« Sketching in Hardware Day 3: David Merrill | Sketching in Hardware Day 3: Ed Baafi and Amon Daran Millner »

July 27, 2008

Sketching in Hardware Day 3: David Merrill

The manipulative is the message

embodied media
the narrowing of experience through windows-mouse interface
how can computers leverage our physical-world skills in a way that is generalizable


  • using existing physical gestures: tilting, swirling, pouring, moving closer

  • sense physical motion

  • display color graphics

  • sense neighbors in close proximity

  • communicate wirelessly with a computer: standalone or interface to a desktop

software: programming a distributed system

  • distributed code/algorithms: robust, enables emergent behavior - but difficult to author complex coordinated bheavior

  • centralized control: easier to author, but can be brittle

  • so siftables combines both

software: events

  • both single events (in a single siftable) and topology changes (a cluster of units, like pouring color from one to another)

interaction paradigm: embodied media rather than "physical cursor"

Sensor Network User Interfaces

  • irst idea: interface to a sensor network

  • but can we create an interface from a sensor network - so that interaction with nodes creates interface?

  • sensor networks are just barely becoming usable, but lots of opportunity for them

  • self-contained/powered mobile nodes! coordinated behavior!

  • lots of architecture problems are being solved - graceful degredation, localization, code change

the problem: how do we lower the barrier to entry?

  • what is the Flash for sensor network infrastructure? what is the Arduino for sensor network infrastructure?

  • Sun SPOTS, SquidBee, XBee


  • dead-simple mesh networking

  • foundation: discovery, position, action,

  • higher-level: global, distributed behaviors, easily see what is talking to what -- discoverability


  • Dave M: Why do you want the multiples?

  • David M: We're good with multiple manipulation, like blocks. We're good at putting things together

  • Jan: Getting students to make the mental leap from the single spot of attention to the multiple units. Also, maybe programming sensor nets is hard is because parallel programming is really hard. Also, why not touch input?

  • David M: no touch input mostly because the original vision was of piles of things, avoiding cellphone

  • Phil: Productivize?

  • David M: Yes! After finishing thesis.

  • Pamela: Is there a way we can all be aware of each other and start a forum for exchange?

  • David M: networks of distributed objects and UI

Posted by egoodman at July 27, 2008 10:43 AM


Post a comment

Remember Me?

(you may use HTML tags for style)

Trackback Pings

TrackBack URL for this entry: