Lecture

1
Lecture 4 Communication and Collaboration(Preec
e 4)
2
What Did You Learn Last Week?

• Which heuristic did you find to be the easiest to
  apply? Which yielded the most useful results?
• What are the three questions that one asks in
  each step of a cognitive walkthrough?
• Will you find all of the usability problems with
  a given software system by applying a heuristic
  evaluation or cognitive walkthrough?
• What are the advantages and disadvantages of
  applying the heuristic evaluation and cognitive
  walkthrough versus running a usability study?

3
Why Study Human Collaboration Communication?

• Traditional HCI studies looked at individuals
  focusing on individual tasks
• However, many modern computer applications
  support collaboration
• Synchronous electronic whiteboards, NetMeeting,
  CyberChair, Chatrooms, games
• Asynchronous e-mail, WebCT, CyberChair,
  PowerPoint, Word
• Studies of collaboration and communication can
  provide theoretical and empirical foundation for
  design and analysis of human-computer interaction
  for groups

4
Body of Relevant Studies

• Conversation analysis (e.g., Sacks, Schegloff,
  Jefferson, 1978 Suchman, 1987)
• Studies of groupwork (e.g., Tang, 1989)
• Comparative studies of FTF and technologically-med
  iated communication (e.g., Isaacs Tang, 1993
  Doerry, 1995)

5
Lecture Overview

• Part I Conversation
• Part II Coordination
• Part III Awareness
• Part IV Theoretical Frameworks

6
Part I ConversationConversation
MechanismsFTF vs. TM CommunicationDesign
Implications
7
Conversations Follow Rules

• Are there rules that people follow when having
  conversations?
• Conversation analysis (Sacks, Schegloff,
  Jefferson, 1978) suggests there are
• Mutual greetings
• A Hi there
• B Hi!
• C Hi
• A All right?
• C Good, Hows it going?
• A Fine, how are you?
• C OK
• B So-so. Hows life treating you?

8
Turn-Taking Rules

• Turn-taking to coordinate conversation
• A Shall we meet at 8?
• B Um, can we meet a bit later?
• A Shall we meet at 8?
• B Wow, look at him?
• A Yes what a funny hairdo!
• B Um, can we meet a bit later?
• Turn-taking rules
• 1 current speaker chooses next speaker
  explicitly
• 2 another person decides to start speaking
• 3 current speaker continues talking
• Negotiation via verbal cues, gesture, gaze,
  prosaic effects, body language

9
Communication Breakdown

• Communication breakdown
• Natural language is inherently ambiguous, and
  relies heavily on context
• Shared meaning must be socially negotiated
• Referents like it, that, one, etc. are
  contextual
• Communication breakdown occurs when people lose
  shared understanding
• Types of breakdown (Suchman, 1987)
• False alarm
• Person A believes she has been misunderstood, or
  believes she misunderstands, when, in fact, all
  is well
• Garden path
• Person A and Person B proceed normally with
  conversation, thinking that all is well. However,
  they have different understandings. They dont
  recognize the discrepancy.

10
Example Script False Alarm

• Topic Asking someone on a date to a movie
• Can you think of a dialog that would illustrate a
  false alarm in this situation?
• Lets write a script collaboratively and act it
  out

11
A Possible False Alarm Script

• (B Boy, G Girl)
• B Um, are you busy Friday night?
• G Im not sure
• B Are you busy Saturday?
• G Um, only at night
• B Well, Ill be seein ya then
• G (after boy has left) That guys so cute. I
  wish he had asked me on a date. ?

12
Example Script Garden Path

• Topic Going to Greece
• Can you think of a dialog on this topic that
  would illustrate a garden path?
• Lets write a script collaboratively and act it
  out

13
A Possible Garden Path Script

• (H Husband, W Wife)
• W Honey, Ive really been wantin to go to
  Grease. When are you gonna take me?
• H How many times do I have to tell ya that its
  too expensive. Im not gonna take you.
• W I just dont get it. Arent I worth it? What
  good is all the money we earn if we cant do
  things like go to Grease?
• (6 months later, after divorce)
• W I just dont understand why he couldnt spend
  40 on a tickets to Grease, my favorite musical
  which finally came to town . ?

