News Papers Covered More Reading project meetings due this Fri 10/23 A review of overview+detail, zooming, and focus+context interfaces. Andy Cockburn, A Review and Taxonomy of Distortion-Oriented Presentation Techniques. Y.K. Leung Amy Karlson, and Benjamin B. Bederson. ACM Computing Surveys 41(1), 2008. and M.D. Apperley, ACM Transactions on Computer-Human Interaction, Vol. 1, No. written proposals due next Fri 10/30 (continued) 2, June 1994, pp. 126-160. http://www.ai.mit.edu/people/jimmylin/papers/Leung94.pdf Lecture 10: Focus+Context SpaceTree: Supporting Exploration in Large Node Link Tree, Design Evolution and Empirical Evaluation. Catherine Plaisant, Jesse Grosjean, and Ben B. Bederson. Proc. H3: Laying Out Large Directed Graphs in 3D Hyperbolic Space. Tamara Munzner, Information Visualization InfoVis 2002. ftp://ftp.cs.umd.edu/pub/hcil/Reports-Abstracts-Bibliography/2002- Proc InfoVis 97. 05html/2002-05.pdf CPSC 533C, Fall 2009 The Hyperbolic Browser: A Focus + Context Technique for Visualizing Large Hierarchies. John Lamping and Ramana Rao, Proc SIGCHI ’95. Tamara Munzner http://citeseer.nj.nec.com/lamping95focuscontext.html A Fisheye Follow-up: Further Reflection on Focus + Context. George W. Furnas. SIGCHI 2006. UBC Computer Science Untangling the Usability of Fisheye Menus. Kaspar Hornbaek and Morton Hertzum, Mon, 19 October 2009 ACM Transactions on Human-Computer Interaction 14(2), 2007. TreeJuxtaposer: Scalable Tree Comparison using Focus+Context with Guaranteed Visibility. Munzner, Guimbretiere, Tasiran, Zhang, and Zhou. SIGGRAPH 2003. http://www.cs.ubc.ca/˜tmm/papers/tj 1 / 24 2 / 24 3 / 24 4 / 24 Focus+Context: Cockburn SpaceTree Focus+Context Distortion Intuition Graphical Fisheye Views DOI: API(x) - D(x,y) focus+context tree: filtering, not geometric distortion move part of surface closer to eye API: a priori interest animated transitions Perspective Wall example D: distance, semantic or spatial x: data element y: current focus DOI for selective presentation vs. distortion infer DOI through interaction vs. explicit selection single vs. multiple foci [A review of overview+detail, zooming, and focus+context interfaces. Cockburn, [A review of overview+detail, zooming, and focus+context interfaces. Cockburn, Karlson, and Bederson. ACM Computing Surveys 41(1), 2008. From Graphical semantic zooming Karlson, and Bederson. ACM Computing Surveys 41(1), 2008. From Perspective Fisheye Views, Sarkar and Brown 1992] Wall, Mackinlay Robertson and Card 1991] demo [A Review and Taxonomy of Distortion-Oriented Presentation Techniques. Leung and Apperley, ACM ToCHI 1(2):126-160, Jun 1994.] 5 / 24 6 / 24 7 / 24 8 / 24 Document Lens, Table Lens 2D Hyperbolic Trees Avoiding Disorientation 3D Hyperbolic Trees/Graphs fisheye effect from hyperbolic geometry problem scalability argument: information density at periphery video: open-video.org/details.php?videoid=4567 maintain user orientation when showing detail hard for big datasets exponential in depth node count, space needed global overview local detail A−B aardvark jerboa first baboon C−D capybara kangaroo the dodo E−F elephant K−L second ferret gibbon lion almost G−H hamster iguana I−J jerboa mongoose third K−L kangaroo M−N quick lion nutria M−N mongoose fourth nutria tiptop O−P orangutang possum orangutang quail O−P Q−R rabbit possum [A review of overview+detail, zooming, and focus+context interfaces. Cockburn, [H3: Laying Out Large Directed Graphs in 3D Hyperbolic Space. Tamara Munzner, fifth scorpion S−T tapir Proc InfoVis 97.] brown unicorn quail Karlson, and Bederson. ACM Computing Surveys 41(1), 2008. From: Document U−V viper sixth W−X whale Q−R Lens, Robertson and Mackinlay 1993. done x−beast rabbit Y−Z yellowtail Table Lens, Rao and Card 1994.] seventh zebra Anteater scorpion a−b Badger [The Hyperbolic Browser: A Focus + Context Technique for Visualizing Large fox Caiman S−T c−d Dog tapir Hierarchies. John Lamping and Ramana Rao, Proc SIGCHI ’95.] eighth e−f