Back home, finally. I enjoy the conferences, but a week is just too much for me. I fall apart without Jess.

Back home in time for Sidewark Arts, too.

Here's J looking at some piggies:

And here's the Best In Show:

And even the Stop The Lie guy came out for some good old fashioned Bible Thumpin':

Here are some pictures from the CHI2006 conference.

[CHI2006 - Unedited Notes]

Scott McCloud
http://www.scottmccloud.com

Been writing about comics for 12, 13 years; in the last few years, been teaching comics. There have been a lot of mutations in comics since 1982; lots of different directions for comics to move. You need to have a large footprint to teach comics successfully. Finding the common denominator is a challenge when teaching things relevant to all of these issues. There is a common vocabulary of words, symbols, and they tell stories with words and images, in a sequence. But it was important to teach comics as a series of a choice.

Choice of moment; which moments will represent a narrative; which should you put in
Choice of frame; when to zoom in or out
Choice of image; what will represent the moment and fill the frame
Choice of word, how words and pictures are integrated;
Choice of flow; how panels are arranged, and how we guide the readers eye through narratives and in panels.

In examining all of these issues, there are few times it is necessary to specify web comics or real comics. Maybe they aren't all that different, when you get back down to the basic principles of comics. All of these follow the same DNA of creativity, which is common across all types. The simple pattern of what to see within the panels, and the reader coming back to imagine what happens between the panels.

That creates an approach to comics that creates a temporal map. This is similar to how a chimpanzee and humans have 96% the same genetic makeup, its true because it says so on the internet. The four percent, however, is enormously important.

This came up when the web was first evolving in the early ninties, when the CDRom and multimedia was beginning to evolve, people were trying new things with comics. How would comics be evolving? Was looking for a durable mutation, which could survive and thrive into the next era. A good phrase - exactly what "the newspaper industry is looking for".

By looking at the early multimedia attempts at comics, you see an attempt to take the visual medium of comics and combine it with things similar to multimedia - sound, motion, interactivity, and cram it all into a frame on the screen. The new media would appropriate the shape of the old media as content. McLuhan. The upright rectangular echo of print would introduce sound and motion, and then you begin to see that, over time, you are interrupting the seamlessness of the presentation. Comics use a temporal map of moving through space and time. This creates a transparency that allows the medium to vanish. By introducing temporal phenomenon that represent themselves in time, you move from one mode to another and the medium becomes visible. People would become aware, at all times, of the medium - and the medium was unable to vanish long enough to create a transparent story telling image.

Comics are an interactive experience. After Understanding Comics came out, Mosaic came along too and people began to try out comics in that environment. They would utilize choose-your-own-path in a hypertext way, in that it is non-spatial. It is here, not here, or connected to here. In comics, every element of the work has a spatial element to every other element at all times.

As a user, one is asked to move from the world of the author back to the world of the user continually. Games are "user-authored" - people do things. Let me tell you what I did. The author, however, sits back and lets things flow over them. Pitching it in the middle creates a bizarre tension between the user and author.

It may be helpful to think of this as a hill connecting users and authors; multimedia narratives in the middle makes things roll down the hill in one direction or the other. Multimedia at its best puts itself at the top of the hill in a birds-eye view, but you are forever aware of what you are looking at.

So we create a challenge; we need to remain consistent to comics' essential nature, while still taking advantage of digital media. The best way was in keeping an eye on the basic DNA of comics - the temporal map. There are protocomics, or precursors to comics, such as certain Egyptian wall paintings. The basic idea of this artform interacted with other technologies, such as paint on stone, relief column, or tapestries. These are cases where, moving through space, you are moving through time.

As soon as print came along, one begins to see examples where the idea of comics comes in contact with the technology, and you end up with square panel borders, the left to right reading sequence. Within 100 years, you see a "true narrative of the horrid hellish popish plot". This leads to the same comics we have today in bound printed media.

If comics pre-date print, if the printed version affected this in such a profound way, then what comes next?

The pre-print examples have an unbroken reading line in a simple, seamless fashion. Single mode navigation. Move through space until you are done. Print makes things change; adjacent moments are no longer adjacent pages. The break was replicated over the years.

A monitor is no more limitless as a canvas than is a page; how is this an opportunity for the content to grow? You can view the monitor as a window. Stories can move in x and y direction, and can take advantage of the various movements of the screen.

But in the early ninties, we are surfing at 14.4 and they aren't ready for giant scrolling abstract poems. A durable mutation, but choice of words was telling. The first google hit was McCloud - this isn't a phrase that biologists use. They use "A successful mutation". Mutation is a tool - durable implies a tool. The mutation does not have a mind of its own, but another mind can plan for and use it. Successful mutation implies that the mutation itself has its own ideas and is "out there". No one is driving the bus.

Comics on the web evolved. Comic strips found an easy place on the web that fit and worked; many communities began to show up on the web. [Copper - Sail] People began to understand the media in a new way, such as full painted color; there was also an evolution of genres. Different styles, depth cues, scanning in traditional content; 3D modeling was used, as well as borderless and color. And then the same visual techniques would show up in print as well, so we see the web actually driving the print innovations. Inertia was plowed aside, and that was a natural evolution of the ideas.

But meanwhile, what do you do about the page and the longform comics, with graphic movement? [nowhere girl] These things had to be reinvented, and did not anticipate changes.

As a results of upbringing, and liking space and minerals and biology, McCloud ends up as the nerd genius creating art. Classicist, Animist, Formalist, Iconoclast; mimics Thinking, Sensation, Intuition, and Feeling.

Form/Content, Beauty/Truth

Those who wanted to make form visible, and those who wanted to make form invisible was being played out on the web. The tinkerers wanted to see the form and the medium and experiment with it and make people aware of it, while others simply wanted to see organic evolution of the idea - they cared most about story.