14
Repairing Communication Breakdown

• Humans are quite skilled at detecting
  communication breakdown
• Once detected, communication breakdown can be
  repaired in several ways (Initiating a repair)
• Nonverbal communication (frowning, puzzled looks,
  pointing, gazing)
• Request clarification
• Repeat what was said with better intonation
• Rephrase what was said
• How could the two scenarios described before have
  been repaired?

15
Research into ConversationsAnalysis Framework

• Viewing a human-computer interaction as a
  conversation can lead to profound insights
  (Suchman, 1987)
• Suchman suggested the following framework for
  analysis

16
Example of AnalysisInterface Under Test
Ventri-cle
Valved pipe
Compass Tool
T-pipe
Straight pipe
Elbow
Muscle
Cardio-vascular system under construction
17
Example of Analysis Transcripts

• Transcript 1 Trouble setting beats
• Transcript 2 Trouble attaching gauge

18
Research into Conversations Implications for
Design

• Design computer programs with conversational
  interfaces? (NOT!)
• Except in highly constrained scenarios,
  conversational interfaces will fail
• Computers view of the world extremely
  impoverished cant access context
• As soon as computational artifacts demonstrate
  some evidence of recognizably human abilities, we
  are inclined to endow them with the rest
  (Suchman, 1987, p. 14)
• ? Misconceptions are bound to arise

19
Research into Conversations Implications for
Design (cont.)

• Design computer software that anticipates
  communication breakdown
• Communication breakdown can be recast as an HCI
  problem (human and computer as conversational
  participants)
• Breakdown occurs when design model diverges from
  user model
• As designers, we can anticipate where such
  divergences may occur
• E.g., in a drawing editor, user clicks
  frantically on drawing canvas, trying to select
  an object, but current tool is not select tool.
  Designer might anticipate such a mode
  breakdown, and present a clarifying dialog box
• E.g., explain why a menu item isnt available

20
Technologically-Mediated Conversation Synchronous

• Numerous software systems have been developed to
  support human conversations
• Synchronous TMC
• Conversations are supported in real-time through
  voice and/or typing
• Examples IM, video conferencing and chatrooms
• Possible Benefits?
• Can keep more informed of what is going on
• Video conferencing allows everyone to see each
  other providing some support for non-verbal
  communication
• Chatrooms can provide a forum for shy people to
  talk more
• Possible Drawbacks?
• Poor video quality
• Difficult to establish eye contact
• Bad behavior behind mask of avatar

21
Technologically-Mediated Conversation
Asynchronous

• Communication takes place remotely at different
  times
• Examples Email, newsgroups
• Benefits?
• Read any place any time
• Flexible
• Powerful, can send to many people
• Can make saying things easier
• Drawbacks?
• FLAMING!!!
• Spamming
• Message overload
• False expectations as to when people will reply

22
Face to Face vs. Technologically-Mediated
Conversation How Do They Compare?

• How would you measure communication
  effectiveness?
• In terms of you definition of communication
  effectiveness, do you think talking face to face
  or via technology is better?
• Why?

23
Face to Face vs. Technologically-Mediated
Conversation Research

• Numerous studies have examined differences
  between synchronous FTF and technologically-media
  ted communication
• Audio-only
• Audio video
• Assumption Bandwidth matters
• Persons connected via high-quality audio and
  video link can converse and perform tasks just as
  well as they can face-to-face

24
Doerry, 1995 Study

• Punchline Bandwidth doesnt matter!
• Transcribed dialogues of pairs of users
  performing shared workspace tasks in 3
  environments
• (1) face-to-face vs. (2) high quality audio link
  vs. (3) high quality audio-video link
• Analyzed dialogues using a systematic
  classification of communication breakdowns
• Topic
• Turn-taking
• Reference
• Results support that of similar studies
• FTF communication has significantly fewer
  breakdowns than either technologically-mediated
  environment
• video link provides no additional advantage

25
Conversations Design Implications (cont.)

• Higher bandwidth isnt necessarily better
• FTF conversation vastly superior to TM
  conversation
• Acknowledge the difference!
• Can anything be done to make TM conversation
  better?
• Peripheral awareness mechanisms
• Mechanisms for establishing eye contact

26
For discussion Will video mobile phones be
successful?

• Will the judder, sudden jerks and
  shadowsdisappear?
• Will it be possible to establish eye
  contactand read lips on such a small image?
• Will people find it socially acceptable totalk
  to an image of someone in the palm oftheir
  hands?
• What do you think?

