Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating Complete User Interfaces by Demonstration. Human Factors in Computing Systems, Proceedings INTERCHI'93, Amsterdam, The Netherlands, April, 1993, pp. 293-300.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....these systems, designers use a drawing editor to draw the graphics, then use the same environment to demonstrate examples of behavior and to execute (test) the program generated by the system. PBD systems include stimulus response demonstration systems that focus on end user triggered behaviors [5,7,8,14,15,21,22], and context based systems that focus on behaviors that are triggered when objects are in a certain context [10,17,18] Though useful for building interfaces such as graphical editors and simulations, current PBD systems do not provide the timing mechanisms necessary to demonstrate timed ....

....and automating iterative behavior. Rehearsal World [8] and Peridot [14] were early PBD systems that inspired the stimulus response method of PBD. The first systems to allow direct graphical demonstration of a wide range of end user stimuli and system responses were DEMO [21] and Marquise [15]. DEMO introduced the stimulus response model and a technique for demonstrating dynamically created objects, while Marquise focused on the demonstration of graphical editors, including those with palettes and modes. DEMO II [5] and Inference Bear [7] are stimulus response systems that allow the ....

[Article contains additional citation context not shown here]

Myers, B., McDaniel,R.,Kosbie, D, "Marquise: Creating Complete User Interfaces By Demonstration", Proceedings of INTERCHI '93,Amsterdam, April,1993, pp.293-300.

....press event. The first demonstration shows Grizzly Bear that pressing down on the Folder icon makes a new red folder appear there (Figure 3) The before snapshot is on the left, with a feedback icon representing the triggering event superimposed. The feedback icons are borrowed from Marquise [9]. The after snapshot is on the right. In this case, the after snapshot shows that a red copy of the prototype folder is created. The red color appears grayish in the screen shot, and is hard to see the new folder is located on top of the old one. If we were to test the behavior now, we could ....

B. Myers, R. McDaniel, and D. Kosbie. "Marquise: Creating complete user interfaces by demonstration." In Proceedings of INTERCHI, ACM Conference on Human Factors in Computing Systems, pp. 293--300, (Amsterdam, The Netherlands, Apr. 24-29) 1993.

....press event. The first demonstration shows Grizzly Bear that pressing down on the Folder icon makes a new red folder appear there (Figure 3) The before snapshot is on the left, with a feedback icon representing the triggering event superimposed. The feedback icons are borrowed from Marquise [4]. The after snapshot is on the right. In this case, the after snapshot shows that a red copy of the prototype folder is created (the red color appears grayish in the screen shot) Figure 1. Grizzly Bear s Control Panel. Figure 2. Layout of the Mini Finder Application. Document Folder ....

B. Myers, R. McDaniel, and D. Kosbie. Marquise: Creating complete user interfaces by demonstration. In Proceedings of INTERCHI, ACM Conference on Human Factors in Computing Systems, pages 293--300, (Amsterdam, The Netherlands, Apr. 24-29) 1993.

....systems, entirely different approaches have also been taken to the general problem of user interface specification by non programmers. These include systems for visual specification of interfaces (see for example [Card88, Huds90] as well as a range of bydemonstration techniques (see for example [Myer86, Maul89, Cyph91, Huds93, Myer93]) which allow some forms of user interface appearance and behavior to be specified without programming using inference techniques. 9. Experience and Conclusions While it is too early to formally judge the performance of the system in the difficult realm of enduser programming (extensive user ....

Myers, B. A., McDaniel, R. G., and Kosbie, D. S., "Marquise: Creating Complete User Interfaces by Demonstration", Proceedings of InterCHI '93, Amsterdam, April 1993, pp. 293-300.

....can draw objects quickly and sloppily. Lapidary lets designers create application specific graphical objects by demonstration without programming. Druid [61] tries to guess the alignment of predefined widgets when the user places them in a user interface. Recently, there have been some attempts [62, 63] to develop complete interfaces based on demonstration. 4.1.3 Model Based Tools We have seen that UIMSs use methods such as special purpose languages and formal grammars to specify the properties of user interfaces. One more popular method in this category is the use of application models to ....

