Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....An ongoing research area in user interface software is creating new kinds of constraint solvers (e.g. Gleicher 1993] Vander Zanden 1995a] Hudson 1996] One Ph.D. research project was to try to add a multi way constraint solver into Garnet that would cooperate with the existing formula solver [Sannella 1994a] The difficulty of this task in Garnet inspired us to create an architecture in Amulet that allows multiple constraint solvers to coexist. Currently, in addition to the one way solver described above, Amulet supports a multi output, multi way solver called a web and an animation constraint ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. Seattle, WA, Dept. of Computer Science and Engineering, University of Washington. 1994a. Ph.D. Thesis. TR 94-09-10.

....carefully performs the necessary computations. 3.2 Programming Languages New programming languages offer some help. Constraint programming languages have been successfully applied to certain aspects of interactive tools, starting with the SketchPad system [38] and continuing with many others [1, 17, 26, 37]. For the most part, these applications have been directed towards building highly sophisticated user interfaces, or in offering an alternative to callback functions for connecting user s actions with functions and values in the tool s back end, rather than facilitating the implementation of the ....

....a Red Black Tree [4] with a standard paging model to cache the marks. 7.3.4 Constraints In MUSE, run time changes are always propagated from function inputs to function outputs. However, more general constraint propagation approaches have been effectively applied to user interface design [1, 17, 26, 37, 38]. Such systems are able to propagate changes bi directionally, and are further able to resolve cyclic constraints with specialized constraint solvers [2] While constraint propagation seems to be very useful when designing complex interfaces, it did not seem to be a useful form of expressibility ....

M. Sannella. Constraint satisfaction and debugging for interactive user interfaces. Technical Report 94-09-10, University of Washington, 1994.

....and Hower, 1996 ] 6 It was a laborious job to collect all these articles from so many diverse sources; the present paper probably provides such an exhaustive collection for the first time. This work has benefited from extensive overviews given in [ Leler, 1988 ] Borning et al. 1992 ] Sannella, 1994 ] Fron, 1994 ] and [ Gleicher, 1995 ] where some information presented here has been derived from. 4 2 Constraints for Graphics and Visualization As it has been shown in previous work many graphical and multimedia automation tasks can be facilitated efficiently using constraint ....

....CS is employed to find all, one, or the best solution of a CSP such as in combinatorial placement and optimization problems. It may be seen as a subset of the more general technology of CBI and can also be used as part of it in a particular problem scope. According to [Borning et al. 1992, Sannella, 1994], this classification reflects the two major approaches: the perturbation model and the refinement model . Most of the mentioned CBI systems are based on the perturbation model that uses the value inference technique. Value inference constraint solvers assign constant values to some variables and ....

[Article contains additional citation context not shown here]

M. Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, WA, 1994.

....no solution is found, since backtracking is not possible. This problem could be circumvented by allowing a constraint to change the value of the variable it was triggered from. Prior to performing local propagation one can perform an analysis to plan the best order to propagate constraints [19]. It is sometimes efficient to plan how to solve constraints before actually doing it. For example if one drags an graphics object which is constrained and must remain constrained, the planning can be done at the beginning of the drag operation, and the execution can be done in real time. 7.1 ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, University of Washington, Seattle, Washington, 1994.

....the layout of windows and other graphical entities, in particular specifying geometric constraints on manipulable graphical objects. Early examples of such research projects include Sketchpad [10] and ThingLab [1] More recent projects include Garnet [7] TRIP II [6] and many others. See e.g. [8] for many additional references. In this project, we wanted to provide a commercial quality constraint library, written in Smalltalk, to support writing interactive graphical applications. Our previous experience with such applications lead to the following choices for the constraint library: ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, 1994.

....then the set of satisfied constraints represents a locally predicate better solution to the constraint hierarchy [49, p. 60] This blocking lemma suggest the algorithm s strategy the propagation of conflict. DeltaBlue s incremental maintenance of the constraint graph plan is straightforward [49, 128, 125]. The algorithm s complexity remains O(V C) as was simple local propagation) As mentioned before, assigning new values given the same configuration (i.e. execution) is especially fast (only O(C) since at most one method is fired per constraint) 26 Extensible local propagation solvers ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, 1994.

....and Hower, 1996 ] 6 It was a laborious job to collect all these articles from so many diverse sources; the present paper probably provides such an exhaustive collection for the first time. This work has benefited from extensive overviews given in [Leler, 1988 ] Borning et al. 1992 ] Sannella, 1994] Fron, 1994] and [ Gleicher, 1995 ] where some information presented here has been derived from. 2 Constraints for Graphics and Visualization As it has been shown in previous work many graphical and multimedia automation tasks can be facilitated efficiently using constraint processing ....

....] CS is employed to find all, one, or the best solution of a CSP such as in combinatorial placement and optimization problems. It may be seen as a subset of the more general technology of CBI and can also be used as part of it in a particular problem scope. According to [ Borning et al. 1992, Sannella, 1994 ] this classification reflects the two major approaches: the perturbation model and the refinement model . Most of the mentioned CBI systems are based on the perturbation model that uses the value inference technique. Value inference constraint solvers assign constant values to some variables ....

[Article contains additional citation context not shown here]

M. Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, WA, 1994.

....constraint. We concern ourselves with the computational complexity of the problem solved by the planning phase that will be called constraint planning problem in the following. Planning algorithms can be divided into three main categories. 1) DeltaBlue [ Freeman Benson et al. 1990 ] and SkyBlue [ Sannella, 1994 ] work by propagating the conAEicts from the perturbations to the leaves of the constraint graph. 2) The propagation of the degrees of freedom scheme (in short PDOF) selects the methods in the reverse order (i.e. executing rst the methods that were chosen last) This algorithm has been used in ....

....3 proves that cp 1 and cp 3 have the same computational complexity. Since the table is now complete, we can highlight interesting points about the constraint planning problems handled by existing algorithms. SkyBlue Problem cp 2 assumes that the constraints must be required, whereas SkyBlue [ Sannella, 1994 ] can handle one type of preferential constraints. Therefore, the SkyBlue problem is NP hard. The NP completeness result given by Theorem 1 makes the exponential worst case time complexity of SkyBlue less surprising. However, Sannella, 1994 ] has proven that SkyBlue could reach this worst case ....

[Article contains additional citation context not shown here]

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, 1994.

....one constraint. This report deals with the computational complexity of the problem solved by the planning phase that will be called constraint planning problem in the following. Constraint planning algorithms can be divided into three main categories: 1) DeltaBlue [Freeman Benson 90] and SkyBlue [Sannella 94] work by propagating the conAEicts from the perturbations to the leaves of the constraint graph. 2) The propagation of degrees of freedom scheme (in short PDOF) selects the methods in the reverse order (selecting rst the methods to be executed last) It has been used in SketchPad [Sutherland ....

....3 proves that cp 1 and cp 3 have the same computational complexity. Since the table is now complete, we can highlight interesting points about the constraint planning problems handled by existing algorithms. SkyBlue Problem cp 2 assumes that the constraints must be required, whereas SkyBlue [Sannella 94] can handle a kind of preferential constraints. Therefore, the SkyBlue problem is NP hard. The NP completeness result given by theorem 1 makes the exponential worst case time complexity of SkyBlue less surprising. However, Sannella 94] has proved that SkyBlue could reach this worst case ....

[Article contains additional citation context not shown here]

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, 1994. Also available as Technical Report 94-09-10.

....of on(front wheel,chassis) and on(rear wheel,chassis) 1. Introduction Constraints specify dependency relations between objects which must be satisfied and maintained by some constraint management system. Constraint systems are used in a wide range of applications, including user interface design [1], animation [2] geometric modeling [3] and artificial intelligence [4] Geometric constraints can fix one or more degrees of freedom for positioning, orientation and dimensioning of the constraint variables. For example, when a circle of fixed radius is constrained to be tangent to a fixed line ....

....variables, depending on its type (see the constraints used in the introduction) Constraints may be multi directional, i.e. each of the involved variables constrains the others, again depending on the constraint type. Multi directional constraints provide a natural way to impose certain relations [1]. For example the constraint distance(point, circle, 0) is multi directional and means the same as distance(circle, point, 0) but the constraint on(point, circle) is one directional and means that the point must lie on the circle, whereas on(circle, point) in our implementation means that the ....

[Article contains additional citation context not shown here]

M. Sannella. Constraint satisfaction and debugging for Interactive User Interfaces. (Ph.D. dissertation), Technical Report 94-09-10, Dept. Computer Science and Engineering, University of Washington, Washington, 1993.

....identify the overall dependency among cells with a partial order among blocks. Then we can perform GLP in a unique manner as is with conventional local 1 The SkyBlue algorithm also realizes simultaneous satisfaction by invoking cycle solvers, but provides no features for relaxing constraints [72, 73]. CHAPTER 4. THE DETAIL CONSTRAINT SOLVER 36 (a) x y strong 3 = y x (b) x y strong 3 = y x weak 1 = y x (c) x strong 0 = x strong 2 = x Figure 4.1: Constraint cells. required t = 1 1 A weak t = u u 1 B weak v = 0 v 1 E strong t v = w w 2 weak w = x required x ....

