CS295J/Class 10

• muse on process
  • we have a "gulf of execution"
    • goal is "good collaborative proposal"
    • we have N people; what should we do to achieve this goal?
  • why do we have this gulf?
  • harder than I anticipated
  • not enough room to {de,}select best ideas (muse on doing this before submission rather than after)
  • experience level of collaborators (muse on dhl collaborative efforts from the past)
  • no single voice/vision (past collabs again)
• let's figure out some things to execute
  • this should become the assignment
  • list what's there and what's missing/broken (dhl hypotheses here)
    • ok: intro (see below -- mix of 2)
    • ??: contributions (some more impact on science areas -- applications)
    • not ok: bg/signicance needs to be tighter and focused only on contributions
    • not ok: preliminary work need to be better related to proposed work
• intro evals
  • metrics
    • C: bring cognitive and perceptual modeling to bear in principled way on HCI design
    • S: scoped to 5 people over 5 years
    • I: high impact
    • N: innovation
    • A: analytical domain (supports hard thinking)
  • concrete global comments
    • more explicit what/how
    • "We propose" at start and "five years ... $2.5M" by Gideon and Jon excellent -- emulate!
    • terser
    • a concrete thing for the $2.5M would be good; what will they get as opposed to what will we do? a theoretical foundation? better UI's? a system for automatically evaluating UI's? new knowledge about cognition?
    • steven + (adam, trevor, gideon | gideon and jon) would capture everything
  • Meta-work and Task-Interruption Support Interfaces (andrew)

    (100) C=20, S=20, I=40%, N=20% (too focused on CS domain, scope too small, needs to appear to move outside HCI more, innovation unclear, good balance between rhetoric and specifics)

  • Collaborative 2 (need better title) (adam, trevor, gideon)

    (220) C=60%, S=40, I=60%, N=60% ("computational models" too vague, good goal for connection to psych, probably need more specifics and tighter language to make them fit, hard to scope)

  • alternative collaborative (still need better title) (adam, trevor, gideon)

    (270) C=70%, S=60, I=70%, N=70% (nice 1st para, 2nd para a little mushy, 3rd para needs work)

  • A Distributed Cognition Model of Human-Computer Interaction (gideon and jon)

    (270) C=70%, S=70, I=70%, N=60% (a bit abstract, nice connection between areas, explicit scope good, question strange, focus on distributed cognition intriguing although it may be risky, "gaps" bit sounds too much like fishing)

  • EJ (need a better title) (ej)

    (200) C=60%, S=40, I=50%, N=50% (too rambling at start

    more solution less problem, hidden "we propose", scope hard to eval)

  • A Behavioral/Cognitive Model of Human-Computer Interaction (steven)

    (190) iC=40%, S=50, I=40%, N=60% (good concrete goals and activities, "doomed to antiquation" great, not enough application)

• preliminary results
• pick out proposal-relevant aspects of new readings
  • neuropsychological assessement. interesting organized around computational themes; very structured around pathology, may need to consider high-functioning differences; not a single score; can we correlate preferences for interfaces or performance with interfaces to strength on cognitive testing measures?
  • interaction costs. norman "gulf of execution" "gulf of evaluation"; add "gulf of formation" (which seems like gulf of execution, at some level, but maybe is just the hard part of deciding what to do next); (is there a "gulf of distraction/interruption" -- or is distraction a feature?