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie, "Marquise: Creating Complete User Interfaces by Demonstration," Proc. INTERCHI 1993, (Amsterdam, The Netherlands, Apr.24-29, 1993) ACM, New York, 1993, pp.293-302.

....convenient means of expressing search strategies in densely connected knowledge bases. The next immediate item on our agenda is to combine the guidar browser with an interactive tool for constructing graphical user interfaces, such as garnet s gilt and lapidary tools, or the forthcoming marquise [5]. However, the project now enters a more ambitious phase, namely the development of a module to support case based design as well as the incremental refinement of specifications and requirements. We view this second phase as essential for the practical development of user interfaces and the ....

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating complete user interfaces by demonstration. In InterCHI'93, 1993.

....in the sample pictures. ffl The system has ready made parameterized templates of commonly used visualizations, such as a tree diagram, an organization diagram, etc. In this case, the PBE mechanism is used to determine the values of parameters. This practical approach was pursued by Marquise[10]. ffl Employing intelligent visualization methods such as FishEye[12] relieves mapping rules from representing complex layout policies. As a result, a PBE engine does not have to worry about complex layout policies such as conditional branching. The problem here is how to let the system deduce ....

Myers, B. A., R. G. McDaniel, and D. S. Kosbie, "Marquise: Creating Complete User Interfaces by Demonstration," in ACM Human Factors in Computing Systems, 1993, pp. 293--300.

....to buttons. Eager [1] watches users perform operations and detects and automates repetition. DEMO [14,2] uses a stimulus response paradigm for demonstrating the behavior of graphical objects. Chimera [7] infers constraints between graphical objects given multiple snapshots. Finally, Marquise [11] uses domain knowledge in order to support building graphical editors. All of these systems use by demonstration techniques but they are not easily compared because they have different goals and use different techniques. Nevertheless, we make an attempt to classify them in Table 1. The first two ....

....No not applicable not applicable not applicable No Event Recording Macro DEMO [2,14] 1991 92 No Yes Yes High Explicit (Aux. Objs) Yes (DEMO II) Compressed Snapshots Response Description Chimera [7] 1991 No No No High Explicit c (Aux. Objs) No Snapshots Two Way Constraints Marquise [11] 1993 No Optional d Yes Low None Yes Event Recording LISP Code Inference Bear 1994 No No Yes Medium Implicit No Ev. Rec. and Snapshots Script Table 1. Overview of Demonstrational Systems if the system reduces the number of objects that it checks for relationships. Some systems use auxiliary ....

Myers, B., R. McDaniel, D. Kosbie, Marquise: Creating Complete User Interfaces By Demonstration, Proceedings of INTERCHI'93, Amsterdam, Netherlands, April 1993, pp. 293-300.

....of a parameter of the selected object, then one example sequence would allow scaling on all objects. Again, this would save a considerable amount of designer time and storyboard space. A critical problem with PBD techniques is the lack of a static representation that can be later edited. Marquise [Myers 1993] and Smallstar [Halbert 1984] use a textual language (a formal programming language in the latter case) to give the user feedback about the system s inferences. In addition, scripts in these languages can then be edited by the user to change the program. This solution is not acceptable ....

....it uses a rule system to infer the widgets, which have previously defined behaviors. Other demonstrational 128 CHAPTER 7. RELATED WORK user interface tools that have followed in the Peridot tradition include Lapidary [Myers 1989] Druid [Singh 1990] Demo [Wolber 1991; Fisher 1992] and Marquise [Myers 1993]. One of the major complaints about Peridot was that it confirmed its inferences by asking questions. One of the principles behind SILK s design is that the inferences of the system should be kept out of the designer s way until they have a need to deal with the actual widget behavior. SILK ....

Brad A. Myers, Richard G. McDaniel and David S. Kosbie. "Marquise: Creating Complete User Interfaces by Demonstration". Proceedings of Human Factors in Computing Systems, ACM INTERCHI '93, Amsterdam, The Netherlands, April 24--29, 1993. pp. 293-300.

No context found.

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating Complete User Interfaces by Demonstration. Human Factors in Computing Systems, Proceedings INTERCHI'93, Amsterdam, The Netherlands, April, 1993, pp. 293-300.

