From: John Canny [jfc@brunello.CS.Berkeley.EDU]
Sent: Monday, July 12, 1999 10:26 PM
To: jfc@CS.Berkeley.EDU
1. Introduction

John Canny, UCB: HCC Retreat Overview
Fac member since 1987. He was in robotics and is new to HCC.
HCC is 30-40 faculty on campus. Not yet formal structure. going after funding from NSF, and others sources.

HCC: ubiquitous, quiet, tightly coupled with the physical world. Implies computing in a human context, rather than creating contexts that human beings must learn and understand. Challenge: engineers must draw in social sciences.

Context: more than when and where an action takes place: includes activities, goals, roles. Wants to understand social behajvior for computer applications that truly assist people. Coomputing is permeating daily lives: changing nature of work, how peole learn, buy goods and recreate. Transformss society. Social scientists benefit from seeing emerging technologies up close and inusing computer tools for large-scale studies.

Changes: caused by info tech. creation of K worker and info copmanies. Agile corporations: temps, outsourcing, offshore labore, retraining. largest employer in the usa is now manpower. average occupation is now 3 years: implies must retrain. ubiquitous networks and comm create new social ties and reshape social networks. promiste of education: learner autonomy and lifelong learning. shift to small, dynamic groups that change their tasks, inc work at a distance. challenge: working at a distance and keeping social closeness. computing is becoming a basic literacy tool. unevenly distributed in the world, which creates inequalities. HCC seeks to create a bridge between computer and social sciences.

Timing: computing seems to be a great success (credited for rise in Dow), but future sccess of info tech depends on scaling barriers that depend on non-tech barriers.

The challenges (the walls):

1. Natural hCI, inc speech, interpretation of dialogue, inclusion of affect.

2. Computer literacy and lifelong learning. A competition: hci wants to make computers easy to use and not require much skill for ordinary people. opposite approach: make sure everyone has a foundation. not just fulltime learning.

3. Face2face vs electronic interaction:  working with educational faculty at ucb.

4. Codified vs. tacit knowledge: how to augment people's natural skills. procedural k is not normally available to computer systems, even tho most people use computers now and it could be trapped. we must understand people's patterns of communicaton.

5. Engineering vs the social sciences:

HCC Themes:
natural inteactions: pens, gesture, speech.
design of learning tools, tools for learning design.
design as praactice, tools for doing it.
cmc tools based on the psychology of interaction.
mining tacit knowledge, social and computer networks.

HCC faculty and staff: some of them are:
from computer science: Includes Robert Wilensky, james landay, john canny.
from sociaology: wellman, bienenstock, castells, fischer.
fromGSeducation: andy disessa, marcia linn, michael ranney.
from psychology: dacher keltner, jjerry mendelsohn.
from mech enginerring
from business: rober e cole, jim loincoln
from sims: marti hearst

2. Telepresence, AR, Tangible Interfaces

John Canny, presenter

€PRoPs (personal roving presennces) and gesturing avatars.
€UPM: a universal planar manipulator
€Bearable computers: turning laptops into AR systems and wearable computers.

1. Props:

[i missed the presentatiaon here]

Taxonomy of interaction behaviors:

verbal: symbolic and prosodic
Nonverbal comm theory says main cues are:
-gaze, proxemics, facial expression, body and hand gesture, posture, touch. e.g., a handshake depends on touch.

Differences between speech and gesture:

Why nonverbal gestures are spatial:

You need social context as well: communiation, persuasion, trust-building. These are very diffrent contests that rely on different nonverbal cues.

Two-way street: props provides selective control over partcular cues. they profide testbed for detailed sutdy of nonverbal cues. they can create a rich variety of media and study invfluence on overall perception of other person. what makes a persona?

Input to Props: how do you achieve rich, expressive input for props/avatars with today's technology?

Future research: gaze cue repair: face2face is not perfect. there are common breakdowns in f2f communication, such as gaze aversion. it may be possible to rectify gaze aversion using image processing and/or a synthetic gaze cue.

2. UPM: a universal planar manipulator

A gadget: a smart desk can move a large number of objects independently.

Bearable Computing: explores issues in personal, persistent computing (AR, worn UIs) using ordinary laptops. Can avoid headmounted display, head tracking and caples. the approach: use optics to overaly computdr images onreality, but use laptops or pdas.

Augmented reality classroom is using the project. sudents work in groups of 5-7, communicating silently via pen or keyboard chat. each group has one main notetaker, the others add their own commetns or question to the transcript. students can mark up the group transcript of lecturer's notes. there is nonarchived chat, also. one student per group works as facilitator or ta, posing quesont to others and testing undestanding.

3. Informal Tools for Multimodal, context-based user interface design

Pen-computing, context inference, MM UIs

James Landay, UCB

Trend in computing: mainframe to mini to personal pc leasds to innovation and integration waves:

post-pc era: what is next: "universal computing"

universal computing:
1. includes covering all or a a whole colelctively or distributed. eveeerywhere:

Away from the average device: to personal, small devices. small, highly mobile, embedded in environment. intelligenxe: immense storage and processing in the infrastucture. everything connected.

HCI challenges:

1. Universal computing devices will not have the same UI as "dad's PC": a wide range of devices, often wi ssall or no screens, or alternative I/O, special purpose to particular applications. special purpose to particular applicatons = info appliances.  lots of devices per user all working inconcert.

2. Design of good appliances will be hard: single device design is easy. hard to design same app in a consistent manner across many devices, eg a calendar app, one speech-based, one GUI based. hard to make different devices work together: which is used when, multiple uis and modes/devices means which should be displayed, building awareness of context of use into design is the key to some of these issues.

3. multimodal input is assumed, but little design support for creating multimodal interfaces.

Their approach:

1. build novel apps on existing appliance, eg, notepals on the palm pda plus crosspad; build new info appliances, eg pda using purely speech.

2. evaluate appliances in realistic settings.

3. iterate: use the resulting experience to build more interesting appliances to build better design tools and analysis techniques.

why is multimodel interaction had, and best practices for designing guis

Multimodeal interaction: when commncting with people, we use more than one mode at a time: gesture+spk, sketch+spk, etc. computers would be esaier to use if we used multimodal hcis and if we used innate perceptual, motor, and cognitive skills. more people would be able to use comptuers inc users with vision or motor imjpairments.

James Landay described his process of researching how to make better interative prototyping tools for UI designers.

What they have done: informal sketch based tools for GUI design, inc. sketch-based tool for website design.
They are also changing a speech-oriented tool. low-fi prototyping of speech=based uis.
The have been able to make some auto generation of uis.

Universal computing is about supporting people, not ubiquitous computing or pervasive computing. we should not focus on bits and servers. we should focus on what are people's tasks and what can we do to support it better.

We need a new kind of appliance = device plus app plus UI.

Major place for success: take advantage of context, used in realistic settings.

4. Computers as Design and Learning Partner

Alice M. Agogino, Mech Engineering, UCB
Sherry Hsi, Center for Innovative Learning Technologies

combines the learning done in design and design done in learning.

Vision hypotheses:

Design process is driven by info
Most of design is redesign
Most design requires integration of hw and computing
Design is a social process: enterprise wide design requires learning form others in a team and use of hot skills:
hot skills: communic, teamwork, leadership...difficult ones.
cold skills: math, analysis....more straightforward.

Imagine the workplace of the future: showed demo.
Manager promised full access and research and experts and money. time and cost management are both important. this is an intro.
designer needs to understand components, review old products, look at old projects. lit search, consult experts, design studio.
May need data mining to relate data that was not combined in the database.

