4 mackroom & grgsqr - glamatthew/hci4papers/04mackroomandgrgsqr.pdf · • contextual resource...
TRANSCRIPT
UbiComp
• Real systems in the real world
• Supporting people with real tasks
• 2 example systems
• Context-aware
• Content types
• Supporting synchronous and asynchronous collaboration
2
Sensing
• UbiComp systems are often utilised in a large number of differing situations
• mobile phones are used at home, work, in the street, in the car... and not just for making phone calls
• To respond appropriately devices and systems must be aware of what situation they are in, and what resources are available... context-aware
3
One challenge of mobile distributed computing is to exploit the changing environment with a new class of applications that are aware of the context in which they are run. Such context-aware software adapts according to the location of use, the collection of nearby people, hosts, and accessible devices, as well as to changes to such things over time. A system with these capabilities can examine the computing environment and react to changes to the environment.
Schilit et. al., 1994
Context-aware
4
Context-aware Systems• Contextual sensing
• Detect environment and state changes, feeding back to user
• Contextual adaptation
• Alter behaviour based on what is sensed
• Contextual resource discovery
• Detect peers and resources, and using them
• Contextual augmentation
• Add additional information to elements already in the environment
Pascoe, 1998
5
Context
There are certain types of context that are, in practice, more important than others. These are location, identity, activity and time.
Dey & Abowd, 1999
6
Devices may have direct or indirect awareness of context. In the case of indirect awareness, the entire sensing process occurs in the infrastructure while the mobile device obtains its context by means of communication. In contrast, a device has direct awareness if it is able to obtain context autonomously, (more or less) independent of any infrastructure.
Gellersen et al., 2002
7
Detecting Context
• Direct awareness
• On device, built-in
• Movement (accelerometers), peers (Bluetooth), GPS (more or less)
• Modern phone (Nokia E90)
• BT, GPS, WiFi, 3G, microphone, camera, infrared, accelerometers
• Indirect awareness
• Asking another device/server
• Weather, WatchBot
8
Context Sensing
• Objective and Contemporaneous
• Dey and Abowd
• Schilit et al.
• Relative and Historical
• Chalmers
• Dourish
9
Lighthouse / Mack Room
• Ethnographic studies of tourists (Galani, Chalmers, Brown)
• 3 co-visitors synchronously exploring a Charles Rennie Mackintosh museum
• Physical, virtual, web
• Shared awareness supporting articulation work (exploring the exhibits)
• Mixed Reality
10
Museum Visitors Study
• Four key resources for visitors
• Voice
• Location
• Orientation
• Common space
11
Technology overview
• Ultrasonic positioning
• EQUIP / Auld Linky
• VoIP
• 802.11b
• Mobile devices always slower
12
Technology• EQUIP
• Shared tuple-space
• Handles periodic dropouts
• Dealt with communication
• Auld Linky
• Delivers different data and data types to different clients
• Adapted to client display/abilities
13
On-site visitor
• PDA with map
• Headset for VoIP
• Ultrasonic positioning
• Compass
• Mixed reality
14
Virtual visitor
• Laptop/desktop
• Virtual environment
• VoIP
• Avatars represent self and others
• Cut-down web view for exhibit detail (text) automatically pops-up
16
System Use
• Navigate museum
• Collaborate around objects
• Share experiences
• More sociable than a “normal” museum visit
17
• Continual awareness of what others are doing in the museum space permits a free-flowing and smooth discussion of the exhibits as all the participants wander around the museum
• Location & orientation
• Where users were and what they were looking at
• Voice
• Shared information space
Awareness
18
Positioning in Mack Room
• Ultrasonic
• 8 transmitters in walls and on ceiling (hidden/invisible)
• User’s device emits chirps, units in walls reply
• Measures flight times of chirps
• 50% accuracy at 0.52m, 95% at 1.83m
• Standard deviation of 1.29m19
Indoor Positioning
• Mainly ultrasonic
• Loooooooooong setup
• RF tags can be quicker but less accurate
• More to come (next lecture)
20
Time of visits
• Mainly only synchronous support
• All visitors must start and end at same time
• Not how people visit, even in trials this was apparent
21
Problems• No peer discovery
• Had to enter IPs statically
• Centralised, single point of failure
• Required long setup for ultrasonics
• Tied to location
• Content had to be manually input
• Result... could not be easily deployed, not mobile
22
The Content Problem
• All information hard-coded into Auld Linky
• Question of expertise
• Information from visits not recorded
• All users always see the same data
23
Content Types for P2P• Static: content is static, always the same
• Mack Room
• External: data changes but is retrieved from external server
• Weather reports, stocks, etc
• Reliance-free: does not require any data
• Simple games, IM clients
• Self-Generating: P2P system continually generates content, often by user contributions
• George Square...