No context found.

Myers, B.A., McDaniel, R.G., and Kosbie, D.S. Marquise: Creating complete user interfaces by demonstration. In Proceedings of INTERCHI '93: Human Factors in Computing Systems, Amsterdam, The Netherlands, April 1993, pp. 293--300.

No context found.

Brad A. Myers, Richard G. McDaniel and David S. Kosbie. "Marquise: Creating Complete User Interfaces by Demonstration," Human Factors in Computing Systems, Proceedings INTERCHI'93. Amsterdam, The Netherlands, Apr, 1993. pp. 293-300.

....dial and gauge widgets can be constructed in this way, similar to Peridot. The primitives of Chimera s inferencing mechanism are lines and other elements of a simple drawing editor. Chimera does not reason about objects that are dynamically instantiated and deleted at run time. Finally, Marquise [Myer93] uses domain knowledge to support building graphical editors. It is a by demonstration system for creating MacDraw style graphical editors. It contains built in knowledge about editor specific behaviors 8. This similarity is not surprising as all graphical snapshot inference mechanisms must ....

....mouse dependent behavior, freeing the real mouse for meta level commands. DEMO [Wolb91] always records a press event and then asks the user whether she really meant a press event, or whether she meant one of three other event types (release, enter and leave in our terminology) Marquise [Myer93] uses the keyboard to start and stop the recording. We use an alternative technique which uses time to distinguish the triggering event from the other events. Figure 4 2 shows Inference Bear s control panel just after the Take Before Snapshot button has been clicked. The Take Before Snapshot ....

[Article contains additional citation context not shown here]

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating complete user interfaces by demonstration. In Proceedings of INTERCHI, ACM Conference on Human Factors in Computing Systems, pages 293--300, (Amsterdam, The Netherlands, April 24-29) 1993.

....specify the layout and structure of the interface as described above, while in behavior mode we could demonstrate possible enduser actions and then specify how the layout and structure should change in response. A similar technique is used to describe new interface behaviors in the Marquise system [17]. A critical problem with PBD techniques is the lack of a static representation that can be later edited. Marquise and Smallstar [8] use a textual language (a formal programming language in the latter case) to give the user feedback about the system s inferences. In addition, scripts in these ....

Myers, B.A., McDaniel, R.G., and Kosbie, D.S. Marquise: Creating complete user interfaces by demonstration. In Proceedings of INTERCHI '93: Human Factors in Computing Systems, Amsterdam, The Netherlands, April 1993, pp. 293--300.

....not inextricably bundled with a widget) and, consequently, can express a wider variety of behaviors than interface builders. Demonstrational systems have also proven popular for specifying the dynamic behavior of objects in an application. Examples include systems for creating user interfaces [41, 47, 32, 5], constructing drawings and animating them [18, 19] and creating HyperCard programs [4] These systems record the user s actions on example objects and then play them back at run time when an appropriate event is received. In some cases they attempt to generalize the sequences of actions to ....

....on a gauge by 30 degrees) with elements of a response (e.g. updating a text box with the new value of the gauge) MetaMouse [18, 19] and Eager [4] try to detect repetitive actions and Lapidary 7 generalize the sequence of actions so that they are performed on all objects in a group. Marquise [32] concentrates on inferring the overall behavior of an application from demonstrations. Inference Bear [5] infers behavior from before and after snapshots of a scene. Lapidary differs from these behavior builders in that it provides a predefined set of behaviors, such as selection from a set ....

[Article contains additional citation context not shown here]

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating Complete User Interfaces by Demonstration. Human Factors in Computing Systems, Proceedings INTERCHI'93, Amsterdam, The Netherlands, April, 1993, pp. 293-300.

....commercially. One aspect of a commercial product that comes close to PBD are macro builders such as found in Microsoft s word processor, Word [24] Currently, the most successful PBD tools in research have been applied to rather limited domains. 3 Tools to make interactive software like Marquise [30], Demo II [5] and KidSim [2] are just emerging and still have deficiencies that prevent an author from creating complete and useful applications. The focus of this research will be to invent ways for the author to express behavior for a tool to understand the meaning and context of that ....

