/[fenfire]/journals/plans/SA_2003.rst
ViewVC logotype

Diff of /journals/plans/SA_2003.rst

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.139 by tjl, Wed May 14 12:14:02 2003 UTC revision 1.140 by hemppah, Wed May 14 12:37:11 2003 UTC
# Line 119  Novel focus+context user interfaces Line 119  Novel focus+context user interfaces
119      as the central item and peripherally the connected locations.      as the central item and peripherally the connected locations.
120      Several of the new technologies in fenfire are related to user interfaces,      Several of the new technologies in fenfire are related to user interfaces,
121      in particular to showing the context visibly yet nonintrusively.      in particular to showing the context visibly yet nonintrusively.
     Some of the ideas aiming to maintain continuity in the Fenfire user  
     interface (such as focus-context) have already been proved to be efficient by  
     earlier researchers. Several ideas presented in this research  
     plan are, however, genuinely  
     new and thus need to be subjected to rigorous experimenting and testing.  
122    
123  Permanent data and peer-to-peer (P2P) information sharing  Permanent data and peer-to-peer (P2P) information sharing
124      Collaboration with even only      Collaboration with even only
# Line 328  Because of the nature of this project, t Line 323  Because of the nature of this project, t
323  Fillets  Fillets
324  '''''''  '''''''
325    
326    TJL
327    
328  One of the basic principles of the Fenfire user interface is continuity.  Fillets[lukka02fillets] for graph rendering were invented
 Continuity helps the user to perceive the interface as cognitively  
 coherent - things are related to each other, not disconnected. This  
 reduces load in the user's working memory.  
   