In the design studio, can animate, simulate, create physical protoypes.

Designer may want to review and select design paramters.

May need to balance time for research vs synthesis.

Multimedia case studies of engineering design.

See website:

Highlight examples, shows best practices.

She claims learning and design are two sides of the same topic.

Internet is perceived by teachers as goo or essential.

Nat Acad of Sciences:
being fluent with info technology
Prepares people

Human-centered computing: design and learning partners.

Techgy frees teacher to prompt and coach students.

Techgy can make learning visible: model thinking. scaffold sudents to explain their ideas. provide multiple variations.

Techgy helps students learn from each other. SpeakEasy Web discussion tool. listen and  learn from each other. promote productive and respectiful interactions. design criteria and standards. emply multiple social activity structures.

Threaded discourse.

Female teacher wanted to ask students how to change class. better to do this earlier, not later.

They noted that female students spoke up more in online courses. shy students spoke up.

Big gender differences.

COOL = next gen discussion tool: Collaboratiave online oportunitites for learning system.

Doug clark explained this asynchronous threaded discussion.

reviewers can rate comments.

In R+D environment, threaded discussion are being subjected to text analysis so one can archive useful topics, to see patterns, to review design history.

Results of online learning.: 15% contribute in class discussion. 85% contribute to electronic discussions.

Coaching improves K integration. They studies 8th graders. May96.

Coaching improved about as equal to cybercoaching; not as good, but close.

To examine situations well: need to look at context, goals, beliefs about the context, eg schooling vs social chatting. This is scheduled asynchronous.

Gender differences: once they used technology, did it change facetoface? their answer: don't know, but anecdotal evidence shows some improvement in taking risk.

Innovative Learning Scenario.

UCB, Vanderbilt, and one other place are involved in innovative use of technology to reform k-12 education.

SimCalc Mathcars

these are explore, apply, reflect simulator on the palm pilot.

Smart Probes
The concord consortium sonic ranger
kids could bounce balls and measure distances and time.

NotePals: Web-based repositrory of annotations:

they have a number of appliances.they are trying them out.

Information Capture: Pens = personal electronic notebook with sharing> Larry Lifer at Stanford works with Xerox PARC.

Techgy promotes lifelong learning.

promotes reflection and integration of knowledge. promotes critical evaluation of diverse sci/eng info. establishes a generalizable inquiry process. prompts for reflection adn integrate asessment in learnaing. simulates work environments.

the transition from the university to the k-12 envienronments may be faster.

Lifelong learning Simulates work environments examples:

multimedia digital library: The virtual Disk Drive DesignStudio, by Yu, Agogino,UCB courseware, 1997.

Brief Presentations by different Participants

5. Siemens TTB, R+D+B factory.

Arding Hsu.
They identify mkt segments, dev techglies, prepare bus plans, do pilot testing. The invest money and domain know how. they get vc money.

they look for conservative, consensus, control of vcs typically. they don't make so much control.

iq property issue: they support research with non exclusive rights. sometimes in case by case basis, they want exclusive rights.

He works with Larry Rowe to set up a good working model.

6. Interval

Bill Verplank
Interval is 7 y old. they wanted to generate a profit within 10 years, but may need to wait longer. they have already spun off 5 companies. central concern: how people communicate.they have some physicists. they have relations to mit, stanford. they have design relaoins to rca in london. they support some ucb people.

His view of UI design.

three basic questions:
1. how do you...affect the world, how you do things.
2. how do you feel? he has been working on active force feedback. not only feeling hard things, but the quality of the feedback.
3. how do you know?

Ontology, via terry winograd.

should a ui have buttons or handles? interfaces not only to info but to expressive realms.

laurie anderson will be in berkeley to do a performance about mobydick.

he is interested in expressive communication.

should a door have a button or a handle on it.

buttons are the trap door to automation.

hi likes cool and hot. cool = incomplete.

when tv converges with cpus, sharp and fuzzy collide.

in the knowing realm: maps vs paths.

One of the spinoffs is Zowie entertainment., san mateo, ca. shows 5 yrs of that will be out in dec 99. smart toys. interactive. they did play research. can sell toy for $50. usijng an antenna. they can track locaton of objects. multiple hands. many kids can interact they pay attention to screen or toy.

mark sheff works with robots. affective interactions. robot can do many postures and emotions.

7. FXPalo Alto Fuji/Xerox Lab

Elizabeth Churchill

An outpost of fuji xerox from japan. 40 people. she's been there about 2.5 y.
Fuji/Xerox is an outgrowth of Rank Xerox and Fuji film in japan. Rank and Xerox dissolved their relationship.

sw R lab. colocated with Xerox PARC. take advantage of silicon valley. links to ucb, stanford, mit, usc. collaborates with james landay. has used interns from ucb.

general approach: user communities, knowledge, technology.

user community field work:
how do people do ordinary work. inspired by enthographic research. interviews in work settings. observations of work ractice. innovations based on work practice observations. broad participation.

One of her projects: VR environments and synchronous, asynchronous  communication.

Process: Create, Use, capitalize.

creation: user communities, tech invention, technology evalutation. use: prototypes and deployment. Capitoalize by licensing and nproduct development.

3 major groups at fxpal:
1. human centered mobile computing.
2. adaptive software.
3. smart spacas.

Sarah Bly works in group 1!!
One group looking at quantum computing, discourse texts, and smart spaces. UIs to video, smart rooms, etc.

Group 1: human centerted mobile computing:

X-Libris: personal document readers; reading, annotation, skimming, how people use paper. how can a portable device use paper concepts.

The Palette: tangible UIs, presentation practices. How do people do presentations. each slide in a presentation could have a bar code.

Virtual Environments @ work: lightwieght communication, CMC and FTF(=f2f) conversations and collaborations

Conversational characters: charaacter-based UIs, gesture and speech recognition and production.

Industry partnerships

Daedalus/Barwan (layered.....

8. IBM

Ted Selker

Showed a trasparent screen using a thinkpad.

User System Ergonomic Research
Works on physical, graphical, and cognitive UIs

When computers are strapped on, then interviewees think the interviewer is the enemy, not just alienation.

But at Hertz checkin out on the driving field, the person with a device in a holster is a friend aiding your rapid exit.

Cognitive issue: can we introduce beliefs.

he has about 30 people in winter, 45 in the summer.

Info techgy directions:

everything everywhere connected. computers will be aware and reduce transcription. computers will understand, teach, augment, proxy.

Relationships,things, and comunication.

culture is what changes slowly. good uis tas taking the tool out of the task: ....but social context: sharing a task, transferring a task, discrete memory or communication or computation aid, sociaal statements: style, K, status.
Be careful: social statement we make may make it antithetical to be ing adopted.

Towards a behavioral motor match.

He mentions the ibm tracking device in the keyboard.

Techniques should support scenarios.

He studied 3D vs abstract icons.

Room with a view:

New integations: showed wallet.

What are pcs coming to: printer, email, server access, context awareness, user awareness, input sensing.

Balancing Info: He showed interest tracker demo, an eye tracker.

Future feel of tools. We have been jsing computers for cataloging. info is beocming personal and personalized. balanced diet of info.

9. Xerox Parc

Paul Durish

Xerox Parc does lots of interaction study, but not in computer science lab. their interest is systems, networks, databases.

his background in cs. he thinks and works with sociologists.

UI guy: it's not about users, its not about UIs. it is about PRACTICE. i.e., what do people actually do, some people think practice is stable. but:  impact of technology,evolution of tecnology impacts practice.

