patrick%20baudisch::alex%20zotov%20ed%20cutrell,%20ken%20hinckley%20microsoft%20research,%20tabletPC

1
starburst

• patrick baudischalex zotoved cutrell, ken
  hinckleymicrosoft research, tabletPC

2
summary
3
(No Transcript)
4
(No Transcript)
5
(No Transcript)
6
intro
7
small targets
8
(No Transcript)
9
touch
10
snap-to-target
11
snap-to-target
12
snap-to-target
13
snap-to-target
voronoi
14
cell
15
plant
16
3D
17
foam
18
art
19
book
20
bubble cursor
grossman and balakrishnan, CHI 05
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24

• this works
• for uniform target distribution

great
25

• this works
• for uniform target distribution

great
26
map locations
27
charts
28
icons
29
links
30
handles
31
uniform layoutsare the exception
32
voronoi
33
starburst
34
test
bed
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
walk
through
39
10,000 foot view

1. identify clusters
2. route claim lines
3. tessellate

40
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
41
O(?)
42
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
Onion peeling O(N log(n)) Preparata Shamos 85
Voronoi O(n log(n)) plane sweep
algorithm Fortune Algorithm
RoutingWorst case n/3 O(n)
Voronoi O(n log(n)) Fortune Algorithm
43
samples
44
tight cluster
loose
uniform
Voronoi
Starburst
45
5 targets
tight cluster
loose
uniform
Voronoi
Starburst
46
10 targets
tight cluster
loose
uniform
Voronoi
Starburst
47
15 targets
tight cluster
loose
uniform
Voronoi
Starburst
48
multi cluster
2
3
4
Voronoi
Starburst
49
visuals
50
bubble
51
bubble
52
starburst
53

• this wont work until pointer over tile (how to
  find the tile in the first place?)
• we need the oppositelet user point to target?
  system shows matching tile

54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
(No Transcript)
58

• if layout not sensitive to clutter
• (or no hover)

59
? multiblending
60
design iteration
61
move boundaries
62
movepixels
63
claim lines!
64
straight claim lines
65
multi-segment claim lines
66
2 studies
67
why?

• 1. doesnt starburst make targets larger
• no, starburst redistributes space(higher
  median, same mean)
• 2. isnt balanced layout obviously better
• yes, balanced is betterbut targets also get
  thinner

68
task
69
task1
start button
target
70
task2
start button
target
71
3 densities uniform, loose, tight
3 target counts
5
2 interfaces
voronoi
10
starburst

• 5 layouts 900 trials
• 12 participants

15
72
(No Transcript)
73
(No Transcript)
74
study2
75
extend
conclude
76
corners
77
corners
78
many
79
conclusion

• do not rely on uniform target distribution

80
thanks!
? patrickbaudisch.com
81
end
patrickbaudisch.com
82
(No Transcript)
83
(No Transcript)
84
(No Transcript)
85
(No Transcript)
86
(No Transcript)
87
related
88
target handles
powerpoint
89
target handles
worden et al, CHI 97
90
target handles
sears et al
91
target handles
mcguffin et al
92
(No Transcript)
93
(No Transcript)
94
labels
plaisant faikete, CHI 99
95
labels
bell et al, UIST 01