Format is important, because comics may not always be received through the medium they are not. Do the ideas scale? Need to look at strange inventions with infinite possibilities, to create something that looks nothing like something anyone has ever seen before, which is a mutation.

We need to be ready for changes in format, because standards may constrict the evolution.

Frequency. A tradition of putting things in rectangles, and varying the size of panels but not their frequency. It's a crime to waste paper; but when there is no premium on blank space, there is no reason not to vary the space between panels. This takes years for people to understand that they've left the cage and are beginning to spread their wings.

We vary between wanting to lean back and let it happen, or control the experience. But most of the design models follow the notion of seamlessness. Conspicuous to let someone play with the construction to understand how it works.

Moving through space is equivalent to moving through time. This is seamless and simple and allows for transmitting content in a beautiful way.

"I want comics to scale because I don't want comics to go away". A movie is a comic's exit route; it can't take on the appearance of another art form. It takes a look of itself, and becomes more like comics as it grows away from other ideas. While we converge in means of distribution, there is room for an aesthetic divergence of shape. A durable mutation.

[CHI2006 - Unedited Notes]

Scaling the Card Sort Method to Over 500 items: Restructuring the Google Adwords Help Center

Yelena Nakhimovsky

Adwords is one of the main sources of revenue for Google. Adwords show up on the right side of the page and are supposed to be relevant to the search. The people who write the advertisements see a different interface, the campaign management interface. The help center provides them with help.

The problem: the help page is huge, and has an overwhelming amount of links. There are over 400+ items, it's a complex product, it has a lot of new features, and the help content was beginning to outgrow the existing structure.

The goal was to make it easier for new users to browse, and to encourage all people to use the help center for common and straight forward questions.

The team was going to use a traditional card sorting approach - placing many items on cards, and having end users sort this material in a way that makes sense to them. This works successfully when the content is strong but the navigational problem is difficult. This doesn't work with close to 400 items. The content was very fine-grained and specialized, in many cases, that a subset card sort would not work either. So what to do?

Method had four stages. The initial research led to a first and second card sort, and finally to an experiment. The goal was to reduce the number of cards at each stage.

The initial research involved accumulating background research from the search box within adworks, what were the most common queries, and what were zero-result queries? Additionally, gathered information from the phone calls in the support center. This provided a look at the verbage used by the actual end users; for example, "How many AdWord campaigns can you have in a campaign?"

Also conducted interviews with the customer support reps, to find out when they refer people to the help center. They were pointing people to the demos and guides in the "learning center", which was outside of the "help center". So the set of cards grew to 500 items.

The first card sort involved three user experience researchers and an adwords expert in the room together. The goal was to reduce the quantity of words that were being used. They included a pile called "unsorted" to allow for elements that are unclear. The adwords expert was there to provide clarity to the vocabulary.

These categories were called "proposed categories" and were put in a separate group, and were taken into a separate card sort. Three customer support representatives were recruited to sort these new piles. These sorts were independent, and the main goal was to use the daily and direct contact with the users and apply it to the sort.

USort and EZCalc allow for the creation of various visualizations of card sorting. It doesn't work with over 100 items. A spreadsheet template was created by J. Lamantia, which allowed for aggregate data, but didn’t allow for the separation of one user to another. Ended up creating a set of linked spreadsheets illustrating all findings.

Finally, created an experiment to understand how AdWords users actually dealt with the results of this card sort. 12 existing AdWords users participated with various levels of familiarity. Two tasks per person.

The metrics tracked included time to task completion, the error rate, and the "give up" rate. The new help center was faster to use and had less errors. This new help center is now live.

Scaled card sort method resulted in an improved information architecture.

[CHI2006 - Unedited Notes]

Understanding Users in Consumer Electronics Experience Design
Samsung Electronics

Joonhwan Jim

Gathering data from users is not enough. Finding the meaningful information "meaningful factors" and converting it to the design is the critical part. User research, in this material, is defined as "all interactions between designers and users, for the purpose of improving the user experience quality". This is balanced with requirements from marketing, quality control, and other development issues that impact the final product.
Background on the study: deals with interactive television, digital television, and a/v productions. Professionals from user experience design, usability and graphic design participated in the research.
Conducting user research does not provide motivation for design; so what are appropriate methods to understand the "meaningful factors"? And, how do you conduct these methods with limited time and resources?
Three stages of understanding the experience design process: understanding user requirements, establishing a strategy, conducting evaluation.
After conducting user research, we often get irrelevant data. Understanding user's requirements involve learning what people want through in-home interviews. A digital television innovation project, through collaboration with a university in England, attempted to study people in their homes. Ethnography allows for coming up with new product ideas and understanding how technology might fit into the overall context. Technical Biography (blithe and Monk) provides several ways to accomplish this: recommends looking at past development and current trends to identify user concerns in domestic activities.
1. Personal History. Participants are asked to describe the experience of use and changes in domestic technologies that subjects had witnessed during their lifetime.
2. Technology Tours (Baillie and Peterson). The participant shows their home and questions are asked about specific technologies.
3. Last Time Questions. Based on Critical Incidents by Flanagan; "When was the last time "
Additional techniques are adapted from the Cultural Probe (Gaver, et al) techniques.
1. Three Wishes. Note pads were left with subjects with "I wish I had" printed on them.
2. What We Watched Last Night. Retrospective protocols were created based around a VCR recording of the television screen from the previous nights.
Looked at many houses, and conducted two visits to each home. Findings included:
36 television use patterns in seven categories, and 5 new features were suggested. Findings related to the methods: these techniques were useful in revealing users' hidden requirements. More meaningful data that was relevant to the topic was gathered, and it took less time.

