The most important result is the recognition and acceptance of the need for a clear distinction between a graphics subsystem {called also the core), and a modelling subsystem (sometimes called also the geometric subsystem).
The graphics subsystem should functionally provide two mechanisms:
The workshop urges a specification of the graphics subsystem as soon as possible.
The workshop feels that a preliminary draft of such a specification could be initiated, starting from the present work, by a working party of a smaller number of people. It is understood that such a specification shall not be the final activity; the normal and continuous but deliberate process of evolving standards applies also to the graphics subsystem. Subsequent revisions might relax the conditions given in the above warnings.
It is hoped, under such circumstances, that the more immediate need for a standard can be met.
The workshop recommends that the specification of a modelling subsystem for a simple but widely used class of applications, be undertaken simultaneously (if possible by the same group which will address the specification of the graphics subsystem).
Transportability and device independence were also addressed from the user's point of view. What should be a recommended methodology for design and implementation of application programs which use computer graphics?
The workshop felt that the basis for such a methodology could be derived through a careful analysis of a representative set of all application programs and a subsequent categorization into what was called program, or skeletal, structures. A preliminary effort toward this goal was initiated, by a subgroup during the workshop.
Several relevant aspects were addressed.
The workshop feels that this is a very promising area for practical results: applicable to design and implementation of application programs using graphics. The workshop encourages research in that direction.
Interactive computer graphics has long been an informal discipline; it should benefit from advances in computer sciences related to more formal disciplines like program correctness, structured programming; it should deal at the specification stage with formal abstractions {functions and data).
The workshop had a thorough presentation by a leading researcher in the field of specification techniques used to specify Data Abstractions.
The workshop recognizes the need for formal techniques to specify concepts unambiguously.
The workshop recommends undertaking an attempt at formal specifications of concepts and constructs in graphics systems.
The workshop suggest that any official proposal of a graphics package as a standard should be accompanied with a set of specifications which provide a reasonable level of formality.
One subgroup of the workshop also addressed the issue of what was called an interim standard. This seeks to respond to those expectations, by the community at large, for immediately and practical recommendations by the workshop. The thread was to consider features of existing graphics packages. As an exercise two packages were examined; to see how they could be altered to improve portability and device independence. Features were characterized such that, when used, they would imply global differences or local differences in the same application using either one of the two packages. It was felt to be outside of the scope of the workshop to make recommendations directed to users of specific packages. However, from this analysis a recommendation could be made.
The workshop recommends that users, designers and implementers of graphics packages should carefully examine their packages to determine where they differ from the conceptual framework proposed here. They should indicate dangerous areas, and how they could be avoided.
Remark: Some packages could be reasonably transformed, to fit in the conceptual framework. This conceptualisation does not necessarily invalidate previous works, but seeks to clarify what the designers had in mind. Transportability should be added.