|
|
© 1997 Copyright on this material is held by the authors.
The particular extension which is the focus of this contribution is the provision of an interactive graphics facility which allows semantic knowledge to be attached to graphical objects. The usefulness of this concept has long been recognised, e.g. Murray et al, [7]. Edmonds & Candy [2] discuss the effectiveness of providing a domain expert with computer tools which support highly interactive use of visual data and also a means of specifying the knowledge both graphically and in text form.
The requirement for Semantic Graphical Objects (SGO) to provide users with a method of graphically representing domain knowledge, was identified during the FOCUS project [3,6] which also identified a requirement for an Editor (SGOE) for such objects. A set of seven essential properties were identified for a SGOE [7] that would enable users to design their own graphical objects for manipulation and drawing and to use these semantically meaningful objects to annotate a bit-mapped image. It is this set of requirements which are being addressed in the work described here. The task was to provide, as a Java applet, a simple object generator and test-bed for trialing graphical interactive objects on the Web. As in FOCUS, the objects were to be constructed from a combination of graphics primitives (lines, circles, ellipses, boxes) with semantic information (constraints & textual descriptions). The data defining such an object was to be capable of being stored in and retrieved from a knowledge base accessible from a Web server. The results of the user's interaction with the object was to be returned to the server.
An object which has been defined can be manipulated interactively on a Web page. To interact with the object the user performs a mouse-down action on a focal point of the object and drags the mouse across the canvas. How that primary focal point is allowed to move with respect to other points within the object is dependent upon the defined constraints and the current state of those points. If the primary point itself is not anchored but is linked to another point which is anchored then any constraint between these points must be evaluated as the mouse is moved. Otherwise, a movement applied to the primary focal point must be propagated through related points until points are reached which either have no further connections or are anchored. The relative co-ordinate position of the non-anchored points is maintained during the move thus automatically satisfying any geometrical constraints. If an anchored point is reached then any applicable constraint must be executed to ensure that the actual move that transpires does not transgress its requirement. Currently, the system by default assumes that the user intends the primary focal point to be unanchored.
Taking the manikin as an example, if none of its points are anchored and the user's primary point of focus is the conjunction of the torso and thigh, then the whole manikin moves to reflect the mouse movement accurately. On the other hand, if the upper body point is anchored and the torso is of fixed length, then the primary point of focus is constrained to a circular movement with the anchored neck point as its centre and body length as radius.
The user might apply constraints with no visible graphical manifestation. Given an anchored point representing a seat headrest position, then the top of the manikin's torso might be constrained to be forward of that position. Another constraint might be imposed to keep the same torso point above the manikin's knee joint. If the user moves his or her primary point of focus to top of the torso then, with the base of the torso and knee joints both anchored, the movement of the torso is constrained to an arc with an outline which lies to the right of the headrest and above the knee.
Initial feedback suggests that, although a little sluggish, the dynamic interactive response time of the Web interface is sufficiently fast to be within the bounds of user acceptability. Thus proof of concept has been achieved.
2. Edmonds, E.A.& Candy, L. Computer Support for Concept Engineering Design: Enabling Interaction with Design Knowledge, Journal of Systems Engineering and Electronics, 7, 2, 1996, 55-71.
3. Edmonds, E.A., Murray, B.S., Ghazikhanian, J. & Heggie, S.P., The re-use and integration of existing software: a central role for intelligent user interface, Monk, Diaper & Harrison (Eds), People and Computers VII, Cambridge Univ. Press, Cambridge, 1992, pp 415-427.
4. Java: Programming for the Internet. Available as http://sunsite.doc.ic.ac.uk/packages/java-http/
5. Medyckyj-Scott, D., Cuthbertson, M., Newman, I., Discovering environmental data: metadatabases, network information resource tools and the GENIE system, Int. J. Geographical Information Systems, 10, 1, 1996, 65-84.
6. Murray, B.S., Candy, L., Edmonds, E.A., User centred complex systems design: combining strategy, methods and front-end technology, Benyon & Palanque (Eds), Critical Issues in User Interface Systems Engineering, Springer-Verlag, London, 1996 Ch.10 pp 169-187.
7. Murray, B.S., Edmonds, E.A., Candy, L., Foster, T.J., Constructing Semantic Graphical Objects. Santo (Ed.), Proceedings of Third Int. Conf. on Computational Graphics & Visualisation Techniques: COMPUGRAPHICS '93, Univ. Lisbon, Portugal, 1993, 46-57.
|
|