Examples: 1. evolution of behavior in media spaces, 2. customization technologies, 3. email patterns of organizational communication.

Implicatoins: 1. foundational grounding [what does this mean? he said something about theory, not only of computation, but also, of human organiaiotn and impact of techgy on society], c2. omputation as a medium, 3. experimental subjects don't have practices.

Having the techgy there changes the practice.

Customization is a social practice supported by techngy.

He is trying to ground his work not jsut in sociology observation but sociology theory.

He tries to do grounded ethnographic work.

He says: Experimental subjects don't have practices. they have only behaviors.[...!!!????]

In his circles, practices means processes.

10. SRI

Luc Paridi??

53 y old company. spinoff factory. (tm). website =

CHIC = computer human interaction center. new group.

He will show some demos. all of the demos are based in their open architecture paradigm.

Multimodal demo: speech and pen

they have lots of recognizers.

agents talk with each other in a kind of natural language so humans can monitor it also.

he showed a map image tha t uses gesture recognition with a pen and circling of items.

he views this as an augmented paper.

He showed Maestro: a way to retrieve info from a giant video database. Has 12 recognizers: speech, frame, and others.

He showed the Fridge metaphor, many words taken from key words of videos.

He asks: for video with blare and iraq.

He sees a diagaram of time, speech recogniziton, speaker id. image shows 12 rows.

one recognizer does ocr recognition. it oculd recognize tony bglair in the caption.

He can put together constructs. the little recognizers cooperate.

He even has a flash recognizer that can recognize flashbulbs going find interesting people.

He is doing a 9/9/99 event at 9:am??? he is giving a big demo.

Question from audience: how can educ techgy use some of this. Answer; they use open architecture. so they can buy recognizers out on the market and customize them into their suite very fast.

11. Nokia Research Lab

Ikka Tuomi

Infosociety and knowledge management.
he is visiting scholar at ucb june 99 to may 00.
He has been at nrl for 10 years.

1. nokia cmc and K managemetn.
2. R on K creation, social learning, and innovation processes.
3. Social implications of info and comm technologies.
4. He is on info tech forum of finland.
5. he will leave out wireless comm devices and technolgies.

Some KM activities in Nokia.

Collaboration competence center. founded in 9e. generalists. about 800 people. developed lotus notes pilot applicatons, bus inteligence info systems. acted as core for nokia rollout of lotus notes to 18K users  (now 40K users) over 100 servers in over 25 countries.

NRC K management group: next gen KM concepts and tools. using NRC as a pilot site, focus on K creation.

KM in their business groups: customer services, r+d, news services, global customers,...

NRC KM architecture

People: communties of practice relate to tam collaboration. They had these two diffrent concepts or kinds of people in organizations.

The technologies or tools they developed to support each group: CoP cafetria, Teamwork support systems.
The created a K interface between the groups and the support tools.

Corporate Knowledge: some key ideas

problems are deep. must use deeper theories and look deeper.

most websites are not good content.

He has written a book on KM theory.

Outline of KM theory

2 epistemologies: analytical, emprical vs. phenominolci=gal, constructivistic, social.

theory of intelligent orgnizaitons
-theory of intelligence; based on phenomenoloigal epistemology, social meaning, processing, an d activity theory
-theory of orgnizations; review of alternative views and the dev of k-based view.

theory of KM:
-analysis of various roles K plays in organizations
-discussion of K creatioin (elarning) models
-analysis of the SECI model
-devleopment of a new model, th e %-A model

He is trying to apply his theory in 2 areas: measure ment in the intelligent oraganization, organization for k creation.

3 dimensions of organizational activity:
1. motive....act = activity
2. input....output = production
3. knowledge....communication = meaning nprocessing

Dynamics of communities: shows diagram.

ephemeral/stable/insitutionalized = y axis downward.

homogenous vs heterogenous = x axis

socialization ba externilation ba combinatioin ba
thought community vs task force
commof relective practiinoners
competence community

community of pratice
thougth style
organizational department function

here is the heterogenous column:

combination ba
....he took away the slide

Living in a Lab: observations on the ihistory an dfuture of the finnish ino society

Finland: 5 million people, 100% digital telcom infrastructure. hitghest mobile comm = 60%. hightest internet hosts in the world. 98% schools hav internet access, 2h/day per pupil. largest populaton of internet banking in abosoultue numbers. finland is siimilar to silicon valley but includes all ages, diverse kind of people.

Some infrasturcture trends: 16 kbps data in gsm already supported. bluetooth = short range wireless connectivity. wap = wireless access protocol. full banking thru fones. can pay using mobile fones in food stores. has bus schedules on the pohones. can order tickets using mobile phones. next yeasr, will have 160 kbps and video on phones.  soon in two years: 480 kbps.  in finland, you can take it for granted that people have a mobile phone...everyone. being phoneless will be like being homeless.

they will rollout elecronic identificatoinsystems by end of 99!!!!! in usa, govt seen as enemy. in finland, govt is trusted more.

they have medical info flowing through the phones.

12. Pixar ANimation Studios

Karon Weber


30 people work on UI tools for film production. her approach is to the pipeline, to get the document out fast.

projects driven in current gen driven by schedule, tasks.

she has 10 y of legacy code in users.

many users have known nothing else but legacy code.

Thre was never an overarching study of what animation production tasks and process are.

Evolution of the shot: evertying starts in preproduction. every scene as a number and a shot numbers

all the character production goes on here. here prod tools don't play a big role. people use oil and water colors.

Technology shows up as soon as production starts.

 the steps:
1. layout wrieframe groundplane. and blocking.
2. root model and anthill installed.  key frames. posing.
3. set dressing.
4. crowds placed in the set.
5. forground elments added.
6.plastic shaders and camera lights.
7. final shaers installed.
8. final light and camera depth.

Her user community: 275 people in many different departments, each with workpractices across departments and users. functionality determined by workflow, work practice and task evoluton. system must continue to evolve to accommodate changing production pequirements.

Each has distintive ways of working, concepts, tasks.

3d animator thinks of splines.
3d claymator thinks of objects and scenes

today's films only give a glimpse of the future needs. where will cg go? what is after  crowds? what about cloth?

How do i design a ui to design wrinkles.

Some of their work practices.

observe and interview user community. storyboard use and scenarios. build paper mockups. run wastcher studies. develop computer based concept mockups. pop quizes. generate ui specs. buld prototypes.

These systems does actually get build. delivering designs for manufacturing is hard. the community is using it immediately. her constraints: techgly vs time vs resources.

Some issues of interest:

1. task analysis and user work flow. some users have allowerd her to install watchers.
2. interaction techniques.
3. customizable work environments.
4. cross platform UI consistency.
5. screen management techniques.
6. context sensitive operations.
7. 3d work in a 3d world.
8. visual design = most controversial topic she has. surveys show for spreadsheet editors: which are right. the users are all over the board. how to come up with a visual design to solve her group needs.

Context sensitive operations: big interest of her.

Thursday 8 July 99

13. PRoPs:
Towards Transparent Interfacaes to the Real World

Eric Paulos Computer Science Divsion University of California, Berkeley,

Personal telepresence devices.
CMC now means...telephone, email, chat rooms, muds, video liveboard, cvideoconfrencing.

What we want: face to face, presons.

Interactions: structred, technolgoically mediated, human interaction. but also unstructured, casual, everyday human-human interaction.

Current technoolgies:

video teleconferencing and video walls: typcally hi quality video and audio, fixed gaze, not chance, casual encoutners, onesided experience.
Telepresence: specific narrow tasks, special oneoff hardware, trained operators

An example: the blimp.