....into modes as well as ways to tell the system how to change between modes. The benefit of providing a direct way to express modal behavior is mostly a matter of convenience. The alternative to manipulating modes is to distribute a condition to each behavior and object as required in Marquise [30] and KidSim [2] Distributing modes when there are a large number of affected behaviors is tedious for the author, both when the modes are created and especially later when one wants to make modifications. 3.1.3 Graphical Paths and Connections Paths are used in board games to determine where ....

[Article contains additional citation context not shown here]

Brad A. Myers, Richard G. McDaniel, and David S. Kosbie. Marquise: Creating Complete User Interfaces by Demonstration. Proceeding of INTERCHI'93: Human Factors in Computing Systems, 1993, pp 293-300.

....or other objects. Then the author assigns behavior to the characters by picking from a list of stock behaviors. These behaviors can be customized by changing some parameters, but in general, the author is limited to built in methods for user interaction and game play. Gamut most resembles Marquise [15]. Like Gamut, Marquise s goal was to create whole applications. Marquise had the ability to recognize palettes of objects and could quickly infer operations such as selecting and dragging. Marquise s major deficiency was an inability to correct guesses by demonstration. The only method for ....

....the beginning and end colors. Other graphical effects will likewise show as ghosts of the original appearance. Mouse input events appear as ghost cursor arrows in the scene with small letters and symbols nearby to indicate the direction of the button click or letter of the keypress as in Marquise [15]. Since ghosts are predominately used for highlighting, ghosts only appear during demonstration. At the moment the author presses Do Something or Stop That, the ghosts will appear showing the state immediately prior to the last event. Showing the ghosts at other times is not desirable since so ....

[Article contains additional citation context not shown here]

B. A. Myers, R. G. McDaniel, and D. S. Kosbie. Marquise: Creating Complete User Interfaces by Demonstration. Proceeding of INTERCHI'93: Human Factors in Computing Systems, 1993, pp 13 300.

....of what can be performed by direct manipulation, by allowing more behaviors to be specified by demonstration [17] We have created many systems which have explored various aspects of this problem. A partial list of these systems includes: Peridot [15] Lapidary [34] Tourmaline [16, 35] Marquise [20], Pursuit [14] Silk [10] Topaz [19] and Turquoise [13] Some of these and many other demonstrational systems are described in a recent book [3] Our latest system is Gamut which allows complete games to be created entirely by demonstration without scripting [12] Gamut is the PhD research of ....

Myers, B.A., McDaniel, R.G., and Kosbie, D.S. "Marquise: Creating Complete User Interfaces by Demonstration," in Proceedings INTERCHI'93: Human Factors in Computing Systems. 1993. Amsterdam, The Netherlands: pp. 293-300.

....of a parameter of the selected object, then one example sequence would allow scaling on all objects. Again, this would save a considerable amount of designer time and storyboard space. A critical problem with PBD techniques is the lack of a static representation that can be later edited. Marquise [16] and Smallstar [7] use a textual language (a formal programming language in the latter case) to give the user feedback about the system s inferences. In addition, scripts in these languages can then be edited by the user to change the program . This solution is not acceptable considering that the ....

Myers, B.A., McDaniel, R.G., and Kosbie, D.S. Marquise: Creating complete user interfaces by demonstration. In Proceedings of INTERCHI '93: Human Factors in Computing Systems, Amsterdam, The Netherlands, April 1993, pp. 293--300. 17.

No context found.

Myers, B., McDaniel, R., and Kosbie, D., Marquise: Creating Complete User Interfaces by Demonstration, in Proc. of ACM Conf. on Human Aspects in Computing Systems INTERCHI'93 (Amsterdam, 24-29 April 1993.

No context found.

B. A. Myers, R. G. McDaniel, and D. S. Kosbie. Marquise: Creating complete user interfaces by demonstration. In Proc. INTERCHI '93, Human Factors in Comp. Sys., Apr. 1993.

No context found.

B. A. Myers, R. G. McDaniel, and D. S. Kosbie. Marquise: Creating complete user interfaces by demonstration. In INTERCHI '93, Human Factors in Computing Systems, Apr. 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC