January 25, 2007Notes: David Knight

Lecture: Sushmita Subramanian

Beyond prototypes

Social and economic challenges

Results from experiments concerning how we can solve these

Advances in technology make UbiComp more possible

  • Cheaper hardware
  • Smaller hardware
  • GPS, active badge hardware
  • Social (privacy) concerns: are people getting more comfortable being videotaped.
  • Discussion
  • London’s CCTV public camera “reality TV”
  • Not doing anything illegal so what’s the problem?
  • Implications on safety?
  • If this gets more and more cheap how can people get access to it (courts, etc)
  • Is there a parallel to when cameras first came out there were serious concerns?
  • Combination cell phone / camera devices
  • Stanford recording lectures and being broadcaster inappropriately focusing on women. Similar unprofessional uses of cameras by security guards.
  • Less scary because there is no easy way to search. (i.e. “where was John today?”)
  • Building security kept local vs. broadcast to public
  • World Wide Web
  • People have become accustomed to web portals. Helped reduce attachment to a single device.
  • Multiple device adoption. We aren’t quite there yet.
  • Mobile Communications
  • SMS approaching 1 billion per day
  • Multiple use devices
  • SIM cards – separate phone as a device from “my phone”

Broader System Challenges

  • How can these devices communication together?
  • Conflicting systems should automatically resolve this. A next generation of plug-and-play is needed.
  • Very difficult to predict user intent. Predicting the intent is key to building a system of systems.
  • Discussion
  • System trust. Systems have to know when to trust each other. Can we do this without user intervention?
  • Example – printer fetching drivers automatically
  • Do users need to know which devices are talking to each other. There might be reasons to show what systems are communicating.
  • Debugging ubicomp interfaces. If some of the devices don’t have a user interface how do developers debug them? Is having a single over-arching system an answer for this?
  • Social and legal challenges: Personal data collection and discarding. Will systems start to suggest when we want?
  • Economic concerns: If there are many services will users pay for them? Different business models.
  • Discussion
  • Parking lot example. Could a system like this pay for itself?
  • Restaurant location used to be very important in order to be successful. If we have systems that tell us these things does location become less important? Can a business model be built are referring customers to businesses to pay for the technology.
  • There aren’t incentives for many businesses to share information about inventory or waiting lines. ATMs evolved like this because once one place came out with one every bank had to.
  • Discussion of social impact of UbiComp
  • Social behavior or recommender systems. How do you decide where to go with friends.
  • Some research has been done with systems teaching users how to take care of their homes.
  • System abuse and cultural differences (values and norms). Misbehavior of people; hackers. Is Wikipedia is a test of this?
  • Having a goal of a perfect society where everyone is under surveillance isn’t realistic.
  • People using pictures and the web to change people’s behavior. (i.e. youparklikeajerk.com, etc)
  • Can an economic model solve problem of society.
  • Speedometer on side of the road does temporarily change driver behavior
  • Social boundaries already exist and how these are shaped by technology
  • Government intervention – GPS in cell phones

What do we need to ubicomp to be successful?

  • Better forum to discuss projects and results
  • Actual deployment, not just a demo.
  • Design open systems so developers can build on each other
  • Over-arching system for debugging
  • Successful business model
  • User interfaces for people to interact with multiple systems

Disappearing Hardware

  • Focus is on task completion
  • Users are accustomed to articulating the task rather than their goal. Makes it difficult for systems to know user intent.
  • Trends
  • Hardware devices have become software. Spellchecker, calculator, and translators.
  • PC has become a generalized device. Platform.
  • Huge progress has been made since 1990s. Memory, storage, device size.
  • Discussion
  • (Now) Q2 2006. 2.2B cell phone users
  • Does our vulnerability increase with technologies like cell phone over land line phone networks?
  • Much of the vulnerability is due to how these systems were designed.
  • Government could mandate infrastructure. Then there are security concerns and the business model has a more difficult time evolving.
  • Phone booths and cell phones in cars vs individual cell phones

Personal Systems

  • Intrusive
  • Limited by computational ability (technology)
  • Personal servers

Infrastructure Systems

  • Shrinking sensors to dust particle size.
  • Discussion
  • Difficult to shrink the battery
  • What is infrastructure? Wallet example. Infrastructure allows the money to be transformed into food, etc. On a desert island, the wallet is useless. Cell phone would be useless in the middle of the ocean.

What we need for UbiComp

  • Wireless needs to be faster and robust
  • Better displays
  • Can we do better than original technology (i.e. pen and paper)? Must get before it will be adapted.

Interaction techniques

  • Using stylus on very small devices.
  • Discussion
  • A device that is appropriate for older people or when wearing gloves isn’t appropriate for young people in normal environment.
  • Interface type depends on the task. Sudoku vs internet navigation on a phone.
  • Nintendo DS – stylus or touch interface
  • Wacom tablets
  • Shrinking technology and different interaction techniques. What do we really want? vs What are we now capable of building?
  • Touch-based devices. Must be task-specific to work. Dynamic touch keyboards – no tactile feedback.
  • Discussion on energy
  • Batteries grow at 10% rate as opposed to Moore’s law exponential rate
  • Small capacitor with power of a AA and can recharge in 30 seconds.
  • What can we learn from biology? Can energy be harvested from microorganism?
  • Use of kinetic energy. Recharge while walking.
  • Environmental vibrations being used to power or recharge devices. Power through light or sounds waves. Essentially this is wireless power transmission.
  • Multiple power sources. Radio example has hand-crank, solar, AC, DC.

Future

  • Quick access to information
  • Integration of computers and real world
  • Proactive systems predicting user needs
  • Healthcare, insulin pump

Optional Reading

  • Scenarios of the future.
  • ORA informs traveler about bandwidth availability.
  • Handheld device allows man to finish presentation. Automatic transfer of presentation to desktop PC. ORA detects private information in presentation and hides it from audience.

Challenges

  • Power. Fuel cells look promising.
  • Size and Weight. Reduces over time.
  • User Interface – How can we design this for invisible devices
  • Interaction Design

Discussion

  • Is wireless becoming a “utility” much like water and electricity.
  • Free wireless for urban environments.
  • Wireless in parks.
  • Municipal regulations over wireless because it cuts into