The group adopted as its purpose, the objectives for methodology suggested by Newell in his position paper Realistic objectives leading to a graphics standard. These objectives are paraphrased here:
The group adopted Newell's notion of program structure as the basis of its methodology for achieving its objectives. A preliminary (loosely organized) effort at designing program structures (i.e. program dynamics, algorithms) was initiated. Several matters were defined relevant and addressed.
Several different cuts were made at useful sets of configurations. The initial set was adopted from the West German position paper by Encarnacao & al. This set was put aside on the basis that it did not appear to provide sufficient distinction between facilities that would be useful for discerning differences in program structures.
A second set was proposed (Newell) in terms of the spanning parameters rather than the terms of the architectural configurations of the first set. The second set used two parameters to distinguish four classes of facilities:
In both cases, these parameters were considered binary in nature (i.e. a facility either did or did not provide for selective changes; similarly for apparent motion).
Later considerations revealed that the motion parameter was of the right sense in that it neatly spanned but was rather insensitive to important sub-distinctions within each binary value. Bandwidth in support of image change was considered as a possible alternative. Preliminary analysis concluded two points about it:
If one also retained the parameter of selective changes in the picture, the set of differentiable facility configurations has become too large to be pragmatically useful for the desired analysis.
The ad hoc decision was to proceed in parallel on two paths. Use the set based on motion and selective change to devise program structures to polish the process. Concurrently, continue to analyze architectural and functional characteristics of graphics facilities in an attempt to discover a set of parameters (richer than the working set but less prolific than the set generated by bandwidth as considered above) capable of spanning the facility universe.
A convenient option was exercised initially as a means to providing examples of applications to be analyzed in order to design program structures. We chose the set appearing on pages 8-9 of Some notes on Software Standardization (Sancha, 20 Oct. 75). A few others were added on the intuitive basis that the convenient set was somehow not fully representative. Five applications were analyzed to one degree or another. Attempts were made to design program structures. The applications addressed were: scientific animation (animated graphs); engineering drawings (mechanical) with hard copy; cartography; data presentation; and interactive preparation of schematic drawings.
One application was abandoned (data presentation). Scientific animation yielded two program structures across the four facility configurations. Engineering drawings (mechanical) with hard copy and cartography were determined to be satisfied by the program structure for simple plotting given as an example by Newell. The analysis of the interactive schematics application exposed serious deficiencies in the methodology of the User Group. Time was taken to correct this and is reported below. When effort using revised methodology was re-directed at this application, one facility context and another (fourth) program structure was devised. In the process of choosing the interactive schematic applications to work on next, it was discovered that the set of applications, conveniently chosen, was not suitable as they did not span the application universe and were not parametrically differentiable. That they needed to have these properties was a late-coming awareness such that full group attention was not brought to its resolution.
However, from the notes of the group, the following insights about the application taxonomy has been derived. Graphic applications appear to be differentiable by considering : the nature of the graphic content of the image; behavior of the user-facility combination regarding image change; and the behavior of the application systems (user + facilities + application program) regarding influence on the environment into which the application system may be connected. The first consideration is an image type parameter with four values : schematic, geometric, pictorial and graphic. The second consideration yields a binary parameter : interactive and non-interactive. The third consideration also yields a binary parameter : on-line and not on-line.
These considerations and the three derived parameters yield sixteen types of application that span the universe especially with the assumption that text may be present in any type of image as may be desired within the application. This latest material was not discussed, as stated/ within the sub-group.
The analysis of applications for program structures yielded another insight. It began to be noticed that sub-structures were occurring in program structures that were useful between application-configuration combinations (i.e. potentially portable units). One undefined sub-structure, input, began to be addressed. Much discussion exposed that there were several views to bring to understanding input and, subsequently, to supporting various inputs within the application system implementation. This insight led to joint meeting and agreement between the input and user groups on this general view (with small modifications). The results of this meeting led to further work at the conceptual level by a revised input group {joined by a member of the user group) as reported by Foley to the whole group.
The initial effort to derive program structure proceeded on an ad hoc basis. Although the process yielded structures, there was a concern as to whether or not what had been derived represented the application fairly. Intensive discussion concluded that an organized methodology was required as opposed to an ad hoc process. Borrowing from software engineering, the following methodology was loosely organized and used:
It was confirmed again that the necessary process is time consuming and poses problems for organizing enough effort at this essential task to produce a sufficiently useful result.
Crude drafts of initial program structures and the applications facility matrix are attached for illustrative purposes only.
It is asserted that analysis of an adequate set of program structures derived in the context of applications and graphic facility configuration will lead to recommendation on codes of practice that will promote improved portability and implementation of application programs that use computer graphics.
Selective Changes to Picture | No Selective Changes to Picture | |
---|---|---|
Motion | Vector General LDS-2 Adage |
Simple Refresh Buffer Picture System |
No motion | Selective erase Storage tube T. V. |
Plotter Storage tube |
The group identified four preliminary structures but realized that further subdivision was needed.
Unfortunately, the notes of this group are somewhat incomplete. However, as we feel that this was a real and significant effort to analyze graphic program structures, we include the working papers as an aid to those interested in further investigations in this area.