Terrestirial Props: surface cruiser or cart.

UIuses microsoft netmeeting. communicates withjava applets and internet.

Noverbal communication:

gestures, speech independent, speech related
facial expressions
gase or eye behavior
vocal behaavior

Props now has an lcd that shows the face of the speaker.

Physical appearance and viewpoint:
-shape, size, height must facilitate intrxns
-friendly and unthreatening form
-conveyeance of personality, mood, or manner
-importance of a few human-like visual cues
-vantage from human head heigght

Proxemics: the study of the natrue, degree and effect of the spatioal separation individuals naturally maintian.

Pointing and simple gesturing: includes backchanneling.

His application domain of choice: social, casual end, not business.

14. Communicating with Avatar Bodies

Francesca  Barrientos, cs, ucb

Communicating with other people through avatars.

Examples of graphical reps of humans: Comicchat, oz, blaxxun. they are graphical worlds put on top of text chat worlds.

Some say text chat is rich enough: sense of place, space, personality [??!!??]

She wants to use human body for communicating.

Nonverbal behaviors: gesture, proxemics, posture, facial expression, gaze, context, appearance.

If you have a body in virtual space, how should i use it?

Previous approaches: expressionselection (emotions on faces), natural language inference (interpretation of comments in parentheses), automatic animation (mit research).

Beyond Chat: UIs for controlling animation of nonverbal behaviors throu verbal comm throu audio with continuous control. Needs to understand how user movements translate onto avatar bodies.

Importance of synchrony: part of meaning of gesture comes form it co-occurrence with speech. Rhythm, flow, stroke, force, shape, are important.

Kinematic mapping. between input devies and avatar bodies.

Evaluation: depends on task, ease of use...

Explore design space: she wants to protoype them and try different approaches. she can explore more quickly than with physical devices like props.

Network issues: synchronizing control and audio data.

Summary: avatars allow commm using surrogate bodies. gesture and speech are coexpressive, design tool for building props.

15. Social/Personality Psychology in Cyberspace

Joanie Connell (UCB) , Prof. Gerald A. Mendelsohn (UCB), Prof. Richard W. Robins (UCD)

Bridging engineering and social psychology.

Looking from the people perspective as to how to use technology. How do people interact using technology.

Basic question: how do people behave differently, and the same, across communication media? How do telephones, computers change the way we interact with people.

Social Psychology: Allport: 'is the study of the effect of actual, implied, or imagined presence of others on behaviors, thoughts, and feelings.' from 50+ y ago.

3 questions about interpersonal interactions:

1. Whar are they doing: behavioral ascription.
2. Why are they doing it: causal attribution
3. Do I like them: interpersonal attraction or repulsion

Each of these has large literature. They are gathering R on all 3 Qs.

Research Projects:
1. Social behaviors: self-presentation, impression formation.
2. Individual differences: personality, roles, gender

Social psychologists look at personalities that drive behaviors, eg. gender roles, shyness, etc.

Research Methodology
1. Series of systematic, controlled, experimental studies
2. Comparison across media
3. Interactive situations
4. Persceptions of won and others' behaviors.

How do we know people are changing systematically across media? We run experiements to study how people interact in different situations.

Eg, Do computers facilitate people behaving more aggressively? They run two different ways for people to get to know each other and measure the number of uninhibited remarks. Interestingly, in business situations, relatively few flames. Context matters. flaming occurs a lot in chat rooms. they try to study consistent

Their Research results

For f2f, telephone, chat.
They study nervousness, warmth and talkativeness.
No difference in warmth!
Gender differences in warmth. in f2f, men rate themselves..women rate themselves.
Warmth = did not define it. asked people what they thougth.
Their rating scale was 1 to 9. their data was measured along these scales.
They did not study how people behave vs how they rate themselves.

Continuing research:
Other measures: partner ratings, observer ratings, linguistic content, physiology.
Other phenomena: causal attribution, interpersonal attraction, trust...
Other interactive situations: groups, classroom, bsiness, other
Other media: asynch cmc in email, video, props

They did track cultural groups but...They studied only americans.

Digital Classroom Presentations

16. Taking Convince Me into the Classroom: Dusting off "The Doorstop"

Michael Ranney, sch of educ, ucb

Riesner's Workbench. Use in k-12 classes.

The Convince Me UI. showes a graphical rep of propositions. node and link depiction. he teaches in math and science. people must categorize evidence and explain reasons and show connections. Can show beliefs. Has underlying connection engine to analyze sets of hypotheses and state beliefs.

Results: convince me seems to be useful.

they are now bringing this system into high schools.

1. Reasoning in science is not too much different formsocial reasonging.
2. Explanatroy coherence theory is readily acpplicatble to social reasongin.
3. echo is a good constraint satifsfaction model.
4. convince me aids one's reasonging  by making it explitia and by deepning ones thinking about aaevicendce.

They tried this in high schools in oakland.with very little tech support. 3 students per computer. novices.
Other difficulties: spotty aaattendance. watered down curriculum, lack of attention span, limited bells and whistles, low SES = social economic status.

Audience comment: UI is very text oriented and demainding of literacay and typign.

17. Designing Ubiquitous Computers for Learning: An Early Study with Palms

Sherry Hsi, Center for Innovative Learning Technologies,

Looked at electronic discussions and supporting communities. Observed movement to create digital libraries.

Wanted to get kids outdoors.
Did a pilot study with palm pilotsand using imagiworks. Berkeley Wireless rsearch center.
Concord Consortium

Researach objectives:

Design for Learning: how to build scafllolds. how to link classroom to daily life. she is collabaorating with Berkeley Wireless research center.

Elliott Soloway and she ran a learning centered design tutorial at CHI98.

usercentered design ->learner centered design->learner and social context design->next gen??

Hypothesis for ubiquitous computers: lowcost, ubiq, good acitivites curricula teaching...wil reformeducation.

Pilot study 1: alameda creek project: diverse hi school. 50% spoke second lang at home.
-improve student understanding ob ecosysms.
-improve inquiry skills
-lik class K with outside learning
-capitalize on techgy for formal and informal use
-build understanding of science.

Used WISE, internet based system

Amanda the panda is the help guide.

spent a week critiquing evidence. made a concept map of linked entities. did field studies using pallm pilot probes.

Preliminary Results: students do improve their understanding. Did studies of pre and post learning concept maps. kids are poor at doing inquiry. teachers bad at using techgy and explaining. she sees the need for building archives. they did not study with a control group. hard to compare.

Cilt conference is an excellent place to learn about what people are doing with technology in education.

18. Using Note-Taking Appliances for Student-to-Student Collaboration

James Landay, cs, ucb

Observation: students leave classes with different ideas about what was discussed. Can we eimprove this by encouragning collaboraton.

Vision: take notes on note-taking appliances (electronic devices suited primarily for writing notes) and combine after class along with lecturer's slides.

NotePals: uses palm pilots and cross pads.

Advantages: inexpensive hardware, easy UI with freeform ink lets students focus on class, no wireless networking required, easy sharing load of notetaking, not limited to a single perspective.

Current Work:
1. performing a larger-scale classroom experiment to study how notetaking behavior changes, results may depend on the style of the professor's slides, testing methodology, and privacy.
2. build a notepals service with new uis. anyone on web could start a notepals group.
3. develop notetaking clients for new devices, eg., vadem clio, palmpc
4. makeing it easier to annotate other tyhpes of objects or media with notes.