24
Lighthouse Lessons
• Support users’ changing roles
• Awareness of others and what they do
• Interface
• Support fluid group dynamics
• Late joiners (synchronous & asynchronous)
• Avoid pre-authored content
25
George Square• Attempt to create a more mobile system
• One that could be used anywhere (rather than in a single museum)
• Supporting fluid groups
• That could be set up quickly
• That would continue to generate new content, rather than failing to adapt to changes and delivering the same content to all users regardless
26
George Square
• Any number of visitors (fluid groups, including single user)
• Can visit at any time (support sync & async)
• Greater shared awareness for smoother collaboration
• Self-generated content
27
Technology
• EQUIP + basic peer discovery
• Recer: recommendations, filtering
• Tablet PCs with USB cameras
• 802.11b/g
• GPS
• VoIP
28
George Square
• Ethnographic studies of tourists
• Awareness can overcome problems
• Errors and inaccuracies in positioning and comms networks often revealed and resolved through talk.
• Highly context-dependent, e.g., sensitivity to GPS error or inaccuracy when a person appeared to be in the road, and insensitivity when still ‘near enough’ to a statue being talked about.
29
1. Location of users2. Photographs and location3. Web & location recommendations4. Picture recommendation5. Own and others’ recommendations
32
Use of System• Sociable interaction
• Tourism and system involves both co-operative and articulation work
• Collaborative photography (articulation work) when icon near statues...
• “I’ll take a picture of the stone lions for you. They’re very majestic”
• Excellent collaboration
• Users assumed roles
• Taking pictures
• Looking up web information
• Would feed back to later users33
Content in GrgSqr
• Recorded as visitors use the system
• Logged and shared
• Users continually contributing to content, system generates its own content
• Always up-to-date (as long as people use it)
34
Self-Generating Content
• Bootstrap problem
• Solutions
• Interesting experience without content
• Expert routes
• Field experts
• Written Guides
35
While location-based information systems for, e.g., tourists, are well known, these systems usually rely on predefining all available information. This reliance on special purpose authored content is unfortunate, as it puts the onus of creating content on the maintainers of the system rather than on its users or the Web in general. Furthermore, users should certainly be allowed to add their own material and observations to a system.
Hansen et al., 2004
36
Opening up a system to user contribution holds the promise of content that is much more informal, opinionated, and even more subversive than content provided by an official source... Notes contributed by an unofficial source such as students or other insiders were valued more than official factual notes that were posted. The content became qualitatively unique and was well-received when it was created by other users.
Burrell et al., 2002
37
Pre & Post visit
• Support for asynchronous behaviour
• Can plan visit and path before arriving
• Experiment with what you could have seen
• Auto-generated blog-like report
38
Using the past
• Continual reinterpretation, relative and historical (Dourish & Chalmers)
• Filtering
• Recommendations
40
Recer• Records activity, location, identity
and time from all users
• Identifies recent activity... or context
• Generates recommendations by looking for similarities in context from previous activity of all users. Focuses data.
