Presentation is loading. Please wait.

Presentation is loading. Please wait.

Support for the dynamic process - history mechanisms Vijayshankar Raman.

Similar presentations


Presentation on theme: "Support for the dynamic process - history mechanisms Vijayshankar Raman."— Presentation transcript:

1 Support for the dynamic process - history mechanisms Vijayshankar Raman

2 General comments Search is a dynamic process –berry picking –useful information = total of all that is collected along the path. “half-finished ideas and partly explored paths can be retained for future use, saved and brought back to compare to later interactions, and the results can be combined via operations …” [Hearst book chapter]

3 Hence, –retain search history to allow easy navigation –save and reuse parts of a search session use output of one part as input to another –manipulate queries, search results, documents, search services etc. using a set of tools/operations -- workcenter/workspace – retain context across sessions

4 SketchTrieve SketchTrieve Goals –mediate communication –allow many different IR technologies –simplify search engine (thru good defaults) –support incremental learning –customization without programming (“radically tailorable”)

5 SketchTrieve (contd) (see Fig. 2, pg 3) documents, queries, result sets, filters, services (search engines) … are all elements data-flow model: connect elements via “box-and- line” notation. anything can be connected to anything else that is compatible: “informality predominates” gives user control over layout & locality (customization, notion of workspace)

6 Other neat tricks –when searcher selects service, other compatible services shown – indicate progress as busy/done/caution (but no intermediate results) SketchTrieve (contd)

7 Itemstash -- retrieved docs Search Event -- current query History -- history of queries Result Item -- view selected docs + metadata InfoGrid (see Fig. 1, pg 2)

8 InfoGrid vs. SketchTrieve Sketchtrieve is more informal –arbitrary connections possible (confusion? ) –user decides layout Difficult to reuse query over another collection in InfoGrid InfoGrid has tree-based annotatable history cluttering avoided manually or automatically via pruning policy InfoGrid gives computational progress indication

9 Comments What is the right balance between flexibility and avoiding confusion? –workcenter? Are well-defined paths and monolithic window layouts better (see SuperBook in book chapter, page 52)?

10 PadPrints Tree-based history of recently visited web- pages (see page 1) history map placed to left of browser window zoomable, can shrink sub-hierarchies] node = title + thumbnail

11 13.4% unable to find recently visited pages only 0.1% use History button, 42% use Back problems with history list –incomplete, lose out on every branch –textual (is this a problem? ) –pull down menu cumbersome -- cannot see history along with current document Previous User study

12 User study of Padprints CHI database, National Park Service website only correctly answered questions considered 20-30% fewer pages accessed lesser response time for tasks that involve revisiting pages slightly better user satisfaction ratings

13 Comments History is useful, what netscape provides is insufficient, but are graphical histories necessary? Why thumbnails How to integrate history maps across sessions (to build personal map of the web -- what the teacher wanted)

14 Visage Basic Objects Visual elements –all data objects in visage (bars in bar chart, points in scatter chart, text labels … ) Frames -- generalization of windows –grouping function for collections of elements –lightweight objects; easy creation/deletion –scripts to govern display of elements appearance, position constraints on user controlled movement

15 User can drag and drop collections of items from one frame onto another –e.g. “Bar Chart” frame -- data dropped shown as bars of certain length and location drag and drop part of “basic physics” -- applies universally example in Figure 3 scripts can add global behavior across frames (brush and link)

16 Slide Frames –other frames dropped on it are minaturized –positionable and annotatable by user Slide Sorter Frame –use can only alter linear sequence within slide Dynamic Query Sliders –add sliders to frames to select subset of objects and then manipulate them –collect only necessary attributes of objects –dropping filter onto retrieved results

17 Comments Which is better to store history of search activities -- tree-based or slide sorter ? Can images in the slide sorter convey enough information? Can slide sorter be used to store overview of “virtual workspaces” (book chapter, page 51)

18 DLITE UI to a digital library direct manipulation interface to a distributed info. system – must show network, remote server status workcenter approach –lots of handy tools for one task –experts (librarians) create workcenters –contents persistent –concurrently shareable across sites Web browser used to display document or collection metadata

19 Components/tools in DLITE Documents (search results, or local documents) collections of components (e.g. result sets) queries -- translator used to apply same query to many sources services -- search services, summarization, OCR, translation … people (for access control, payment …)

20 Interaction See Fig 1 (page 144), screenshotsscreenshots pointing at object brings up tooltip -- metadata activating object -- component specific action –5 types for result set component drag-and-drop data onto program animation used to show what happens with drag- and-drop (e.g. “waggling”)

21 Comments Users seem to have lots of problem with flexibility (result set icon activation) workcenter -- customization, acts as reminder asynchronous, concurrent practising liked by user animation used to track progess, (partial) results -- but why try to share this across machines?


Download ppt "Support for the dynamic process - history mechanisms Vijayshankar Raman."

Similar presentations


Ads by Google