user perceptions of drawing logic diagrams with pen-centric user interfaces
DESCRIPTION
User Perceptions of Drawing Logic Diagrams with Pen-Centric User Interfaces. Bo Kang, Jared N. Bott, and Joseph J. LaViola Jr. Interactive Systems & User Experience Lab Department of EECS University of Central Florida. Outline. Related Work Motivation Experiment Results Discussion - PowerPoint PPT PresentationTRANSCRIPT
User Perceptions of Drawing Logic Diagrams with
Pen-Centric User Interfaces
Bo Kang, Jared N. Bott, and Joseph J. LaViola Jr.
Interactive Systems & User Experience LabDepartment of EECS
University of Central Florida
Outline• Related Work• Motivation• Experiment• Results• Discussion• Conclusion
Related Work• Pen-based Interfaces– DENIM (Lin et al. 2000)– CrossY (Apitz et al. 2004)– LogicPad (Kang and LaViola 2012)
• Evaluation and Perceptions– MacKenzie et al. (1991)– Wais et al. (2007)– Forsberg et al. (2008)– Vatavu et al. (2011)– Bott et al. (2011)
Motivation• LogicPad– Hybrid interface for
Boolean logic problems
– Seemed faster than sketching
• Is speed more important for these diagrams?
Experiment• Performed study comparing 3 pen-
based interfaces for creating Boolean logic diagrams
SketchPure sketch, 100% accurate, “ideal”
Drag-and-DropTraditional WIMP-based, stylus and keyboard
HybridRadial menu for gates, sketch labels and wires
Subjects and Apparatus• 18 college students participated– 3 female, 15 male– Ages 19 – 30
• Worked on tablet PC– HP EliteBook 2760p
Experimental Task• 3 copy-and-verify
tasks (one per interface)– 6 problems per task
• Given a diagram-equation pair
• Copy a diagram using interface, get a Boolean equation back
• Compare given equation with one from interface
Experimental Design• Wizard of Oz approach
– All 3 interfaces programmed with ordering of tasks, which equation to show
• 3 by 2 within-subjects factorial design– Independent variables:
user interface (sketch, drag-and-drop, hybrid) and diagram complexity (low, high)
– Dependent variable: completion time
Metrics• Measured
completion time• Rate each interface– Making gates– Making wires– Making labels– Arrange gates– Create diagrams– Speed– Frustration
• Rank interfaces– Ease of use– Speed– Naturalness– Overall preference
Hypotheses• Primary: Participants will prefer the
sketch interface over the hybrid and drag-and-drop interfaces
• Secondary– Hybrid interface will be faster than the
sketch and drag-and-drop interfaces– Sketch interface will be rated more
natural than the hybrid and drag-and-drop interfaces
Results - Rankings
Results – Completion TimeInterface Drag-and-Drop Hybrid SketchComplexi
tyσ σ σ
Low 92.9 19.5 70.8 12.1 70.0 14.8High 226.4 38.7 186.6 33.3 202.4 34.6
Overall 159.7 74.1 128.7 63.7 136.2 72.1
• T-tests on completion time– Sketch faster than drag-and-drop– Hybrid faster than drag-and-drop– Hybrid faster than sketch, except at low
complexity (no significance)
Results – Ratings• Significant tests– Ease of use in labeling, arranging, and
creating diagrams• Easy label: sketch > hybrid > drag-
and-drop• Easy arrange: drag-and-drop >
sketch• Easy diagram: hybrid > drag-and-
drop
Results – Hypotheses• Primary hypothesis – Did they prefer
sketch interface?– No
• Was sketch most natural?– Yes
• Was hybrid fastest?– Yes…
Discussion• Speed and user perceptions– Difference in rankings/ratings and
completion time–Why?• No task switching with sketch interface• Internal versus external mistakes• Drawing style slows down sketching
– No easy way to spatial arrange drawing– 100% sketch accuracy not as fast as
hybrid
Discussion – cont.• Why rank an interface as best
overall?– Spearman’s rank correlation between
overall ranking and other rankings and ratings
– Highest correlations with ease of use ranking, naturalness ranking, speed ranking
– Sketch “was fast for small diagrams”– Sketch “was easy and natural”– Hybrid “easier than the others”
Conclusion• Would users prefer a sketch interface over a
faster interface?– Study comparing three pen-based interfaces for
creating logic diagrams– Sketch was well-liked, but not decisively so
• User perceptions and measurements– Perception of speed and our measurement differed
• Should we continue research into pen-based interfaces for structured 2D languages?– Yes
• Pure sketch might not be the most powerful, but clearly desirable traits
Acknowledgments• This work is supported in part by NSF
CAREER award IIS-0845921 and NSF awards IIS-0856045 and CCF-1012056.
QUESTIONS?
Bo Kang: [email protected] N. Bott: [email protected] J. LaViola Jr.: [email protected]
RatingsInterface Drag-and-drop Sketch Hybrid
Statement σ σ σEasy gate 5.72 1.447 6.17 0.924 6.39 0.698Easy wire 5.00 1.680 5.89 1.231 5.94 1.305
Easy label 4.67 1.815 6.89 0.323 6.33 1.085Easy
arrange6.22 0.878 4.56 1.854 5.78 1.263
Easy diagram
5.39 1.243 5.67 1.138 6.28 0.752
Quick 5.61 1.335 5.83 1.249 6.11 1.023Frustrating 3.06 1.731 2.50 1.249 2.11 1.023
CorrelationsCorrelation with Overall Ranking ρ p
Ease of Use 0.778 0.000Naturalness 0.694 0.000Speed 0.639 0.000Frustration 0.411 0.002Easy Diagram -0.409 0.002Easy Label -0.397 0.003Completion Time 0.397 0.003Quick -0.342 0.011Easy Gate -0.314 0.021Easy Wire -0.281 0.039Easy Arrange 0.056 0.690