INTERACTING WITH GRAPHIC OBJECTS

Document manipulation is one of the fields in which computer-based systems are proving themselves to be most effective. Nevertheless different typesetting tasks require different degrees of control on the behavior of a typesetting system, and different users need different kinds of computer feedback in response to their commands and these are important barriers to the acceptance of computer tools. A solution to the lack of a "best" user interface might be to have a powerful virtual typesetting machine and easy-to-use facilities for defining interfaces.

High resolution displays and pointing devices suggest a style of using computer systems that is analogous to that of the familiar paper and pencil. In addition, the capability of using stored knowledge--both data and procedural information--and the use of adequate structuring methodologies can provide a great deal of power and flexibility in handling images and the objects they represent.

Taking these considerations into account, the author has designed and programmed GOB, an interactive document processor that operates in terms of graphic objects whose structure and format are stored separately. GOB handles both text and graphics in a uniform way, and it can produce documents of quality similar to that of the best batch text processors. The positioning and the redisplay mechanisms of GOB are two of its most original features. Positioning is done by solving sparse systems of linear equations that are constantly changing. GOB provides a programming language that can be used to program complex formatting functions and to define user interfaces.

A subset of GOB has been made to run to the point of generating pictures like those appearing in several places of this dissertation. The experience with the prototype suggests that it is indeed feasible to build interactive systems of this kind, that have a value by themselves (using built-in user interfaces), and that also allow for customization of the document manipulator.