He believes: sharing notes using notepals can help groups of students collaborate more easily.

He builds systems. not skilled in evaluating them.

Support for Creative Design

19. Digital Desk: Worspacae for Architectural IR

Ame Elliott (Architecture) and Marti Hearst (SIMS), UCB
She is in design theory and methods group. what is design, can it be taught? can one teach creativity.

Architecture: a problem finding domain, not a problem solving design.
Multidisciplinary, collaboarataive domain.
Cultrually learned practiaces of design lend themselves to ethnographic research: novices are different than rofessionals, hierarchical.
Characteristics: visual thinking, ill designed tasks, problem finding, not problem solving.

Creative Design:
Design=acat deliberately, to plan.
Designers use images of previous projects (precedents) to inform design process.
Designers design all the time, a lifestyle, not a job.

How can everyday computing support a designer.

IR in Architecture:

Currentlyl Web is not useful for IR tasks:
1. little content, poor quality
2. collections are narrow and deep.
3. searches emphasice precision on not recall.

Architects prefer paper images.
1. easy to juxtapose imagesand make collages.
2. Sketching on top of images helps them think.
3. Designing is a dialogue between designer and problem, and media is part of the convesation.

Digital Desk
from vision maker, big, can sit or stand by it. 25-37"
Size: drawings are large, 24x36, juxtgaposing many images easier.
Luminous: easy to trace, functions as a light table.
Flat and horizontal: good for drawing and spreading out papers.
Lightweight UI: pen input intuitive to designers.

What should Desk UI do?
1. support visuzlization
2. support search.

Example images. Images show collections, searches, relevant documents.

Hypothesis: users will havea an easier time creating a presenttion from previously created images using the digital desk than with a desktop or paper.
Task: users will creaste a 24x36 poseter presetning preliminary design ideas using 2 of 3 environments.
Comparison: Digital Desk, conventional dekshop, papersicssorsglue
She's using ethnographic, anthropological techniques to analyze architectural practice.

Experiment Design
3 meetings with participants: get task, look at 40 paper images to get ideas, take system tutotiral. hour and a half to make first poseterin one environment. 1.5 to make a second poster in second environment.
Comapare mental workload (posttest questionnnaire) and time for test.

1. DD will help IR inarch.
2. 2 aspects of function are integrating paper and dig docs and searach viz.
3. create an experiment

20.  DENIM
A sketching tool for prototyping web and desktop UIs

Mark Newman and Jimmy Lin, cs/ucb

Based onSILK, cmu project of james landy.

DENIM goals:
-real, deployable systems
-support for website design
-expanded capabilities
-even more informal than silk.

Web design study

ethnographical study.
visited 4 design firms, 1 large internet company.
interviewd 4 ui designers, 3 graphic designers, and X others.

Design specialties
Info design: structure, categories of info
Navigation design: interaciton with info structure
Graphic design: visual presentation of info and navig

Insome firms specialties in different people.

Design phases:
Discover=assemble info relevvant to project.
Design exploration: explore alternative design approaches:
Refinement=select one approach and iteratively refine it.
Production=create protoyeps and specs hand off to implementers.

Design Phases and Denim: Denim will focus on early phases of design: info design and navig design.

Design implications
1. designers collect and structure large amounts of information; need support for info design.
2. designers produce intermediate artifacts: focus on creastion of artifacts appropriate to design phase: schematics site maps, storyboards
3. Designers sketch to rapidly explore design alternatives: a sketch-based tool will be helpful inearly design; expression more important than precision.

Design of the tool
1. major difference from silk: only one window. at the bottom.
3. left part of tool bar is semanatic slider to view a project.
highest level is site map, logical view of project. text labels of org chart. with arrows.
 Not trying to show all relations, only major
4. Storyboard view: shows rought layout. arrows show more precision and show transitions via hyperlinks.
5. sketch view: single page, all detail visible. not yet graphic design with details of color.

This UI in oriented toward the pen UI.

1. pan and zoom.
2. pencils: gneric pencil for sketching. event pencils for sketching arrows associated with particular events.
3. rubber stamps: for inserting components, built incomponents.

Story boards:
1. transition...



Using components

How to track and store alternatives.

21. IC2D
A drwing program for the visually impaired.

Hesham M. Kamel and James Landay

Lots of R on blind to access graphical data. his tool lets blind people draw.

Example: drawing a side view of a car. Tookhim 13 minutes.

-problems with uis for dwg
-design of ic2d
-future work.

Major problems with drawing UIs.
-guis rely on visual feedback. imagine drawing witht he monitor off: where is the cursor, what's on the screen. how to di get back to wehre i was.
-haptic uis are hard to carry and expensive.

-based on telephone keypad
-two ways to navigate. direction keys or numbes.
-gives feedback via voice. gline users accustomed to screen readers.
-showed example of drawing.

grid recursion
-allows precision of 27x27 cells.
-permits drawing objects at different scale, eg, the right arrow.

Example of drawing a circuit
3x3 grid used for menus, also, to select graphic element, color.
shows navigating.

He prefers using keyboard and physical input rather than speech input. for him it helps to create a better mental map.he does not like giving over to the computer too much of the activity.

Question: using tones to display when he crosses grid lines. to help him navigate. Reply: plans semantic feedback. also can create additional

21. Social Network Analysis.

Elisa Jane Bieneerstock.stanford

Social networks analysis:
1. interdisciplinary method for mesuring cnxns between individuals, using sociology, antrho, math, physics.
2.soch and atnth ask interesting questions.

Social Network Theory
-sociology and SNT do not focus on individuals but on larger groups.
-microsociologigst study interactions or relations between individuals, not process inside indivdual.
-macro socioogists study large societies.
-interest is collection processes traditional in terms of culture, structure.
-socialnetworktheory is an explicit structureal theory whose focus is the micro/macro link: links between individuals, but that person linked to someone else.

1. individual's attitutdes and opportunities are constrained by their social context. context or environment is the social sspace that person occupries.
3. social space explicitly is who they are connected to in wocial world
-who they talk to, like, dislike, influence, boos around, engaged in business with.
4. these conextions are relations.

Measuring relations:
1. 2 useful ways to measure or conecptualize relations that allw math meodelling: graphs, matarices.
2.graphical rep indivisual are rpessesntedas nodes and relations as edges.
3.A==>B implies a relation between a and B. can be directional can have value.
4. For a matrix rep each indivdual is represented as a veactor of possible relations with all others. no relation is represented as a zero. presence of a relation can be directed or valued.

Interpreting relations
1. while absence and presence of relations can be represented in same way, diffrent types need diffrent interpretatins: value and direction, communication and exchange.
2. some network techniques focus on finding the relatons among relations.
3. some focus on individual attributes within the network (centrality)
4. Others focus on the structure poperties of the network itself.

1.Knowing which positions are active or instrumental is one question network analysis addresses. differnet types of centrality to mesure diffrent types of influence.
2.Degree-number of noteds.
3.closeness-ability to get to all nodes in a short path.
4. betweenness-property of lyong on paths linking others.

eg, if a person ins betwenness, might be a bottleneck, might be the most efficient  for distributing messages, if the person disappears, might cleave network.

1. people withsimilar realtional patters are similar: block models, correlation, geometric distance, other clustering techniques.
2. grouping people withsimilar pattersn can be useful.
3. methods are be used to help people to share.

Applications for computing
1. computers allow network researchers to collect better more comprehensive data about computer realtaions thanreal life realtions permit.
2. network analysis proides people intersted in the structure of computer comunities a powerful monitoring too.
3. diffrent cetnrality measures can be used to detiermine if network s are fefficient and where or whey they are not.
4. these techniques can be used to look at the natural use and growth of communities but can also aid to highlllighting where the structure of the computer architecture is causing inefficiency.

