Transcript CMC - people.csail.mit.edu

16.499: cscw
Computer-Mediated
Communication
Michael Bernstein
14 February 2007
CMC
2
Cognotor
3
Meeting
Central
4
CMC
5
Computer-Mediated
Communcation
Studies “the social effects of
different computer-supported
communication technologies.”
[Wikipedia]
6
Computer-Mediated
Communcation
The Social Effects of
Communication that relies on
Technology
7
Examples
Text Chat
Video Chat
Audio Chat
Group Chat
E-mail
BBS
Communication in online games,
MUDs
8
Why is it difficult?
Edward: [laughs]
Michael: …huh?
Edward: OK, gotta go!
Michael: You are so weird.
[click]
9
Why is it difficult?
Different communication systems
have different affordances.
These affordances have direct
effects on what is or isn’t
social acceptable.
How do we understand what
design decisions will carry
what meaning?
10
Cognotor
11
What was Cognotor?
A tool for facilitation of plan
and outline creation – a
brainstorming device
Support for both private and
public workspaces
Emulate a whiteboard through
networked computers and a
shared display
12
How Cognotor Worked
13
Design Decisions
Separate screens for each user
Private edit window – nobody
can see items you’re trying to
post to their screen until you
finish
Anonymous contributions,
anonymous changes
14
Problems
Group 1:
gave up, used paper
individual work without
communication
Group 2: stopped using input and
just shared video streams
Users wanted to see things in the
workspace that were currently
invisible
Users mistook each others’ spoken
15
Parcel-post Communication
Model
“I want a
Hamburger.”
Encoding
001101011010
1010101010111
101010100011
Decoding
Stacey desires a
hamburger
16
Parcel-post Communication
Model
Information comes in discrete
units, like packages
Each parcel will be understood
There is a queue of unseen
parcels
17
Interactive Model of
Communcation
Conversation is a highly coordinated
activity…
- Continuous
- Requires frequent feedback
…in which meaning is attained by a
variety of mechanisms…
- Multiple channels of information
…that have context dependent
functions.
- Each move projects onto the next
one
18
- Midcourse corrections
Problems identified
Parcel-post model impeded the
interactivity of the
conversation
Significant delay times
Loss of context with anonymous
contributions and unannounced
changes
Individual screens mean less
visual context and mutual
19
engagement
Fixes
Shared editing
Shape and positions consistent
across machines
20
Meeting
Central
21
The Problem
Distributed meetings continue
to have difficulties
22
User research
1,752 Sun employees responded
to an internal survey
Audio problems: cannot hear
people, poor audio quality,
extraneous noise
Behavior problems: speakers do
not check for understanding
Technical problems: difficult
to identify speaker, not
everyone could view materials,
23
no way to tell who was in the
Meeting Central: The
Facilitator
Speaking
Invited but
not in
attendance
Hand raising,
to indicate
desire to
speak
Passive
feedback
indicators
Indicate
support or
confusion
(back
channel)
Whisper
feature
24
Later work: MC 2.0
Renaming
the
whisper
feature to
“voice
chat”
[CHI 2005: posters]
25
16.499: cscw
Discussion
[email protected]
Thanks to Scott Klemmer, Ron Yeh and Stanford