So are subjective design preferences qualitative or quantitative measurements? It's both. We can establish design direction and strategy by adapting the methodologies to objectively measure subjective qualitative data. The goal was to establish a direction and strategy for a graphical user interface for Samsung televisions. Applied two question topics and a decision making method to make lucid measurements of aesthetic preferences.
First impressions were a measurement of a user's degree of preference upon their first encounter with the conceptual design. Conceptual suitability measures the level of a user's concurrence with the factors of the conceptual design. Analytical hierarchy Process (AHP) was applied for relative comparison ensured the objective validity of the answers.
From three country tests, found that they could discriminate between superiority and inferiority among the prepared design concepts. The proposed two question topics were advantageous with regard to time, and easy to conduct. AHP overcame the limitations of existing quantitative methods that rely on statistical analysis.
Analyzing video tape in usability evaluation testing takes a great deal of time. How can you skip it? A high fidelity prototype was used in a simulated test, and used a "check paper", a paper containing a pre-determined correct path for each task: Action Purpose / Correct Path / Participants Path / Observer's Notes. If the subject performed an action outside of the correct path, the observer makes a note. After the task is complete, the observer leads a discussion based on the incorrect actions. Discussion is made when the subject still remembers what they have done; this reduces analysis time. There is less bias in the time-on-task measurement because less talk is require during the task (no thinking aloud).

[CHI2006 - Unedited Notes]

Developing User Interface Guidelines for DVD Menus
Karin Kappel, Martin Tomitsch

A DVD menu is an interface to a movie. Users experience it like a bonus to the movie, and hope that it gives them a good experience. They expect that it is easy to operate, and they expect their skills to transfer.

The manufacturer views it differently; they see it as selecting different content elements and navigating content elements. They also provide the ability to change settings, display copyright notices, and provide advertisement for other content.

Research Focus: How can we implement a usable DVD menu?
"The failures of web design recreated yet again" - Don Norman. Frustrating experience for users navigating through DVD menus. 75% of all US households have a DVD player, which is a huge amount. DVD menus differ from desktop interfaces, in that they require a remote control. The output device is different, too - a television, with low resolution.

The process: literature review, expert knowledge and user studies, along with card sorting, technical feasibility, and usability testing led to the development of guidelines for the creation of usable DVD players.

The literature review showed no scientific writing, but a great deal of content relating to phone menus and interactive television. There were some articles and blog entries, and many entries and comments relating to frustrating experiences. "Whoever designed the menu system on these DVDs should find another job!", a quote from Amazon.com.

The expert walkthrough looked at 70 movie DVDs. The result found many usability issues, and discovered best practices. The main issue is the lack of consistency and standards. Additionally, there is a lack of mapping between the remote control and the selection technique.

Usability testing; 20 participants, 10 DVDs, and 5 typical tasks for DVD usage. Tests were conducted in the home. Self-declared expert users had no problems, but there were major problems for intermediate and novice users. People appreciate seeing the movie sequence that they can map to the movie; having only text doesn’t allow people to see the appropriate scene. Scene selection is the most complicated aspect of the DVD menus. Another problem that arose was the sixty seconds of introduction that users were unable to skip. This includes the copyright, the company intro, the movie intro, and frequently, the previews.

Asked a survey with specific questions. 350 respondents. Results were mostly best practices, but bonus material was the selling point for many people in electing to purchase a DVD (but they don’t watch it!)

Created a set of 44 guidelines: http://deco.inso.tuwien.ac.at/dvd.html

After creating the guidelines, they tested them to see if they were appropriate. Used a card sort to group the categories into eight categories; also gave them to a DVD menu designer, who created a prototype using the guidelines. Some features are difficult to implement: these include technical constraints, presenting the current status of a selection, and providing audio feedback. The number of menus to put on a DVD is limited.

Usability tested the prototype, and found the same test setting as the gathering phase. The results were "extremely satisfying"; no major problems, and allowed for the guidelines to be revised. Then, allowed the guidelines to be heuristically evaluated through ten student teams. This allowed for improvements of the guidelines through the inclusion of examples and better document navigation.

Why are DVDs unusable? They are bought first and then used, which is opposite from websites. Additionally, there are technical constraints. There is cheap production, and many people have fancy ideas.

Back from a three hour dinner at Fondue Mentale with Bill and Dave; ok food, good conversation. I'm beat. Bed time.

[CHI2006 - Unedited Notes]

Does Think Aloud Work? How do we Know?
Elisabeth Cuddihy
Laboratory of Usability, University of Washington

Trying to validate usability testing mechanisms and creating technology that can assist and augment in usability testing.

Few standards exist for conducting tests or for analyzing the gathered data. Turn to the literature to see what has been said and what work has been done, and we find a number of papers that compare different methods to each other. There aren't, however, a lot of work that look at the underlying theory to understand if the methods are measuring what we think they are measuring. To what extent is Think Aloud Protocol biasing or changing people's behavior, and to what extent does the data represent actual behavior in real life? How good is it at locating actual usability problems?

Elisabeth's research investigates, through methods like eye tracking, if the verbal protocol matches the physical protocol – are the people saying what they are really thinking?

Research challenge presents work to be investigated. Rigorously looking at how Think Aloud affects performance and behavior, and how large cognitive load changes the results of these tests.

What are the best prompts to keep people talking during Think Aloud?

Acknowledgement tokens, such as "uh huh", "ok", etc can get people restarted. If this is not enough, you may ask an open-ended question such as "and now you are …" Unfortunately, "Please keep talking", as specked by Herb Simon, would actually fluster people. Participants would apologize. So these acknowledgement tokens may actually prove to be more useful than the traditional vocabulary. However, the acknowledgement tokens should be used sparsely.

If, during a practice Think Aloud session, the participant doesn't "train well", what do you do?

Don't correct them; the goal is to create a connection, or some kind of bonding, in order to get the best sort of data. Don't try to be critical; don't explain what they could have done.

How do you get participants from merely describing their behavior to giving reasons for their actions? And, ideally, which sort do you want?

You can't trust people to tell you what they know – only what they are doing. It is the job of the designer to interpret and understand why people do what they do.

[CHI2006 - Unedited Notes]

TAP: Touch and Play
Park Duck Gun
Basics Research Laboratory

Will intra-body signaling be useful in future of ubiquitous computing? Intra-body signaling is communication between one hand to another hand; you detect something with the left hand and perform something with the right hand, using your body as a transfer mechanism. Ten year old technology.

Why use inter-body signaling when you can use wifi? Is wireless communication perfect? Bluetooth photo printing requires at least five steps. The number of devices waiting to serve in a given space is increasing as ubiquitous computing is on the rise. The users have a lot more to learn, which be can be a problem.

There are currently 18 ways to print a picture with a camera and a printer. Why are there so many ways, and why is so much learning required? Context is vague; the identify of the user, selection of device, selection of the service, data of interest - all must be delivered from user to computer. Context awareness computing - the applications do the right thing at the right time without direct manipulation. How can the computer know what the user wants?

Touch holds the answer.

We invented the remote control of the television in order to avoid touching it. But touching is familiar - it can provide context.

Key idea: the user can print the photo he was seeing the camera by touching the printer while holding the camera. The touching can deliver the information using intra-body signaling.

The conceptual model in Bluetooth is to transfer data from one device to another using wireless communication network. TAP is more direct. Select data; find the most likeable or logical device; evoke. Bluetooth must be learned; TAP also has to be learned, but has a more intuitive and lower-learning curve. Bluetooth requires typing; TAP requires touching.

This can be expanded to show a context-aware room, where each item knows what it is for when connected with another device via you.

[CHI2006 - Unedited Notes]

Embedded phenomenon: supporting science learning with classroom-sized distributed simulations
Tom Moher
Department of Computer science, University of Illinois at Chicago

http://www.evl.uic.edu/moher/

Many view the calculator as one of the most interesting things to be introduced into the classroom, but there have been a whole range of digital items embedded in classrooms – from palm pilot and on. While there has been this large growth of computing in devices, there has also been a large growth of computation embedded in the environment. This is a more ubiquitous manner that affords equity and "whole classness" where the technology is provided to the student by the school.

This is intended to explore scientific inquiry. Scientific facts, or declarative knowledge, are not the goals of science. Scientific inquiry is intended to understand the other forms of learning, such as conceptual learning or attitude issues towards science. Science involves processes that take a great deal of patience and happen over time.

Teachers try to make these sort of things available to students through the use of ant farms and other long-term phenomenon – activity centers within classes. Natural processes, however, don't always indicate the intended outcome of the course of study. Digital phenomenon have been attempted to allow multiple encounters over a long period of time. Environments can be set up and reused in order to gain insight.

Embedded phenomena seeks to extend the digitalization into the entire physical classroom in a way that relies on imagination rather than digital technological advancement.

Simulated phenomena are "mapped" onto the physical space of the classroom. Requires imagination. Then, a thin-layer of digitalization is added around the room through web browser based portals. The state of the simulation is represented through distributed media located around the classroom representing location-specific portals. The simulations are persistent and enduring – they run day and night for weeks and weeks, and they continue to run. Kids can tune in to see what is happening, but don't need to be present to cause phenomena to occur. Students conduct investigations of the concepts by monitoring and manipulation simulations.

Learning implies situated, embodied, social, episodic and opportunistic; the phenomena must occur with all five of these issues. But also, this provides an opportunity to introduce science into a classroom focused on mathematics or reading.

Participatory simulations are a powerful metaphor; the learners are part of the simulation, and participate as scientists studying what is going on around them. They participate in a virtual environment in an attempt to allow for authenticity. Using the physical and entire room, and the notion of ambient media – these are windows into the phenomenon. Mixed reality plays a role as well, as students must rely on their imagination to make sense of the events.

Roombugs. Imagining that the classroom is a small farming community; attract the bugs we want, and repel the bugs we don't want. In order to do this, tablets are laid flat and are considered sand traps. Students can understand the local bug population and use pesticides.

Helioroom. Students in third grade learn about the solar system. They then exercise this knowledge so they can solidify the material and understand it. The simulation says that the sun is in the center of the room, and tablets are attached to the walls. All planets are the same size, and given different colors. Occlusion was used to indicate the difference of front and back, and orbital periods were shown. Students had to determine which planet was which. Encourages transitive comparisons.

Roomquake. The classroom is the locus of a seismic field, which will experience several earthquakes over time. Students can read the seismographs, and earthquakes are signaled by a subwoofer making a deep sound. Students must understand the epicenter of the earthquake within the classroom and mark the placement. Over time, we can find the fault line.

[CHI2006 - Unedited Notes]

Beyond Record and Play
Backpacks: Tangible Modulators for Kinetic Behavior

Hayes Raffle, Hiroshi Ishii
MIT Media Laboratory

web video

Objective of the project was to allow children to experiment with sensors, feedback and motion parameters through use of a tangible interface which promotes direct manipulation and collaboration. Backpack creates modulate recordings created with a 3D construction assembly system.

Similar to central pattern generators - mechanical models dealing with similar patterns of motion. Stream processing has a precedence for people dealing with real-time modulation of data streams - analog mixers, for example. Building on tangible interfaces for learning. A form of digital manipulative would try to look at what people are already learning - what can be added through the use of computation?

Best known example is the Lego Mindstorms, which uses a GUI to indicate procedural programming which is then downloaded into the toy. These decouple the programming from the physical model making. Tangible interfaces try to overcome this by bringing the two together.

One approach is to make the physical structure look like the control structure; Duplo blocks with different functions allow for the exploration of digital and physical structures. The other approach is to use a record and play style to indicate a limited type of programming to offer limited behavior. The child can express desires and aesthetics in the model.

Backpacks build on Topobo system - a constructive assembly system. This is a set of building blocks that can be put together to move in a certain way. The Topobo has Passives, Actives, and the Backpack - physically passive but computationally active. A representation of the motion. Four kinds of backpacks. A time delay phase shift; a position offset (orientation); Faster Slower which changes the frequency of the motion; and Bigger Smaller, which allows for amplitude.

Modular and Composable - you put them on and they work right away. Default mode is local; a button is pushed to allow for a global modification.

Foundation idea was in augmented queens - explored waves, spirals and resonance. Very confusing as there was hidden state and it was difficult to understand what was going on. The backpacks tried to make these changes tangible. The technical implementation was based on sensor networks and amorphous computing models.

Users wanted to be able to sculpt the logic in the same way they could sculpt the physical form. Replaced the knob with a light sensor. Design process also incorporated ideas of feedback into the system. Knob changes state and motion.

Domains of Knowledge. The Time Delay Backpack which incorporates phase shift and wave propagation. Time delay backpack can be used to adjust time between legs when walking. Demonstrates both positive and negative phase shift. Illustrates that a wave is the same motion as a phase, just with different timing.

Harmonic resonance is indicated through faster-slower motion; slow movement can translate to a faster walking - resonance and balance between structure and motion.

Incorporating mechanical connectors to the knob allow it to become part of the structure, which indicates feedback.

Kids aged 6-15 were incorporated into the design kit creation. There were informal evaluation sessions used in home environments to understand if these were working. Backpacks were demonstrated, and then there was an afternoon of free play. Also, eight grade physics by design class was given homework to understand locomotion through the ideas. The backpacks that described ideas the kids were already aware of, faster/slower, bigger/smaller, were easy for the students to understand. Students had difficulty understanding the relationship between the light and the movement, when using the input states.

The kids who were most successful in constructing and understanding the tools were those that used an iterative design cycle.

The backpacks were an accessible interface and introduced concepts that ranged in complexity. All the children were able to use the backpacks; showing the children built creations which used the backpacks greatly accelerated children's conceptual understanding. From play to abstraction, backpacks are intended to raise the ceiling on learning. They introduce abstract ideas. Backpacks may support knowledge transfer in order to teach kinematics models.

The whole goal of the project is that coupling control and representation can support hands on design and experimentation of kinematic systems. Introducing parameterized control, sensors and feedback in a way that doesn't involve complicated programming. This indicates complex relationships like phase relations in a way that can include knowledge transfer.

It's now 8:20pm and I'm only starting to feel normal again. I think eating a big old plate of smoked meat helped a bunch - we went to Schwartz's, a bizzare mix of Seinfeld-style eating rules and communal dining.

Last night started with a few drinks and a strange cirque du soleil themed CHI opening. Although cool, I'm not sure the geeks were the right audience for a half naked man swinging from the rooftops. Who knows.

Chris, Mike [scroll down], Crow and one of his Toronto friends went out to dinner with Karen, who I've known online for a bunch of years but never met. We started at Kandybar. Our hostess was the waitress and the cook and the owner, and I'm sure we were the only party in the restaurant for the last few weeks. She was crazy - we suspected meth or speed. The food was lousy, the company was great, and I drank too much.

Then we headed to another bar, and I don't remember much after that. There was a hotel with a fireplace fountain, another bar with giant beers and close to fifty original Eames chairs, and another bar with shots of Jagermeister. What a terrible idea.

Fast forward to 2:00am, and somehow I ended up in my little hotel room snug in bed.

Here are some pictures of the fiasco that was last night.

[CHI2006 - Unedited Notes - alt.chi]

Tokyo Youth at Leisure: Towards the Design of New Media to support Leisure Planning and Practice

Anne Elliott

Tokyo is a dynamic environment with many leisure options and planning resources. There is a good public transport system which affords a great deal of moving around. Target user population was young adults, aged 18-25 - these people have more free time in the "golden years" between school and joining the work force.

Research questions pertaining to tools used to plan leisure outings; what resources do people use to plan leisure outings, and what is the structure of a typical leisure outing?

Methods used included research, ethnography, surveys, interviews, focus groups. Found that relaxation and companionship are the key leisure qualities for the target in Tokyo, while finding romance was the least important quality. People and television are the top resources for planning, while there was little use of the mobile phones in order to access the internet for planning. Communication is a leisure activity, and shopping was the most popular of these activities. Trends that emerged: relaxation and communication are key to understanding Tokyo. Most people have very full schedules and are very mobile. They sleep less than eight hours a night, and commute more than an hour. Downtime is essential.

Communication and Companionship are key leisure priorities - "hyperconnectivity" is a sense of constant connection. This was maintained via email and the mobile phone, and many people would have between 5 to 20 sessions of internet enabled communication.

Serendipitous, effortless discovery of things to try. People would use the internet to locate, but then the mobile phone to coordinate. There is primarily one planner and organizer, who then plans a meeting in a specific place. This is intended to minimize the commute and maximize the choices. The key to understanding if something is worthwhile is recommendations from trusted sources.

Found that space is crowded; there is a desire for personalization; recommendations are important.

[CHI2006 - Unedited Notes - alt.chi]

Playgrounds are a unique island of unstructured play in children’s over structured lives. Can digital technologies augment the playground experience without compromising existing play activities?

MIT Media Lab

When we think about augmenting traditional venues, how do we go about doing it? We don’t want to compromise the value of the tradition, only augment it.

Design process involved personal playground experiences, observations in two playgrounds in the neighborhood, drawing exercises with four year olds, and narrowing down the different ideas. A big theme was related to pathways and running.

Quick prototyping, using handy materials. A play-session at the preschool. Focused on motors, trying to hold them and use them. What can we do to make the children focus on something else? How can we include the children in the process, instead of bringing something into their world?

Adding paper spinners into the creations – participatory design within the kindergarten. Stepping on the steps make the spinner run. Started to see that the children invent their own games. The kids wanted to know if it worked with toy trucks, for instance. Kids realized they could customize their tools and work with each other, creating new ideas. Improvisation, fantasy, and advanced physical activities. Create a natural way to play with digital objects.

What did we learn?

Design should enhance existing play activities, not replace them. Keep play open-ended, children will invent their own games and rules. Design should support a variety of play-patterns. Limit features to promote imagination and invention: simplicity works.

[CHI2006 - Unedited Notes]

Share and Share Alike: Exploring the User Interface Affordances of File Sharing
Stephen Voida
Georgia Tech

It is presently common to exchange files via email; the use of attachments. Additionally, network file shares are used, as are corporate or personal websites. Collaborative websites are increasing in popularity, like wikis, and there are specific-media sites, such as flickr, intended for particular types of content. Corporate groups may use shared virtual desktops.

Common thread - there are essentials to each of these. What should be shared, with whom should it be shared, and finally - how should the sharing take place. The last question is interesting simply because of the amount of tools available. Users select the type of tool that is available based on affordances and features of each tool. Yet tools impose limitations on how file sharing can take place. It is difficult for a sender to manage access controls, especially if they change frequently. Additionally, certain features may dictate the choice of tool, even though other affordances may not be an idea match. An example of this is the use of email - even though the tool is hardly ideal for sharing files, it is essentially ubiquitous and people can be sure their file sharing wil be successful. Also, existing tools don't cover all sharing styles.

Related work: peer-to-peer sharing; role-based access control; domain-specific sharing; preferences for sharing and privacy (Olson, et al). Users preferences vary widely, and users treat information in clusters of types. Additionally, users think of individuals in the same way - they think of people in groups when it comes to sharing content. A survey tried to understand the 'how' of sharing - how do people figure out which tool to use and the people with whom they share. Asked for specific sharing instances, and the type of information shared. Asked how they went about sharing the data, and also asked about access control systems. Finally, asked for breakdowns and deviant cases, instances where it didn't work as was expected. Followed this with attitudinal scale.

Population was ten expert computer users in a medium-sized research organization. Average of seven sharing recipients, including individual and groups. There were 34 different types of files shared, including common things like business documents to the very strange like "ideas" or "tv shows". Relatively few sharing tools were reported; 43% of all sharing was done via email. Overall, however, there were multiple and simultaneous mechanisms used in 13% of the responses. An example would be posting to a website and then sending an email to the recipient telling them that information was available.

Breakdowns found:
- Forgetting what files had been shared and with whom.
- Difficulty selecting a sharing tool with desired features available to all recipients. Unclear if the end audience had access to the tool.
- Frustration with having to send out-of-band notifications.
- Resistance to sharing caused by a lack of feedback about changes to the sharing state. One user was resistant to posting papers to her personal website, as she wasn't able to track what happened once the information was there.

List of common mechanisms used in the study developed a set of sharing characteristics. Scope of sharing, addressing, visibility, notifications, location of files, access control, logging, redistribution, versioning. The first four turn out to be the most significant.

Scope deals with the quantity of potential recipients; availability of the tool, compatibility of the underlying transfer mechanisms, and more universal tools were more popular.
Addressing discusses how the sender identifies the end audience.
Visibility shows how well the items are reflected back to the sharing user.
Notifications are on the recipient side.

This identified, bottom-up, the characteristics of sharing. Found a divide between the items in common mechanisms. Email is a push model of information sharing, where the rest seem like a pull model of information sharing. Push-oriented sharing requires attention and effort by the sender, which may require attention and effort by the receiver. It requires a priori knowledge to name the recipients. Pull requires more attention and effort by the receiver.

Design goals were identified, to position between push and pull. It was necessary to also address issues of sharing discussed above.

Sharing Palette User Interface was developed [video shown]

SPUI represents a Unified Model of File Sharing. Requires little attention and effort by the sender: just drag an icon to initiate sharing. Requires some effort by sender, and some by receiver. Users select a sharing tool based on how well its affordances and features map to the sharing situation at hand.

[CHI2006 - Unedited Notes]

Support for Activity-Based Computing in a Personal Computing Operating System.

http://www.activity-based-computing.org/

Jakob E. Bardram
Mads Soegaard
Jonathan Bunde-Pedersen

Bannon, Cypher, Greenspan, Monty, CHI 1983. "Current human-computer interfaces provide little support for the kinds of problems users encounter when attempting to accomplish several different tasks". Computers are working on an application level, rather than an activity level; most contemporary computers are data and application centered, rather than activity centered. When there are a lot of interruptions, these types of systems break down.

The desire and goal is to multi-task based on activities.

In order to work, this system needs to integrate with applications, data, services and network in the OS. It needs to be fully and natively synched with applications - it needs to extend a well known operating system or user interface.

Healthcare and hospitals were used as a case study, but the question was raised - can this work with other disciplines as well? Can this be used in a "general-purpose OS"?

Activity Principles:
Activity Centered.
Activity Suspend and Resume (heterogeneous devices)
Activity Roaming
Activity Adaptation
-
Activity Sharing
Activity and Context Awareness.

Activities have Services, and Services have Data. An Activity may take several weeks. Several Services support this activity, and all access different data items.

Each user has a list of activities - each of these activities:
- maintains state information
- maintains look and feel
- maintains access to data
- stores state persistently
- can be suspended, and later resumed (different than virtual desktop)

This also addresses multi-tasking and interruptions.

Activity Roaming deals with moving from one device to another; the goal is to move across heterogeneous devices. Support mobility and exploiting computer resources in different places.

Activity Adaptation deals with large interactive displays, large desktop clients and tablets - all used in the hospital setting.

Essentially, swapping out Applications for Activities - the start bar no longer has a list of open programs, but instead has a list of open activities.

Implemented with AML - Activity Markup Language. An Activity ontology in XML, using the ABC Protocol (ABCP) - HTTP like, which allows for publishing and subscription, as well as client server and ABC to applications.

Potential problem: is it really necessary to bundle everything into activities? What if there is just a small, casual task that occurs - where do you put that? And if you categorize something into the wrong activity, how do you find it again?

Conclusion: ABC is useful as an approach to general purpose computing. The ABC principles and technologies can be embedded in Windows XP and work as a natural extension, with special focus on a single user experience. It makes sense to move ABC to the operating system.

[CHI2006 - Unedited Notes]

Laura Davish, for Luis Von Ahn
Carnegie Mellon University

Peekaboo

Human Computation

There are many things that humans can do that computer programs can’t. We want to solve these problems by making use of human processing power. Over nine billion hours of Solitaire were played in 2003. Empire state building was built in 7 million human hours (or 6.8 hours of solitaire). The Panama Canal was built in 20 million hours, or less than a day of people playing solitaire.

We want to make good use of these wasted human cycles.

Imagine the human brain as an advanced processing unit. It can solve problems that computers can’t solve.

The first problem is labeling images with words. Input an arbitrary image from the web, and output certain keywords. Computers can’t solve this problem. However, methods of accurately labeling images with words would be valuable. Image search on the web is an example of this value - google search on the web. Google images uses filenames and html metadata to find particular images. The problem is that this is not always accurate; "dog" may imply a particularly poor looking man, rather than the animal. So we could improve image search if there was a way to label all images on the web.

Using humans, cleverly, to label images on the web. If you just asked them, you would have to pay them a lot of money. Human computation is better and cheaper - get them to want to label the images for free. Get the people to pay you to do it. Tom sawyer.

The ESP Game!

A multiplayer online game that people like to play. As a side effect, they are labeling images for them. And they are labeling them accurately, and quickly. Could label all images on google in a few weeks.

The ESP game is a two-player online game. Partners don’t know each other and can’t communicate. The goal is for you and your partner to type the exact same word, given that the only thing you have in common is an image. The best strategy is to type words that are similar to the actual image. When the first and second player’s words connect, we find a very good label for the actual image. http://www.espgame.org.

Is this game fun?

The game has created 15 million labels, with only about 75,000 players. There are many people that play over 20 hours a week. People play for over 15 hours straight.

The ESP game is an algorithm. The input is an image, and the output is a set of keywords. We can prove its efficiency. The game is an algorithm - they are "Games with a Purpose".

Peekaboo is another Game With a Purpose. www.peekaboom.org

Locating objects in images. The above game tells us that there are words associated with the images, but don’t tell us where the elements are within the images. We can use this type of data to improve computer algorithms for visioning.

A two player game. Two people on the web, don’t know each other, and can’t communicate. Peek and Boom are two separate players. Boom is given an image and a word. The image is, for example, a butterfly, and the word is "butterfly". Peek gets nothing. The goal is for Boom to help Peek guess the word "butterfly". Boom can click anywhere on the image, and Peek gets a glimpse of a circular area around the click. Players can reveal and also point, allowing for a very minute detail about a particular image.

Is it fun?

27,000 players
2,100,000 pieces of data

Some people played for 120 hours in 10 days - 12 hours a day.

Is the data any good? Yes - use segmentation and cross over between multiple evaluators in order to get strong data. Sometimes the segmentation algorithm doesn’t work, but can easily create a search engine with the data highlighted in the images in an approximate manner.

[CHI2006 - Unedited Notes]

Scott Cook
Co-Founder, Intuit

Creating Game-Changing Invention

Cirque Du Soleil built a larger revenue-model than Ringling brothers' 100 years of existence; they did this in a market that is generally flat-to-down. They charge many more per seat than Ringling brothers; they did this without pulling share from Ringling brothers. They did this by creating a brand new business.

3M; Dick Drew. Was working on a new kind of sand paper; visited auto-body shops. Looked at two-tone paint; heard the body shop individuals talking about the problems with two-tone paint. The masking would remove paint, requiring a lot more sanding in order to fix it. Dick saw this as an opportunity to sell more sand paper, but also to fix the masking process. Required an adhesive that would leave a cleaner edge but would remove cleanly. Tinkered on a "wimpier tape". Bill McNight, CEO, asked him what he was doing. The CEO told him to stop, and get back to the sand paper. But he ignored the boss; he went back to the adhesive tape. Dick Drew invented Masking Tape; then Dick Drew invented Scotch Tape. He studied the customer first. "The bottleneck is always at the top of the bottle" (Peter Drucker).

What kind of company are you building?

By what method will you develop the next set of innovations that will lead you to bring bigger and better things to customers.

Models of Invention

1. The Lone genius
2. The boss is the genius
3. Copy Competitor's Inventions
4. Cloister the geniuses in a lab
5. Make your people the geniuses

Recognize that the source of invention is unlikely to be some big executive buried in the middle of the company; the raw material happens at the surface of the company, where the company connects to the customers. This is most likely to be done by the people who have the most contact with customers.

Managing becomes creating a way to nurture this level of team building.

Principles:

1. invention comes from mindset change. This requires a paradigm shift; consider the shipping business. This is a business that usually requires ships to move faster, thus lowering cost. By 1960, the "NS Savannah" was the goal - Nuclear Ship. Was a giant flop. This had no effect on the world. Malcolm McCleen. A trucker from North Carolina. He would deliver goods to ports to be shipped. He observed a tremendous amount of delay, work, cost, to transfer boxes into the cargo holds of the ships, and then the process was repeated on the other end. He sold his trucking company, bought a tanker, added a flat bed, and loaded it with boxes. His maiden voyage traveled from NY to Florida. He invented containerized freight, which revolutionized shipping. The cost to load and unload materials onto a ship was about $6 a ton. Containerized Freight was .16 cents. This changed the world, as there are hundreds of millions of people in Asia that owe their middle class status to Malcolm McCleen. Changes ripple through culture.

Malcolm didn't focus on the speed of the ship; he focused on a mindset change. Mindset is a paradigm, or an expected belief, or a mental model, or an assumption set. Science gets stuck in paradigms, where progress stops. Then some young person, generally under 30, invents a new paradigm and changes the world.

Intuit; change lives so profoundly people can't imaging going back to the old way. Employees => Customers => Shareholders.

April, 1992, launch of Quickbooks. Quickbooks were selling at twice the price of the market leader. Quickbooks had a smaller feature set. The advertising was terrible. Marketing was a flop. And there were bugs. Lots of bugs. Goal was, obviously, to fix this.

In fact, when Quickbooks launched, in one month, went to market leader. never got less than twice the sales of the market leader - at twice the price. What explains this?

Conducted research; actually began to call the customers who said they were using Quickbooks in their office. This was a home product after all; what are you doing with it in your office? Found out that most people hate debit and credit accounting. In a small company, most people can't afford an accountant, and the small business owners hate it.

Quickbooks was the first piece of accounting software that had no debit and credit.

Quickbooks Revenue Growth was enormous. An exponential curve, with innovation after innovation. New inventions that have kept the business going. Albert Gyorgyi, novel Laureate, "Discovery consists of seeing what everyone else has seen and thinking what nobody else has thought". Success starts with humility. Empathy is not just about walking in another's shoes; first you must remove your own shoes.

500 million people business; volume over $1 trillion annually; has grown over 10,000 since 1970; grows at 20% per year; operates in over 100 countries. VISA. Dee Hock, founder of visa: the problem is never how to get new, innovative thoughts into your mind, but how to get old ones out. Every mind is a room packed with archaic furniture. You must get the old furniture of what you know, think and believe out before anything new can get in".

Our minds have filters, and filter what we see around us. Get out face to face wit customers, in their office, in their home, and spend time watching them.

Typical development process.

1. Define requirements
2. Design solution
3. Involve UX
4. Uncover customer problem

Instead, flip that. Do it backwards.

1. Involve UX
2. Find customer problems
3. Define requirements
4. Design solution

IRS mails 8 billion pages of tax forms and instructions each year. If laid end to end, they would stretch 28 times around the earth. Americans spend over $200 billion each year, and 5.4 billion hours, to prepare their taxes. That's more hours than it takes to produce every car, truck and van in the United States.

While TurboTax is doing well compared to other software packages, they are doing poorly compared to other methods (HR Block, by hand, etc). So it was necessary to go out and look at customers. They created five segments of users, self-directed, simple and know it, can't be bothered, worrywarts, too hard for me.

The worrywarts are people who don't mind doing taxes but are afraid of the IRS. They want control but don't want to do it alone. "I'm afraid of the IRS" "I want someone to hold my hand". New mindset: Worrywarts want to do their own taxes together with a tax preparer. They want a mixture. Combine human experts, with technology, into one package.

Victoria's Secret. They have a flat market position of just about 20%. They hired some outside individuals to go to woman's homes, and pawed through the woman's Lingere drawers. Found that consumers who had Victoria's Secret underwear thought it was uncomfortable, and were reluctant to wear it. Came up with a new form of fabric that stretches more than 25% of other fabrics; share of market doubled to 45%.

Executives study users in their context too! Listen carefully when they call, hold town hall meetings, deck the halls with findings, photos, findings, artifacts. A culture of user research and ethnography.

How does this change mindset? The source of truth is out there with customers. Seven sources of innovation, according to Peter Drucker.
New technology and science; changes in public perception; demographic change; industry and market structures; process need; incongruities; the unexpected.

Invention comes from mindset change; mindset change comes from seeing differently. Savor surprises, as learning.

SC Johnson makes Raid, a bug killer. The pack a lot of materials in it so it is slow acting - so the bug goes back to the nest and then transfers the chemical to other bugs. They went out and watched consumers use the spray, and guess what - they drowned the bugs in spray, and of course, the bug didn't make it back to the nest. They changed the nature of Raid - added a lot of water - and created a new kind of Raid. Sales up 50%.

Sold two versions of Quickbooks, one with hardware and the other without. The one with hardware - contrary to CEO expectations - outperformed the other - all due to listening to customers.

It is necessary to fail in order to swing from the fences. Celebrate failure. "Greatest Failure Award". Intuit actually gives an award for the greatest failure.

Focus managers on a customer metric. How do you get creativity out of a metric? Financial metrics already exist. Most companies have lousy customer metrics, though. customer metrics lose when they are tame; need to create a powerful customer metric and focus all managers on it. Net Promoter is a way of quantifying experience. The only way to grow a business to get customers to come back for more and to tell their friends. Ask people to rate the tool on a 10 point scale to determine if they will tell their friends about it.

Nurture and protect teams. The role of the CEO and management team is to protect the user experience groups.

Here I am in Montreal, and damn, is it cold. And wet. And cold. I registered for the conference, ran into Katie Minardo which was nice but weird, and then met up with the SCAD kids.

After walking about for a while, Ryan, Bri, Nick and I found a place to eat some authentic Canadian food (ok, we had German-style pub crud, which was plentiful and fattening).

Now I'm going to figure out what I'm doing tomorrow. And then go to bed. I'm beat.

It's definately 3:58am and I'm wide awake.

Here are some thoughts I'm putting together on the distinction between the m.Des and the MFA degrees (30k pdf) - nothing like academic semantic brainfood for the wee hours of the morning.