Affilitation of twomode networks.
1.  people by people analysis is only one type of work that is done.
2. peple canalso be related to each other thru commonaffiliations.
3.these techniques can be used to classify people by common choices or selections as well as affailitation.

Exchange relations
1. strucute of different types of realtions may be diffrent, so some measures are not appropriate for some types of relations.
2. karen cook et al in 83 found that acentrality is not a good measure for determing who is powerful for certaintypes of exhcanges, eg info changes. when i give info to you i still have it. not so with dollar.
3.there is a whole subfield that looks at measures to summarize characteristics of exchange networks.
4.this is instuructuive becasue osome of those measures may have applications to different relations in computing.'

22.Social Networking on WEb

Danyel  Fisher

Motivation: people spend a lot of time on web in isolation.

Erickson, Dourish.

Spaces: urls, locations dry.
Places; human

Web places, associated conceptually. see other users, chat, follow, hide.

Webplaces implementation.

see WebPlaces.

Social network, connection.

Mapping to places: difficult.  Url+user--->placae.

implicit structure: social networking notation: grpahs of informaton and users.

Uses: lets people find other peole with saeme interest, needs.


Questions: webplaces, ambiguous places, the thridvoice controversy. talk of the net: melissa virus. social networks and mailing lists.

General Discussion

1. what is hcc? composite of hci, cscw, cae
2. how to evalueate, asses, what are the criteria.
3. what role does techgy play?
4. what other apps can we focus on besides email? notetaking? design? education, IR?
5.Process vs practice, task vs. ???
6. different goals of sociologists vs engineers.
7. need for a project.
8. could be the project of ui's for collaboration about collaboaration.


How can we build up relationship between computer sciences and social sciences?

Primary Questions 

1. What is the human-centered component?
2. How can we collaborate more? With whom?
3. How can we relate to industry better.
4. How can we bettter user funding models to foster collaboration?

1. Design
2. Multimodal
3. Sociology and computer science
4. Education
5. Evaluation, assessment criteria
6. Technology, process, practice
7. Concept visualization
8. Concrete projects, terminology
9. Human-machine, human-human interaction

Final Groups
1. Human-machine/human-human interaction: 13
2. Design: 6
3. Multimoda: 7l
4. Sociology, anthropology and computer science and education: 14

Design Breakout

Aaron Marcus
Simi Turek
Flavio Azevedo
Ame Elliott
Arding Hsu
Jimmy Lin
Ted Selker
Hesham Kameßl

1. What is design?

To make things
we use computers, not the computers using us
the design process: going from abstract to concrete
design process involves looking at things form many perspectives.
art is something designed to create a reaction.
science is something presented to ahve people refer to.
design is presenting things to facilitate function.
design is for function for someone
design must always keep customer in mind.
design: must keep user in mind.

2. What is the human-centered component?

User centered, usability, market driven implies fitting into existing structures.
What is central to human beings is searching for new powers, relations?
Designers can be megamaniacs or at the other end completely market driven.
Designers make tradeoff decisions.
Designers  need to understand how the world works.
Designers need to have fresh eyes, to be knowledgable, to understand realistic constraints.
Designers need to be able to think imaginatively, think crazily, create something out of parts.
Desginers are not omniscient.
One major aspect of design is learning.
One major aspect of design is communicating ideas to others.
Designers can design evolutonary or revolutionary.

3. Key research issues

How do I know when a technique is effective?
How do I translate an idea to space, from one space to another?
How do I translate from concept to artifact.
How can I see a simulation of the social impact of my protoype?
How can I design a user experience?, i.e., the feel to the user?
What are the tasks of the design process?
How can the computer support these tasks?
How can I capture the innovation process of the designers?
How can the computer help in the innovation process?
How can computers help manage the design process?
How can computers manage the assets and artifacts, without losing the original?
How can we evaluate the tension between functional design and stylistic design?
How can we measure the merit of functional, style,  of a desgin?
How can we find problems of users? How can we determine technical aspects of the solutions?
How do we solve a problem elegantly?, ie.meets
What is elegance?
How do we determine where we should be on the designer continuum between maniac and victim of marketers?
How do we evaluate designs?
How do we support collaborative design?
How do we manage design process?
How do we avoid least  common denominator results?

4. With whom should they collaborate?

At Stanford, CSLI, is not a good model.
How can we make them be hungry to share resources?
How can we get people to want to share resources and credit?
We need to figure out how to help the sociological sciences to get, run equipment and to fund their students?
How can CS offer "locker space" to less well financed groups?
What else can CS people offer to the Soc people?
Help put in wireless leave behind infrastructure in the Soc areas.

5. What are the funding models for HCC or for human-centered design research?

Consumers fund designs.
Uninterested benefactors support  or fund design.
MIT Media Lab: from partners and partners have no control, neutral territory.
NYU Interactive Technology Program: educational program, little control.
Three-year model: first year is fun, second year, people ask where is the benefit, by third year, must prove results.
Quarter-by-quarter results is most often a business demand.
1.5 years might work for some.
SRI model: integrates things from other departments.
Columbia University dilemma: wound up competing with industry, too close to real applications, not practical issues. results in possible conflict.
One dream of the center: two tiers: one attacks difficult research issues. each industrial partner puts in a contribution. Second level: application specific. each partner takes some from basic level and puts in for specific targeted results.
Al Gore: we must increase funding for human-interaction research: government support.
What is the current financial situation of the HCC?
IBM historically may give about  $1m  to UCB/CS.
Marketing research is important.
Good presentations are important.
Good demos and public relations are important.
Good access important.
Good PR is important.
This culture must be built and maintained without destroying the research value.

Definition of Design
Glossary from Aaron Marcus and Associates, Inc. ,
May be used with citation.
Basic Terminology as of about 15 Feb 99
Aaron Marcus, President, AM+A

User interfaces

UI development process

UI design components

Metaphors, semiotics
User Interfaces: A Definition
The user¹s interactive experience
of command-control over functions
and data conveyed through
human-computer communication

Emphasis on communication

[o Note: One aspect of UIs = Information visualization, so in the discussion below
UI => UI+IV in almost all places, where IV = tables, charts, maps, diagrams]

Steps in UI Development Process:
Define problems or opportunities
Establish objectives and tactics
Set budget, schedule, tasks, resources
Research: Investigate techniques for steps
Examine results of research
Refine criteria for success
Determine primary usability criteria
Define design brief (statement of goals)
Steps in UI Development Process:
Visualize alternative solutions
Select the best design(s)
Prepare documents for implementation
Implement: Build design, complete product
Evaluate: Test results at any UI dev. step
Record history, issues, decisions
Make specifications, guidelines, recomms.
Steps in UI Development Process:
User-interface development cyclical
and partially or completely repetitive

Example: evaluation may be carried out
prior to, during, or after the design step

Example: Separate processes may exist
for each component of UI design

(Metaphor development has same steps!)
Components of UI Design: 1/3
Users: Demographics, experience (from
naive to expert), workflow/productivity
criteria, needs, desires, context, tasks

Subject Matter: Data, functions, tactics, strategies

Technology: Hware, Sware, Net, UI Env.