Earthworm Flamingo 9 / 24 10 / 24 11 / 24 12 / 24 Noneuclidean Geometry Parallel vs. Equidistant Exponential Amount Of Room 2D Hyperbolic Models Euclid’s 5th Postulate euclidean: inseparable room for exponential number of tree nodes Klein/projective Poincare/conformal Upper Half Space hyperbolic: different exactly 1 parallel line 2D hyperbolic plane hemisphere area spherical embedded in 3D space geodesic = great circle hyperbolic: exponential no parallels 2 π sinh 2 r euclidean: polynomial hyperbolic (torus.math.uiuc.edu/jms/java/dragsphere) [Three Dimensional Geometry and Topology, William Thurston, Princeton University Press] 2 π r 2 infinite parallels Minkowksi [Thurston and Weeks 84] 13 / 24 14 / 24 15 / 24 16 / 24
Distortion Challenges Untangling Usability of Fisheye Menus Menus: Fisheye, Overview, Multifocus Menus: Hierarchical how to visually communicate distortion compare fisheye, overview, multifocus, hierarchical gridlines, shading measurements target acquisition problem performance time, errors lens displacing items away from screen loction preferences eyetracking unsuitable if must make relative spatial judgements design issues mixed results comparing to O+D, pan/zoom distortion vs. O+D vs. hierarchical temporal landmarks fine-grained navigation: focus-lock when needed [Untangling the Usability of Fisheye Menus. Kaspar Hornbaek and Morton Hertzum, ACM Transactions on Human-Computer Interaction 14(2), 2007. Fig 2.] [Untangling the Usability of Fisheye Menus. Kaspar Hornbaek and Morton Hertzum, [Untangling the Usability of Fisheye Menus. Kaspar Hornbaek and Morton Hertzum, ACM Transactions on Human-Computer Interaction 14(2), 2007. Fig 2.] ACM Transactions on Human-Computer Interaction 14(2), 2007. Fig 5.] 17 / 24 18 / 24 19 / 24 20 / 24 Results F+C Without Distortion Fisheye Followup Generalized Fisheye Requirements troubles with focus-lock mode degree of interest (DOI): a priori importance (API), specialized hardware static structure, allowing distance defn demo: www.cs.umd.edu/hcil/fisheyemenu distance (D) LOD/API at points within structure hierarchical (baseline) outperformed for known-item task distance can be semantic or spatial interaction focused at point/region distortion vs. selection faster, more accurate agnostic to geometry smaller screen footprint DOI for selective presentation vs. distortion no differences for browsing tasks what to shown vs. how it is shown eyetrack: transition and context regions not used much how shown for fisheye geometric distortion: TrueSize as implicit API readability important - multifocus ZUIs: temporal/memory harder than side by side give up on showing entire context? multiple views: topological discontinuity at edges less space for transition regions? [A review of overview+detail, zooming, and focus+context interfaces. Cockburn, multires displays: big and heavy... Karlson, and Bederson. ACM Computing Surveys 41(1), 2008. From: Baudisch 1992.] [A Fisheye Follow-up: Further Reflection on Focus + Context. George W. Furnas. SIGCHI 2006.] 21 / 24 22 / 24 23 / 24 24 / 24 Future Goal: 10M node Tree of Life TreeJuxtaposer Phylogenetic/Evolutionary Tree Common Dataset Size Today • side by side comparison of evolutionary trees Animals Plants You are here Protists Fungi 1 2 3 4 M Meegaskumbura et al., Science 298:379 (2002) M Meegaskumbura et al., Science 298:379 (2002) David Hillis, Science 300:1687 (2003) Paper Comparison: Multiple Trees Accordion Drawing Guaranteed Visibility Guaranteed Visibility Challenges • rubber-sheet navigation focus – stretch out part of surface, • marks are always visible • hard with larger datasets the rest squishes • easy with small datasets – borders nailed down • reasons a mark could be invisible – Focus+Context technique • integrated overview, details – old idea • [Sarkar et al 93], [Robertson et al 91] • guaranteed visibility – marks always visible context – important for scalability – new idea • [Munzner et al 03] 5 6 7 7 8
Recommend
More recommend