 Fillets [lukka02fillets]_ for graph rendering were invented  
329  in the Fenfire project at the time when we were  in the Fenfire project at the time when we were
330  still working with Ted Nelson's ZigZag structure.  still working with Ted Nelson's ZigZag structure.
 Originally the term filleting  
 is used in mechanical engineering  
 to denote  
 or rounding corners of surfaces  
 in order to improve  
 the properties of cast objects.  
331    
332    Filleting, or rounding corners of surfaces, is
333    used in mechanical engineering to improve
334    the properties of cast objects.
335  Our use of the filleting for connections in graph-like diagrams  Our use of the filleting for connections in graph-like diagrams
336  is motivated by the Gestalt principle of good  is motivated by the Gestalt principle of good
337  continuation. Smoothly changing contours enable more efficient perceptual  continuation. Smoothly changing contours enable more efficient perceptual
338  grouping of visual elements, in this case,  grouping of visual elements, in this case,
339  grouping of the node and the connection.  grouping of the node and the connection.
340  Especially in situations where the graph layout is constrained,  
341    In situations where the graph layout is constrained,
342  e.g.~Focus+Context views or views where the coordinates of the nodes  e.g.~Focus+Context views or views where the coordinates of the nodes
343  are informative, fillets can clarify the relationships considerably without  are informative, fillets can clarify the relationships considerably without
344  altering the layout.  altering the layout.
345    
 Fillets are already in use in the Fenfire project, but our earlier  
 research [lukka02fillets]_ has only scratched the surface; the idea has  
 opened several new possible directions for basic research.  
 There are several possible extensions of the basic idea: for example, using  
 some texturing to express the stretching of the filleted connection  
 to the user, or lighting and shadows  
 to provide a sense 3D shape of the smooth surface.  
346    
347  We plan to test the usability of the ideas first in basic cognitive tasks  One of the basic principles of the Fenfire user interface is continuity.
348    Continuity helps the user to perceive the interface as cognitively
349    coherent - things are related to each other, not disconnected. This
350    reduces load in the user's working memory.
351    
352    Some of the ideas aiming to maintain continuity in the Fenfire user
353    interface (such as focus-context) have been proved to be efficient by
354    earlier researchers. Several ideas presented here are, however, genuinely
355    new and thus need to be subjected to rigor experimenting and testing. We
356    plan to test the usability of the ideas first in basic cognitive tasks
357  before implementing and testing them in a larger context.  before implementing and testing them in a larger context.
358    
359  The first-phase experiments include tasks varying from very simple  The first-phase experiments include tasks varying from very simple
360  perceptual tasks (such as visual search and recognition) to more complex  perceptual tasks (such as visual search and recognition) to more complex
361  problem-solving tasks. The experiments will be conducted with about ten  problem-solving tasks. The experiments will be conducted with about ten
# Line 380  interconnections, compared to more commo Line 372  interconnections, compared to more commo
372  phase, fillets will be tested in a more realistic situation requiring the  phase, fillets will be tested in a more realistic situation requiring the
373  user to trace visual connections between information items.  user to trace visual connections between information items.
374    
375    Textured backgrounds implement continuity by providing unique, permanent
376    visual context for interface elements to facilitate search and recognition
377    of the elements in prolonged use. Textures will be tested to show that the
378    user can implicitly memorize textures during interaction with the
379    interface elements and use this information to recognize the elements
380    faster and more accurately. We will also conduct readibility tests on
381    textures to show that a textured background of text does not necessarily
382    impair either the legibility or the readability of the text.
383    
384  Finally, we will carry usability tests on a simulated interface, in which  Finally, we will carry usability tests on a simulated interface, in which
385  several ideas are implemented simultaneously. The users will be observed  several ideas are implemented simultaneously. The users will be observed
386  in realistic use situations. Behavioral data will be gathered in verbal  in realistic use situations. Behavioral data will be gathered in verbal
387  protocols, videotaping, and surveys.  protocols, videotaping, and surveys.
388    
389    
390    Our earlier reseach has shown that filleting significantly
391    improves the perceivability of certain graph layouts.
392    We plan to obtain experimental justification and
393    to formulate principles of filleting
394    in more general layouts including curved connections
395    and smooth animation.
396    Generalizing filleting to three dimensions ...
397    
398        - quantitative experimental results from general graphs layout
399    
400    Goals???
401    
402    
403  Buoys  Buoys
404  '''''  '''''
405    
# Line 445  Break lines Line 460  Break lines
460    
461  .. make sure about the concrete stuff  .. make sure about the concrete stuff
462    
463  Break lines are another one of Fenfire's user interface techniques  *Break lines* are a technique used in technical drawing
 that have been influenced by technical drawing.  
 In technical drawing, break lines are used  
464  for indicating  for indicating
465  where an object extends beyond what is drawn in the current  where an object extends beyond what is drawn in the current
466  diagram. It is visually clear since it uses a shape that  diagram. It is visually clear since it uses a shape that
467  is obviously not a part of the object's own shape  is obviously not a part of the object's own shape
468  (wiggly freehand line, see Fig. [ref-fignasa]_).  (wiggly freehand line, see Fig. [ref-fignasa]_).
469  We apply this technique by drawing the buoys as  We apply this technique by drawing the buoys as
470  non-photorealistic pieces torn off the target document.  non-photorealistical pieces torn off the target document.
471    
472  An important part of the design of break lines in Fenfire  To allow for fluid animation,
473  is the fluid animation:  the shapes of the break lines need to be carefully designed.
474  for example, animating a fragment to a full document  For example, animating a fragment to a full document
475  should not look like the edge just gliding  should not look like the edge just gliding
476  over the document, but rather as if larger and  over the document, but rather as if larger and
477  larger parts were magically torn off the original document.  larger parts were magically torn off the original document.
 In the software engineering part of the Fenfire project, we have  
 just accepted that this type of animation  
 seems to work and moved on; however, this has opened a number of  
 interesting questions for basic research.  
   