• Most interesting information becomes highly recommended
• Based on social use of system
41
Roles
• Supporting for fluid groups
• Disconnection and reconnection
• Users could...
• Stop and start easily
• Change roles
• Change machines
• Change locations
43
Adaptive Data
• Context recorded, logged and employed
• No two visits the same
• Visit customised for each visitor
• System learned new data from all visitors
44
...since state is stored on the server clients have no support for continued operation upon disconnection from the network which may be fairly frequent in a wireless setting.
Mitchell et al. , 2003
45
P2P System Types
• Centralised: peers rely on external server to hold data and/or mediate comms
• Hybrid: peers decide or take turns on which will perform additional tasks
• Pure: all peers are always equal
• More later...
Triantafillou et al., 2003
46
In some sense, an ad hoc mobile information system is the ultimate peer-to-peer system. It is self-organizing, fully decentralized, and highly dynamic. However, current peer-to-peer systems[...] are designed for stationary hosts connected to an Internet-like infrastructure.
• Self-organizing: the topology of a mobile peer-to-peer system constantly adjusts itself by discovering new communication links.
• Decentralized: each peer in a mobile peer-to-peer system is equally important and no central node exists.
• Highly dynamic: since communication end-points can move frequently and independently of on another, mobile peer-to-peer systems are highly dynamic.
Korteum et al., 2001
47
Problems
• Problems with GPS availability (more later)
• Still not fully mobile
• 802.11 network issues/connectivity
• Hybrid system
48
Summary• Think back to CSCW core concepts
• Context-aware systems
• Content types
• P2P types
• A lot of important items are in bold (but not all)
• Read George Square paper
• Mack Room paper optional
49
References
• References on next slides are not examinable and you are not expected to read any if you do not want to
• But can help give more detail and make things clearer, so you may want to read one or two if you find something interesting or difficult
• http://portal.acm.org/dl.cfm
• only works on campus
50
Schilit, B. N., Adams, N. I., and Want, R., "Context-Aware Computing Applications," in Workshop on Mobile Computing Systems and Applications Santa Cruz, CA, USA: IEEE Computer Society, 1994, pp. 85-90.
Pascoe, J., "Adding Generic Contextual Capabilities to Wearable Computers," in 2nd IEEE International Symposium on Wearable Computers: IEEE Computer Society, 1998, pp. 92-99.
Dey, A. K. and Abowd, G., "Support for Adapting Applications and Interfaces to Context," in Multiple Users Interfaces: Engineering and Application Frameworks, Seffah, A., et al., Eds.: John Wiley and Sons, 2003.
Burrell, J., Gay, G. K., Kubo, K., and Farina, N., "Context-Aware Computing: A Test Case," in UbiComp, Gothenburg, Sweden, 2002, pp. 1-15.
51
Gellersen, H. W., Schmidt, A., and Beigl, M., "Multi-sensor context-awareness in mobile devices and smart artifacts," Mob. Netw. Appl., vol. 7, pp. 341--351, 2002.
Hansen, F. A., Bouvin, N. O., Christensen, B. G., Gronbaek, K., Pedersen, T. B., and Gagach, J., "Integrating the Web and the World: Contextual Trails on the Move," in Hypertext, Santa Cruz, USA, 2004, pp. 98–107.
Triantafillou, P., Ntarmos, N., Nikoletseas, S., and Spirakis, P., "NanoPeer Networks and P2P Worlds," in 3rd IEEE International Conference on Peer-to-Peer Computing, 2003, p. 40ff.
Kortuem, G., Schneider, J., Preuitt, D., Thompson, T. G. C., Fickas, S., and Segall, Z., "When Peer-to-Peer comes Face-to-Face: Collaborative Peer-to-Peer Computing in Mobile Ad hoc Networks," in International Conference on Peer-to-Peer Computing, Linköpings, Sweden, 2001.
52