....equations and multi way constraints. Previously, the problem was mainly considered to be due to the inability of local propagation to simultaneously satisfy multiple constraints, which would appear as directed cycles of methods. Researches, including DETAIL, have been made to overcome this problem [72, 73], but they turned out to be insufficient to solve all the problems regarding linear equations. 1 By local propagation, we mean algorithms for graph formalized constraint systems, not generalized local propagation (GLP) Unlike local propagation algorithms, GLP can express various strategies for ....

[Article contains additional citation context not shown here]

Sannella, M., "Constraint Satisfaction and Debugging for Interactive User Interfaces," Tech. Rep. 94-09-10, Dept. of Computer Science and Engineering, University of Washington, Sept. 1994.

....management system 1 INTRODUCTION A constraint is a relationship to be maintained among a set of variables. For example, A B=C represents the relationship that C is the sum of A and B. Constraints are widely recognized as a useful tool for user interface construction [1] 4] 10] 14] [20], 21] 23] 26] Through constraints, relationships among user interface components and their applications can be defined declaratively, leaving the task of maintaining the relationships to a constraint solver, an algorithm that determines a plan for computing constrained values in a way that ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. Ph.D. Thesis, University of Washington Department of Computer Science and Engineering technical report 94-09-10.

....First, if a constraint hierarchy has one or more cyclic solutions, existing algorithms may not be able to construct an acyclic solution, even if one exists. Second, the fastest known algorithm for satisfying multi way, multi output constraints, SkyBlue, requires worst case exponential time [41]. Our experiments with real applications indicate that cyclic hierarchies frequently arise in practice, and that they can seriously degrade the performance of a constraint solver. This paper presents an incremental constraint solver that surmounts these shortcomings. It can 2 satisfy in worst ....

....these relationships in all directions. Like multi way constraints, multi output constraints also have important uses in interactive applications. The programmer frequently views several related computations as a single computation (e.g. unpacking a data structure into multiple variables) [38, 41, 21]. Multi output constraints allow a programmer to express such a computation naturally as the sum of its constituent parts. In contrast, single output constraints force a programmer to subdivide the computation artificially. For example, most programmers prefer using a multi output constraint to ....

[Article contains additional citation context not shown here]

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. Ph.D. Th., Department of Computer Science and Engineering, University of Washington, Seattle, WA 98195, 1994. Also available as Technical Report 94-09-10.

....are rarely available for compiled languages like C , since the required information is normally not available at run time. Furthermore, none of these tools support inspecting the dynamic behavior of objects. Debugging tools for constraints have appeared in a few research systems, such as CNV [8], but these are mostly graphs of the dependencies and have not been applied to toolkits in widespread use or to debugging interactive applications. Since most of the toolkits which provide high level models are research systems which are never distributed for real use, it is perhaps not ....

....when the list of objects changes. Sometimes, bugs in constraint code arise from calculating the dependent objects incorrectly, rather than from the computation done with those objects. Therefore, the Inspector provides a facility to display the current dependencies of the constraint. Whereas CNV [8] used a node and arc graph to display the dependencies, our experience is that this kind of display breaks down for realistic applications because there are too many constraints and the names on the nodes are too long. Instead, we use a hierarchy display with elision, as shown in Figure 4. ....

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. Ph.D. Th., Dept. of Computer Science and Engineering, University of Washington, Seattle, WA, Sept. 1994. TR 94-09-10.

....and variables are prohibited (if a cycle is found, the cycle is broken by removing a constraint) and constraint methods can only have one output variable. The SkyBlue algorithm presented in this paper is a successor to the DeltaBlue algorithm that supports cycles and multi output methods [12, 13]. Cycles of constraints can occur frequently in user interface applications, particular when geometric constraints are created among graphic objects. DeltaBlue s prohibition of cycles places an undue burden on the programmer who has to worry about inadvertently introducing cycles. Several local ....

....in the set. The following sections list some of the applications built using SkyBlue and describes user interface benchmarks comparing SkyBlue to another solver. The rest of this paper presents the details of the SkyBlue algorithm. The author s dissertation describes SkyBlue in more detail [12]. Contact the author for information about implementations of SkyBlue. 2 SKYBLUE APPLICATIONS SkyBlue has been used in a number of applications produced by different research groups: ffl The Multi Garnet package [14] uses SkyBlue to add support for multi way constraints and constraint ....

[Article contains additional citation context not shown here]

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, 1994.

No context found.

Michael Sannella. Constraint Satisfaction and Debugging for Interactive User Interfaces. PhD thesis, Department of Computer Science and Engineering, University of Washington, Seattle, 1994.

No context found.

Sannella, M. 1994a. Constraint satisfaction and debugging for interactive user interfaces.

No context found.

Sannella, Michael. Constraint Satisfaction and Debugging for Interactive User Interfaces. Ph.D. Thesis, University of Washington Department of Computer Science and Engineering technical report 94-09-10.

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