Metaphors: Ideas conveyed through
terms, images, sounds, movement

Components of UI Design: 2/3
Mental models: Organization of data,
functions, tasks, roles, jobs, people

Navigation: Movement via menus,
dialogue boxes, windows

Appearance: Perceptual characteristics
(visual, acoustic, tactile)

Interaction: Input, status display,
and feedback techniques
Components of UI Design: 3/3
Special data visualization: Charts, maps,
diagrams of structure and process
require metaphor design, also!

Special corporate or product identity:
Specific adjustments of all components

Metaphors, Rhetoric, Semantics,
and Semiotics, 1/2

Rhetorical technique of semantics dimension in the discipline of semiotics

Fundamental part of communication,
often spatial

Component of user-interface design

Essential: can¹t be not-used!

Metaphors, Rhetoric, Semantics,
and Semiotics, 2/2
Semiotics = the science of signs
in communication

Semiotics emerged from linguistics

Semiotics evolved to include all forms
of communication: architecture, film,
literature, painting, music, product design,
even user interface design

Semiotic Dimensions
Lexical = how signs are produced

Syntactic =  how signs co-relate

Semantic =  how signs refer
to other things

Pragmatic =  how signs are used
or consumed

Knowledge Design
Wisdom: significant knowledge
with tested experience

Knowledge: significant patterns of
information with action plan(s)

Information: significant patterns of data

Data: simple perceptual or conceptual input

What Is Visually-Oriented
Knowledge Design?
Information-oriented visual design, or graphic
design, is the systematic use of typography,
symbols, color, and other static and dynamic
graphics, in two, three, or four dimensions,
to communicate knowledge: facts, concepts,
and emotions that enable people to make
effective decisions about their or others¹ lives.

Friday 9 July 99 Breakout reports

Multimodal Breakout: James Landay

Multimodal Research:

-What is 'recognize'?
-What is 'natural'? What would supernatural be?
-What are 'tools'? learning might be hard, but would make you powerful; are tools use once and throw away, or is it changeable.
-What is 'universal.'? People with different abilities should be able to use computers. People with different skills should be able to excell and use some together.

Multimodal Domains

Homes: embedded in products, learning in the home, how we interact with humans
Design: as everyday activity, would multimodal help, eg planning a trip; for real production systems
Communities: collaboration
Factories: sensing


CCAC: involved already by person from UC in new product design, inc Web projects, and they already started new companies!
Industry: projects, teaching. have industry partners teach courses, lead joint projects.

HCC Breakout 2

About the space
How do we work
Establishing credibility
An example project

About the space

Who: profs, grad students, users, visitors
Space models: war room, projects/demos, cubicles/maaahine room; zoo of visistors,users
Attractions: toys, real users, events, reading groups, seminars, free food

How do we work?

Cross-discplinary projects
Short term studies: 3month timescale, milestones with concrete results, class projects to capture new students.
Longterm visions
Novel user group as project seed: insurance, lawyers, doctors, wide range with no relation to univ.

How to foster collaboration: threats, buddy system, seminars, reading groups, intellectual capitalism

How to establish credibility?

Publish, but where: chi, cscs, uist, jperssocpsych
What will UCB/HCC be known for? What is the niche? eg, penbased computing?
Alumni go on to great things: multidisc, collab skills v attractive to employers
Organize metings/workshops and publish proceedings
PR: contests, science fairs


Diffusion of repsonsibility: keep it small, someohe has ownership (first author)
Island mentality: collaborating is too hard; need collab buy-in
Failure to deliver results: establish credibiilty
Students frightened off: career damage?

An example project

Research Q: what happens when interpersonal intearcitons are disembodied?
CS Qs: props-how to design a remote presence for the home? avatars-what are good controls for avatar gestures?
Customer? at+t, users = consuemrs for new home videowall product.
EE questions: required quality of voice commus to preserve emotional content?
Core resources: custom equip, testing set up, user=subjects, experimental data/results

HCC Breakout 4

Researach is centered around people.

Always think about how the techgy helps people.
Issues involve soci engineering, politics, markets, industry and instituions, cscs and HCI, moraltheical issues, sociocultural implications.
Alleviate harm and increase benefits.
focus on a system of people plus tools.
We should focus on social fabric.
CSCW shoves aside political and social-cultural issues in the deployment of technology.
Market forces impact techgy.
Moral and ethical issues would be emphasized at HCC.
Legal issues might be involved.

We were not necess saying that HCC should do soc engineering, but bring critical reflection to the fore.

We need to engage communities with whom we design techgies and bring in critical reflection.

What is important: the practices of the ongoing activities of users, and think about the impact on practices while deploying technology.

Goal: avoid technolgy driven systems design.

Goal: put political, social impact in the foreground of design.

Goal:focus on system of people, with systems of tools.

Innovation: both conceptual and concrete that feed back. not just on technical innovation. we should innovate concepts.

Funding models

Industry collaboration: how can this be done where for industry and acadmy is both good. for academy, conceptual innovation might be the most important. How can we do this?

Alice from UCB: self-reflection important. retreats important. some of us were worried about some terminology that is used. "it's not your father's computer. it's a computer even your mother could use"

How we talk about technology is important.

Breakout 6

CS and Sociology Group

What sociology, antrhopology have to offer.

Technology as status

What do people value?

Sense of identity
Sense of achievement
Sense of mstaery
Sense of importance

Castell, and other writers, recognize increasing imporance of the first.

Next stage of computer: ubiquitous, surrounds us. [we are in the machine. we are the soul of the machine]

Creating a Center:

Find people
Establish a center
Create a movement first
Some evangelism by core peole.

Identify hot areas in other disciplines, offer help, and resources:

Change in work and home lifestyle: Challenge: conducting study of how people live.
Health care, maintenance
Ageing, lifelong learning
Models of learnign
K creation and management
Activity theory: related to learning and practice. Deft = The way you behave is colored by a social environment. Learning takes place through a mentor who scaffolds the learner. learning is always in a social context.

Open Discussion

Identity: Create a t-shirt,domain/web server. don't be too inclusive. be clear; avoid vague terms.

Mision statement:

Put hcc retreat discussion on a website
Send stuff to
Website for HCC =
HCC reading group:

Political and moral issues: understanding vs. prescription. what is appropriate? be careful about political correctness? political and moral issues are a research topic?

Cf. Donald Day's lecture at last IWIPS conference: technology may be good, bad.

Cf. Univ of Toronto: carries on real projects

Legitimatediscourse: don't need to solve all.
We must partner with industry. not be a thinktank servicing industry.

Heart with morality, industry mind says, financial controller want: convert excellent technology into products.

At Siemens collaboration is in three categories:

1. good will, contrib to society.
2. intersting projects. transfer results to siemens
3.start ups

HCC can work through technology. Outcome is new ways of working, new ways of developing technology, new....

HCC provides an opportunity to achieve these goals.

Outreach: can go out to depts, industry, general public.

HCC Long-range Plan

John Cann, ucb

Our reason for being = mission statement:

-to make computing more useful to people.
-to use computing to learn more about people.
(social science studies)
-computing is permeating our daily lkives, and becoming our environment.

(he reads post modern philosophy. moving into the way we behave is increasingly impacted by brands. we are becoming more virtual in our comm and society. we confuse physical with vr. physical has been partly vr for some time).

The HCC model

-we want to have social sciences participate as primary partners.(we want critical dialogue withthem aobut unintended consequenceas of techgy. cs people think vertically. ss people think horizontally)
-interdisciplinary collaboration is hard, and happens sporadically by iteself.
-collaboaraton happens much more often when facilitated by: shared space or proximity, reg face2face contact, shared funding.

What we want to deliver
-HCC provides resources and seeks to generate:
-focused researche projects involving collab between 2-7 faculty (nsf, industry, etc. funded).
-industry partnerships in those projects.
-opportunities for graduate students: contact with or co-advising by faculty acroos departments. contacts with industry.

Three projects in the pipeline

1. Being there: (2cs, 2ee, 2psychologists). expansion of props work.john canny,  jerry mendolsohn, and ...keltner of nonverbal comm. will participate. how to use nonverbal cues and use compression to get the right things there with a savvy tech system. need to learn what's most imiportant and effective. will dissect and study nonvebal cues. will try to interpret facial expression and ship it to props or avatars.

1. tacit knowledge mining (2 cs, 2 sociologist, 1 anthropologist). came out of discussions with ucb bus school. how to understand tacit knowledge. illka says: 90% of K in a corp is tacit, not explicit. computing does not do much with that now. hard to get data out of people's heads. maybe analysis of practice can extract it. team will use several disciplines.jim lincoln is sociologist is in the bus school. project is organized around corp practice.

3. knowledgescapes (2cs, 2 socio, 1 sims): how to improve the document space by organizing it not by content, but do an org according to info needs and quereies of users to cluster documents. some evidence shows that this helps give info to people. The other question: web demographics. one of the social difficulties in most civilized countries: disadvantaged neighborhoods are geo and social network disadvantage. people do noe know others in other communities. this restricts their mobility. as people rely more on info from the web for shopping and getting services, will we see similar patterns. will there be a closeness of their exploration of the web. they will do demographic sampling of certain sites with known demographics.

These projects have $1m research, so they can do some progress.

Lessons from the GVU

GVU is similar. John Canny talked with the current director. about history. Jim Foley founded it. Jack Rosniak is the current director. Jim got a big grant and put it into creating the center. created a board. started pr and a web site. jim delegated internal activities. jim concentrated on industrial outreach. john thinks similar is correct.

-Create a core group of faculty and industry people with specific responsibilities.

-Hire some staff people to manage daytoday functioning of the center.

-create partnerships with industry.

-go after government support.

-create shared labs.

GVU succeeded. About that time, GIT declared it a formal center and gave about $100m as an official center.

The Bekeley Brand

-Balance and strength in SS and CS.

-Of the 30 main graduate research disciplines, UCB has depts in the top 10 in 28. according to NCR. Even more that the private schools, has more in top 10 than any univ in usa .

-Strength in Theory and Formalism.

Alice: What about education and design?

Next Semester

-HCC seminar continues: speakers fromcs, social scienes, academic and industry.

-New grad corse onHCC covers:
1.presence and non-verbal comm
2.socialnetworktheory and paplicaions
3.tacit knowledge
4.activity theory, social learning.

How can you help?

-UCB: be involved in semianr, retreasts
-Direct support at center level.

GVU started with $200K, HCC does not have this.

HCC does not have discretionary industrial funds. Need bootstrapping funds.

-Partnerships: focussed projects, fixed duration. IP contracats need to be figured out.

-Steering committee or IAB=industry advisory board.

Alice: Anonther funding source: digital library. she has two grants now. Gore wants to give: $150M for  national digital library.

Final Feedback

Interval, Bill Verplanck: Personal and professonal motives. Interested in design and sketching. His Interval interests: cable companies are being bought up. broadband into the home.  Viewing on demand. how to watch 500 channels. convergence of cable and pc and movies and tv and web. alternative telephone channel to the home. computer mediated communication. focus is: not on how to hookup the networks, but what people do. looking at the future of remote controls. is the remote, or the couch, or the whole room. how to lean forward, fly. when are you engaged in the control. CSCW has work in it. we need CSCPlay, entertainment, mobile. What happens when you do not think about the computer as a computer but a bunch of stuff.

Pixar, Karen Weber. Impressed by diversity, different disciplines. passions. Another motive: goal of sharing her knowledge and experience. would like to continue her relationship. One of her concerns: HCC heard a lot about C anc C but not about humans. Need real people to think out of the box. Other research models have fought against the tension between people and machines. Cauation: don't become an MIT media lab. they give good demos, but don't think about people using technology. don't become just a good demo center. You need to study people and how they work. Then how to apply the technology. Tension in academic departments, better to be inclusive, but maybe with a different model. another approach: when people visit the center, do they come with the knowledge, they are just a team member. are the grad students going to be just cs people or a new brand of academic.

FxPal,Elizabeth C.: Some people talk about mission, pr, branding, putting people first. She also says, don't become a demo center. Should focus on a particular project. Think about the benefits and actual activities. Industrial partners: difficult, but possible. Interns, seminars in the eve are good. IP should be considered.

XeroxPARC, Paul Dourish: set up situations which encourage debate. Stat a movment, domain, a buzz. Look for a small focussed collab projects. eg grads help each other. Share the data and observations. need for new ways of thinking about collab, about working together.

Nokia, Illka: does not yet know the berkeley system yet. sees people trying to go into a new area. computers are embedded in social life. meaning, not info in the future. meaning processors should be thought about. we have macro level topics: new social patterns, languages, trends. these people must be found. proplem is not technology, but social organization, change processes, power structures, how to communicatenew ways of doing things so they are understandable, how to package innovatoin and market innovation. re MIT: demo or die. you must have concrete artifacts to concentrate discussion. demos seem often trivial. if you compare this to deploying it in real life, the deployment is more challenging. the hard and difficult  task: understand regulation and laws, how people use it. they have support for competence management  of skills in orgs at nokia. but there are different laws in different countries re collecting data. this impacts more than techgy. example: collecting id info in finland is ok. in usa is a hot topic for debate. big topic: rethinking new architectures of computers, not digital. thinking how you can deploycognitive artifacts in society: emotional, material aspects are more important.  new: knowledgemanagement. access is not imporant in future: relevance is imporatant. in future: avoidance of info is important. No one understands these concepts well. Must be able to communicate model for how this HCC will happen. operational model, funding, etc. must comm this clearly.

Inverval Mark S: skill sets issue: cold/soft science. What about design, art. Look at very complex problems and try to find a solution by intuition. i am concerned that more attention is paid to this group. CCAC has people like that. There is an academic instinct to debate too long. designers can make useful prototypes.

Siemens, Arding Hsu: good chemistry here. re research:  he spent 15 y in siemens central research for worldwide multimedia research. business center in berkeley: CIA. want to marry comm, info, and automation together. if you want good comm, you need broadband. you'll see many needs in factories. in factories, you see big networks, processes. this is a big future for HCC development. Another direction: bridge physical world and virtual world. Bottom business line: no money, no business. How can we help hcc get some $ to get to the right place. In his earlier life, he was liaison for many collaborative projects.  He knows the 3y cycle. he suggests: must come out with a 3yr financial model. can divide govt and industrial funding. maybe 50% govt at first. For govt funding can look at the big picture. for industry, must be focussed. He has $500K for berkeley. he needs a quick success story. needs berkeley help with success story, then he can bring in more $. He has talked with Siemens semiconductor people.

Glad to see HCC arising.
Depends on leadership
Exampes: UCB,Yale interdisciplanary, EWC; interdisciplinary
Lack of design input.
Good demos important for both conceptual analysis and presentations.
Need to look at history, other cultures, semiotics, visible language, design

Mr. Aaron Marcus, President
AM+A California Tel: 510-601-0994, Ext 19
Fax: 510-547-6125
AM+A New YorkTel: 212-625-1813
Intelligent Design:
User Interfaces and Information Visualization
Our WebWord interview:

Our paper about user-interface design for the Web: