Longitudinal Study on Rapi by yurtgc548

VIEWS: 0 PAGES: 15

									Longitudinal Study on Rapi
Jesse Prabawa Gozali
WING Meeting Feb 15th 2008
Presentation Outline
 Overview of Rapi
 Longitudinal Study (Dec 07 − Jan 08)
    ◦   Motivation
    ◦   Methodology
    ◦   Findings
    ◦   Discussion
   Future Works
Overview of Rapi
   Rapi (\’rəpē\)
    ◦ From the Indonesian word meaning tidy,
      organized or presentable
    ◦ A rich Open Public Access Catalog (OPAC)
   Four key features:
    ◦   Dual pane: overview + details
    ◦   Customizable results table
    ◦   Tabs
    ◦   Embedded suggestion bar
   Available at http://linc.comp.nus.edu.sg/
Overview of Rapi
Presentation Outline
 Overview of Rapi
 Longitudinal Study (Dec 07 − Jan 08)
    ◦   Motivation
    ◦   Methodology
    ◦   Findings
    ◦   Discussion
   Future Works
Motivation
   Iterative design and test cycle:
       Date                           Description              Phase
 Jun − Jul ’06      Literature review                             −
 Aug ’06            Paper prototyping                             −
 Sep 8 ’06          Expert review session                     Snapshot
 Dec 22 ‘06         Preliminary usability study (JCDL ’07)    Snapshot
 Feb 21 − 23 ’07    Usability study                           Snapshot
 Mar − Oct ’07      Server side codes and API development         −

 Dec 16 ’07 − Jan   Longitudinal study                       Longitudinal
 16 ’08
Motivation
        Snapshot          Longitudinal

          First-time
                          Change over time
         experiences

       Feature adoption   Feature adoption

                          Changes in search
         Learnability
                              behavior

                            Novel usage
        Discoverability
                             patterns
Methodology



   Two-pronged approach
    ◦ User study (nine students)
       Interaction logs
       Questionnaires (scenario-based tasks and diary entries)
       Interviews
    ◦ UI kiosk placed in situ
       Interaction logs
Findings
Findings
Findings
   Feature adoption                          Feature              First Task   Last Task

    ◦ Number of participants using                                 1   2    3   1   2   3
      various features in the three           Tabs                 8   9   9    9   9   9
      scenarios of the first and last         Overview + Details   3   0   0    7   0   0
      tasks                                   Suggestion bar       7   1   4    3   1   2
   Novel usage patterns                      Sorting columns      3   5   3    3   3   1
    ◦ Multi-type,                          Editing columns         0   0   1    0   0   0
      multi-combination searches
    ◦ Parallel searches prior to relevance assessment
    ◦ Multi-tasking
    ◦ Subtask caching
   Changes in search behavior
    ◦ Parallel searches
    ◦ Suggestion-prone
    ◦ Relevance assessment
   More details in the paper available from CVS
Discussion
   Three proposals for a rich OPAC
    ◦ Centralized search environment
         Tabs as the control center
         Search and browse
         Mixed results (e.g. type, medium)
         Single page environment
    ◦ Context-aware filters
       Advanced syntax may be un(popular|usable)
       Faceted metadata may be insufficient
       e.g. software version in how-to books, domain-specific
        keywords
    ◦ Context-aware suggestions
       Task-centric suggestions
       e.g. Macromedia Flash == Adobe Flash?, Rowling != Rowing?,
        Foo Bar?
Presentation Outline
 Overview of Rapi
 Longitudinal Study (Dec 07 − Jan 08)
    ◦   Motivation
    ◦   Methodology
    ◦   Findings
    ◦   Discussion
   Future Works
Future Works
   Supporting quantitative data
    ◦   Task completion rates
    ◦   Average time on task
    ◦   Average task error counts
    ◦   Average task satisfaction scores
Thank you
Any questions?

								
To top