27
Part II CoordinationCoordination
MechanismsDesign Implications
28
Coordination Mechanisms

• Coordination required when group of people work
  together toward common goal
• Coordination mechanisms
• Talking
• Used in service of orchestrating joint activity
• Gestures, nods, glances
• Can even form formal language, as in orchestra
  hand signals
• Schedules
• e.g., university lecture schedule
• Rules conventions
• e.g., throw away trash, keep quiet in library,
  drive on right Normans cultural constraints

29
Coordination Mechanisms (cont.)

• Shared representations
• Examples
• Calendars
• Task lists
• Documents (e.g., scholarly papers)
• Source code
• Concept maps
• Enable group members to
• Make inferences about group progress and status
• Share and refine ideas
• Develop a common perspective on a problem

30
Coordination Mechanisms (cont.)

• Example empirical studies
• Rogers, 1993
• Conventions imposed by software arent always
  followed people may see them as a burden if
  additional work is required (productive
  laziness)
• Belotti Rogers, 1997
• Examined role of shared reps in collaborative
  work
• Found that information represented on-line often
  becomes re-represented physically On-line
  version is lost or overlooked

31
Coordination MechanismsDesign Implications

• Dont assume that people will want to use
  software to coordinate collaborative work
• Extra work involved may inhibit use
• Dont assume that computerizing external
  representations will help
• There needs to be easy means of accessing,
  updating, and continuously reviewing shared
  computer documents
• Careful empirical studies of group practices in
  target domain are absolutely essential!

32
Part III AwarenessAwareness MechanismsDesign
Implications
33
Awareness Mechanisms

• Awareness means knowing whos around and what
  theyre up to
• Awareness facilitated by
• ability to move around in space
• knowledge of socially acceptable behavior
• peripheral vision
• Perception of other peoples cues
• Door slightly ajar vs. door closed
• Person makes eye contact vs. looks away
• overhearing and overseeing - allows tracking of
  what others are doing without explicit cues
• Awareness mechanisms in IM programs?

34
Awareness Mechanisms (cont.)

• Empirical Study (Heath Luff, 1992)
• Studied control room operators in the London
  Underground
• Controller A responsible for movement of trains
  through station
• Controller B responsible for announcements to
  passengers
• Controller B often initiated announcements by
  eavesdropping and anticipating consequencesno
  direct communication was involved in these cases

35
Awareness MechanismsDesign Implications

• People use awareness of others to act in the
  world
• How can technology support awareness in teams
  whose members are distributed?
• Peripheral monitoring Portholes (Dourish Bly,
  1992)
• Show snapshots of people as they work at their
  desks
• Clicking on snapshot brings up a menu of options,
  including e-mail the person)
• Informal study showed that Portholes led to
  shared sense of community

36
Portholes (Xerox PARC)
Regularly updated digitized images of people in
their offices appeared on everyones desktop
machines throughout day and night
37
Awareness MechanismsDesign Implications (cont.)

• Explicit Notification
• Enable users to notify each other about certain
  events
• Automatically notify users of events of interest
  (e.g., status of shared tasks)
• Example Yahoo IM

38
Awareness MechanismsDesign Implications (cont.)

• Create a virtual space for non co-located people
• Goal make people feel as if they are in the same
  virtual place even though they are in physically
  different spaces
• Extremely difficult to support peripheral
  awareness and mutual gaze
• Example Hypermirror (Morikawa and Maesako, 1998)
• See next slide

39
Awareness MechanismsDesign Implications (cont.)
(woman in white sweater is in a different room
from the other three)
People in different places are superimposedon
the same screento make them appear as if in same
space
40
Awareness MechanismsDesign Implications (cont.)
How its done
Two in this room are invadingthe virtual
personal spaceof the other person by appearing
to bephysically on top of them
Two in the room move apart to allow person in
other space more virtual personal space
41
Summary Points

• Various social mechanisms (e.g., turn-taking,
  coordination, awareness) enable us to communicate
  and collaborate
• Empirical studies of these mechanisms have
  important implications for how to design
  technology to support communication and
  collaboration