean that eventually there are effectively no contributions being made to the system? Do you assume that the amount of uncharted territory is constant?A. (The answers confused me. I think Dan basically said "yes, sorta" to both questions.)[...]
Wed, 26 Apr 2006 13:13:40 GMTEdward Bosma: My company, Océ, is a kind of Xerox for the Netherlands. We have >24,000 employees in 80 countries.
At Océ we know how to design for use (fit to the task, make it easy to learn and efficient to use)... but this does not guarantee product success.
Sometimes the product is not installed correctly, does not fit with how people work, etc.
How do we design, not for use, but for experience?
We see experience with a product as a process. Identifying phases in this process can help to understand experience. We identified these phases, partly based on communication science and experience:
We looked at our past mistakes. We postulated that all preceding phases must be passed successfully to reach a given phase.
Design for exposure: ensure your product will become available and fully operational for the intended users. When we introduced a digital multifunction unit that would print as well as copy but looked almost identical to the analog, copy-only unit, we realized that people were still printing on the inferior, small desktop printer. We did a redesign that made people more aware of the product's affordances.
Design for motivation: seduce intended users. Think about how a product can address current problems, basic human needs.
Design for orientation: We threw in a tutorial and some small tests for better retention.
Design for adoption: make users apply the product in their real-life situation. What can withhold them from doing so? We must raise trust and confidence that everything will go right, at each step along the way. For this printer (shows behemoth piece of machinery) we implemented an initial phase in which the machine will only print the next job when the operator confirms that the previous job was successful.
Design for incorporation: build a long-term relationship between product and user.
Q. How do you balance ease of learning with efficiency? Have you considered having different interfaces for different users? What about the idea of having a barcode that I can use to reinstate settings I use often.
A. Requirements may conflict; our paper says you have to strike a balance among all.
Comment. Alan Cooper suggested a sliding panel. I'd love to see a copier that has something like that - a fat copy button and a panel that hides/reveals the complexity.
Comment. (Xerox guy). Do you have a product planning department? My observation has been that sometimes it's a battle between the usability guys and the marketing folks who want the new features.
Q. How do you measure adoption rates for this kind of product?
A. We have no quantitative data. This is based on field experience, sitting down with customers.
Q. In your scenario, probably the buyers are different from the users.
A. Good question. We focus on the end users. Our goal is not that the product be sold, it's that it be used properly.
Tue, 25 Apr 2006 21:35:10 GMTAlone 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: Weaving ubiquitous computing into 'the fabric of everyday life". What happens if you use a system over a long time?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 ric[...]
Tue, 25 Apr 2006 15:31:42 GMT(Notes from a talk about IBM's Dogear.)
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.
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.)
Tue, 25 Apr 2006 15:16:11 GMT(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.)
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.
Mon, 24 Apr 2006 17:03:29 GMTThis year the CHI conference is experimenting with how the (impressively heavy, 8-track wide, and that's just the papers) day schedules are presented to attendees, and they came up with a refreshing idea. Every morning they will line up 42 speakers to present 21 hours' worth of scholarly content, each being alloted 40 seconds to attract listeners to their talk. The end product is packaged as the CHI Madness session.
I find the idea deeply clueful. This gives people an occasion to briefly pitch their findings to an audience on the scale of a thousand. Normally you're able to do this only with a handful of people, during hallway or reception conversations. and I expect it will help better match presenters with attendees and perhaps spread attendance a bit more evenly. I also think the exercise of distilling to a 30-second pitch will probably help improve the presentations themselves.
We just had the first such session. I feared that it would be tedious, but the pacing and variety make it just right. It's fascinating to see how different researchers condense and pitch their papers, both visually and aurally. Academics are often slow to get to the point, so this is an enlightening exercise. Some are more skillful than others. Most are pretty creative, so the few Powerpoint lists filled with bullets and people who read their abstracts look a bit stodgy.
One researcher hilariously pitched his paper as an 8-round wrestling match between - wait for it - documentation methodologies. The Phlat guy closed with the slogan, "It's Phlat, it' phat, yo." One guy's temporal compression work was ideally suited to the medium, so his pitch was pretty effective.
Mon, 24 Apr 2006 15:30:08 GMTThe CHI conference is in my hometown this week, and Socialtext is graciously sending me downtown to learn some stuff. Intuit founder Scott Cook opened the conference. Here are my notes from his talk.[Update: what follows is more of an impressionistic transcript; Antonella Pavese put up a really nice "annotated reconstruction" of Scott Cook's CHI talk, complete with pictures and all - go see it!]Speaker intro (paraphrased): In 1984 Scott Cook founded Intuit, which makes the famously successful Quicken software and is among the first companies to have a public wiki, the Tax Almanac.Scott Cook: Je suis heureux d'etre ici aujourd'hui. Mon francais... not so good. In preparing for this, my wife told me not to worry about being charming, clever or intellectual. "Just be yourself", she said. One of the best examples of game-changing invention hails from here in Quebec. Guy Laliberte founded Cirque du Soleil 25 years ago. It became a huge business. They did this in a market that anyone would have said was down. They were asking for more money than the Wringley Brothers were asking. They didn't steal Wringley's customers. They invented a new kind of entertainment. 3M used to sell sandpaper. Employee Dick Drew was in an automobile shop, heard a worker complain that he couldn't remove the tape without taking the fresh paint off, too. Then he'd need to sand a lot. This led him to invent the masking tape (even though his boss didn't want him to), and later on, the cellophane tape, which turned 3M into a huge business. Said Peter Drucker: "The bottleneck is always at the top of the bottle". What kind of company are you building? That depens on your model of innovation. Here are five: (1) the lone genius (the Edison / Bell model); (2) the boss is the genius (popular among some Silicon Valley CEOs); (3) copy competitors' inventions; (4) Cloister the geniuses in a lab - but the innovation never seems to make it into a product; (5) make your people the geniuses. The model I prefer is the last one. The people who are in close contact with customers are the unit of innovation. Managers' jobs are to create the greenhouse, enable those little teams to form around customer needs. Dilbert slides in which a VP of marketing gets punched and spit on elicits general laughter. Look at the evolution of shipping. By 1960 this industry had reached an apex with the NS Savannah, nuclear-powered cargo ship. Malcolm McLean was a trucker who observed tremendous amounts of delay in transferring merchandise: unloading and loading to and from ships. In the mid-50s he sold his truck company. He bought a rusty old tanker, the SS Maxton, which he put a flat bed on so he could put rectagular containers. SS's maiden voyage was 50 years ago last week. Containerized freight revolutionized shipping. The cost used to be $6 a ton. Containerized freight costs 16 cents a ton. There are today millions of people in Asia who owe their middle-class status to Mr. McLean. He didn't focus on the speed of the ship, he focused on speed of loading and unloading. Anyone with a science background should read Kuhn. It explains how some young scientist, generally under 30, invents a new paradigm that nobody believes in at first but is eventually embraced by all. Germ theory, Copernicus, etc. That's science. Intuit's mission is "to change lives so profoundly that people cannot imagine going back to the old way". Here's the story of how we moved from Quicken to QuickBooks. We had less than half the features and were selling at twice the price of the market leaders. The advertising was terrible. Our first ad, I wrote. It was a catastrophe. Second was an ad firm, which produced probably the worst ad in history. Exposed to a million readers. We got 4 responses. That's the brilliant launch of QuickBooks in 92. Her[...]