478    
479  .. figure:: ../../manuscripts/xupdf/mercury5part2  .. figure:: ../../manuscripts/xupdf/mercury5part2
480     :width: 8.45cm     :width: 8.45cm
# Line 481  interesting questions for basic research Line 489  interesting questions for basic research
489  The concrete research goals are to evaluate the practical value of  The concrete research goals are to evaluate the practical value of
490  break lines as implemented in the Fenfire project and  break lines as implemented in the Fenfire project and
491  to study more formally the perceptual properties  to study more formally the perceptual properties
492  of different types of break lines.  of different types of torn edges.
493  Especially the properties of the animation are interesting,  Especially the properties of the animation are interesting,
494  because the motion appears quite natural even though  because it looks somewhat natural even though
495  there is no real-world analogue.  there is no real-world analogue.
496    
497    
498  Text filtering  Text filtering
# Line 510  of per-pixel operations has been increas Line 518  of per-pixel operations has been increas
518  possible to run small *fragment programs* for each rendered pixel.  possible to run small *fragment programs* for each rendered pixel.
519  This flexibility allows us to overcome the restrictions in the  This flexibility allows us to overcome the restrictions in the
520  texture filtering.  texture filtering.
521    
522  However, designing such filters in an *ad hoc* fashion is  However, designing such filters in an *ad hoc* fashion is
523  extremely cumbersome and error-prone; a more  extremely cumbersome and error-prone. What is needed is a suitable
524  formal approach for evaluating the results of  mathematical framework for modeling the
525  different filtering methods is needed.  *perceptual* qualities of rendered text.
526    
527  The goals in this area are the development of a mathematical framework  The goals in this area are the development of the mathematical framework
528  for modeling the *perceptual* qualities of rendered text  for qualitatively
529  as well as experimental evaluation of the models.  XXX
530    As one theoretical approach for modeling the readability
531  As one theoretical approach for the readability modeling  we can use the correlation function of the rendered text.
 we plan to use the correlation function of the rendered text.  
 There have not been much research on this approach in  
 graphics context.  
 However, we hope to be able to apply our earlier experience  
 on Bayesian methods and similar models.  
532    
533    
534  Unique background textures  Unique background textures
# Line 595  is constructed deterministically, which Line 599  is constructed deterministically, which
599  overlay and the placement of services is controlled tightly. This feafure provides $O(\log{n})$  overlay and the placement of services is controlled tightly. This feafure provides $O(\log{n})$
600  resource discovery efficiency and the number of packets grow with $O(\log{n})$. However,  resource discovery efficiency and the number of packets grow with $O(\log{n})$. However,
601  the tightly structured approach may not be reliable in highly adverse conditions and the  the tightly structured approach may not be reliable in highly adverse conditions and the
602  performance may suffers (e.g., sudden network partition or hostile attack).  performance may suffer (e.g., sudden network partition or hostile attack).
603  Thus, either of the approches do not support efficient resource discovery in  Thus, either of the approches do not support efficient resource discovery in
604  peer-to-peer environment in *adverse* conditions.  peer-to-peer environment in *adverse* conditions.
605    
606  In our research, we study mathematical models which can be used for analyzing the impacts of adverse  In our research, we study mathematical models which can be used for analyzing the impacts of adverse
607  events both in the loosely and tightly structured systems. Currently such impacts are not well known and mathematical  events in Peer-to-Peer environment. Currently such impacts are not well known and mathematical
608  framework could help researchers to understand these events and effects better. Also, we  framework could help researchers to understand these events and effects better. Also, we
609  study algorithms which provide $O(1)$ efficiency in Peer-to-Peer environment. In this case,  study algorithms which provide constant or locarithmic efficiency in Peer-to-Peer environment
610  our focus is to investigate how well these algorithms outperform in highly adverse conditions.  (e.g., [maymounkov02kademlia, gupta03kelips]_. In this case, our focus is to investigate
611    how well these algorithms outperform in highly adverse conditions.
612    
613        
614  Research schedule  Research schedule

Legend:
Removed from v.1.139  
changed lines
  Added in v.1.140

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26