Last: DIS2004: reflection
Next: DIS2004: design for hackability
August 02, 2004
DIS2004: ubicomp at home and on the move
ubicomp at home and on the move
the 'missing what' of Brand's model - social interaction
ecological habitats of devices
activity centres - where devices are used is often different to where they live
coordinate displays - places such as where mail is distributed, a noticeboard
produce a space-plan showing all 3 on a physical plan. Shows assembly and manipulation of stuff at functional sites
Two parts to ubicomp in the home: placement and assembly (reassembly). 'Piecemeal' migration (Edwards and Grinter). Allows stuff to assimilate and disappear from view (already happened with TVs, telephones)
Configuring ubistuff:
compositional approach -
physical to digital transformers
digital to physical transformers
digital transformers
emerging physical and digital spaces
a basic interaction mechanism - the jigsaw editor - only services that are compatible can be fitted together
can people understand this abstraction? is it intuitive and efficient?
cooperative analysis carried out using paper-based mockups and seed scenarios (Ehn & Kyng 1991), married with situated evaluation (Twidale et al 1994)
3 criteria of analysis (morgensen 1994)
users provided ideas for new devices, components and applications
Broader lessons:
Inhabitants as designers and developers
Reasoning with diverse elements
Interleaving the new and the old
Linking outside the home
****
Gate Reminder
Samsung Electronics
Reminds user of objects they should take when going outside
Reminds users of messages they should be aware of before leaving
Installed at the front door
Based on Smart Home project from 2002
scenarios - books you borrowed from a friend, DVD rentals you should return, take out garbage on a particular day, forgot to buy bread even they passed it several times
Don Norman - Signal + message
but, should be signal + message + context
the right moment (person/time/place/atmosphere)
(don't remind about bread when in a meeting at 3pm)
two types of contextual-aware reminder:
mobile reminder - carried by a user
ubiquitous reminder - installed into your surroundings
front door - right place
before going outside - right time
user identification - right person
3 common reminding methods - post-it on a door, place the item by the door, ask someone to remind you
potential needs -
missing objects
weather information
regular reminding
writing reminders from outside
RFID used, tags attached to objects (medium-range type - 1m)
User detection - motion detector, door sensor, face identifier, speaker identifier, RFID again
cultural factors affecting design - shoes placed near front door in S Korea, people have to spend a few seconds in the area to put shoes on
mobile phone control system
whiteboard application used as a lightweight reminder
evaluation results -
mirror put near face identification cam to encourage usage
fixed eight of screen a problem
RFID - if it works, provides natural & transparent interaction
everybody loved whiteboard
showing photos and videos of objects - not everyone got this
[this was a good talk, at least they built something and carried out UCD - seems to be getting hard pointed questions from academia]
****
privacy risk models
Carnegie Mellon
"what about my privacy?" - especially with location and ubicomp
privacy means different things to different people
new issues - genetics, telemarketers
need a design method to identify and manage privacy risks
analogy - security threat model - what you are trying to prevent and what your adversaries have
analysis - common questions, like a task analysis
privacy risk management
help teams prioritise and manage risks
this model provides *reasonable* levels of protection for forseeable risks
questions:
social and organisational context
who are the users?
who shares information?
who sees it?
(difference between sharing info with coworkers or spouse, what people will share (e.g. health))
start with *most likely* users
e.g. people in your mobile phone book
what kinds of personal information are shared?
current location vs. home phone no vs. hobbies
some information already known between people
different ways of protecting diff pieces of info - location can be withheld/revoked, name or birthday cannot
what is the relationship between sharers and observers?
risks w/friends are unwanted intrusions or embarrassment
risks w/paid services are spam, 2nd use, hackers
incentives for protecting personal information
mechanisms for recourse (in the real world!)
what is the value proposition for sharing personal information in the first place?
e.g. nurse locator badges - initial reaction is bad, but has let them prioritise their work
risk vs. benefit of privacy info
technology questions
how is personal info collected?
do users have practical control over their information?
network based approach vs client based approach
push or pull?
do you send the info?
can others pull information about you?
need to make people aware of requests
provide understandable level of control
one-time or continuous disclosure?
e.g. observer gets snapshot
granularity of information shared?
spacial granularity (city, neighbourhood, street, room)
temporal (at boston next month, 5-6th August)
identification (a person, a woman, me)
keep and use coarsest granularity needed
risk management
estimate:
likelihood of unwanted disclosure
damage that disclosure causes
cost of adequate protection
focus of high likelihood, high damage, low cost risks first
not about the numbers, but relative, and just carrying out the process in the first place
how does the disclosure happen?
accident? bad UI? bad conceptual model? malicious?
what kind of choice and control and awareness are there?
opt-in, opt-out
what are the default settings?
Better to prevent or detect abuses?
case study - location-enhanced IM
can request a friend's current location
automatically show your location
invisible mode, reject requests
default location is 'unknown'
what about ethics?
that's different to a design judgement
should i be able to lie?
[really good talk]
Trackbacks
TrackBack URL for this entry:
http://www.anti-mega.com/cgi-bin/mt/addmttb.cgi/1089
Listed below are links to pages that reference DIS2004: ubicomp at home and on the move:
» ubicomp at home and on the move from Konomi Blog
http://www.anti-mega.com/antimega/archives/001115.html...
more | August 3, 2004 06:01 AM
» RFID powers ubiquitous computing from RFIDbuzz.com
Chris Heathcote is blogging from the 2004 conference on Designing Interactive Systems. One of the bits that caught my interest was his notes from a presentation of Samsung's Smart Home experiments: [...]
more | August 4, 2004 11:58 AM
» Motorola uses RFID to power ubiquitous media from RFIDbuzz.com
''Using a technology called 'Liquid Media,' four monitors at each station would play and stop video or audio content when Mr. Zander walked into or out of their RFID reception zones.''
more | August 5, 2004 05:44 PM
