Conversational programming system for NORC
Conversational FORTRAN programming system
In the first part of this presentation, I would like to spend a few minutes on background information on NWL and on the goals and history of the project of which DISPLAYTRAN is a part so as to put the work which will be described in its proper perspective. I will also describe briefly several other parts of this project which are of interest. The second part of the presentation will discuss DISPLAYTRAN at some length.
NWL is a laboratory working under the direction of the Bureau of Naval Weapons. It has as its function research and development directed toward weapons and weapons systems as well as testing and evaluation of these. This work led quite early to a requirement for high speed, high precision digital computers, especially for work involving trajectory computations. As a result, the Aiken MARK II relay calculator was installed in 1946. This was followed by the MARK III in 1950, and the NORC in 1955. Currently the major computing facility is the STRETCH. A Stromberg Carlson Charactron tube printer-plotter was installed on the NORC in 1959 and still serves on its principal output unit. A similar device is available off-line to STRETCH users and is extensively used for both printing and plotting. Besides the large digital facility, there also exists at NWL a small analog computational facility, which is used in a variety of simulation problems arising in weapons development. This is currently being expanded.
Besides its continuing interest in having the most up-to-date computing equipment available for use by its scientists and engineers, NWL has, of course, long been interested in those developments which would allow this hardware to be more effectively used by the programmers, scientists and engineers of the laboratory. While currently most programming is handled on a closed shop basis, by a staff of about 65 professional programmers, it was recognized several years ago that studies such as those at project MAC indicated both the potential usefulness and practicability of direct access and reactive or conversational computing.
A project was therefore set up in the Programming Systems Branch of the Programming Division to investigate and study these developments and to see how they could be exploited at NWL to improve the productivity of scientific manpower. It was early decided to in particular explore the utility of display or graphical devices. To further this work, it was decided to obtain a limited number of graphical terminals. A 360/40 system was acquired for the purpose of driving these terminals and also to replace one of the two 1401's then installed in the laboratory for peripheral processing. The configuration is as shown in this slide. (Fig. 1). You will note that besides the two user terminals, there are two terminals to be used respectively for preparing STRETCH system input tapes and printing STRETCH system output tapes. After the acquisition of this system, the Data Processing Division of IBM expressed an interest in joining with NWL in this project and have since January 1965 actively joined with NWL in the work undertaken as part of it.
In pursuing this project the four areas in which NWL feels particularly interested are summarized on this slide. (Fig. 2). Currently most emphasis is being placed on the first two areas with lesser amounts on the latter two. Extract: Utilization of Existing Systems
Utilization of Existing Systems
So that we might gain practical experience as soon as possible we decided to utilize several existing systems - principally the QUIKTRAN and ALPINE systems among the conversational systems, and the MIDAS simulation language (implemented in a batch mode on STRETCH) to explore the interesting area of digital analog simulation languages as a means for aiding in the solution of problems typically handled on analog or hybrid computers. This language has found acceptance with our engineering staff.
In employing QUIKTRAN, our principal goals were to evaluate the utility of conversational FORTRAN to the scientists, engineers and programmers of the laboratory, and to study the typewriter type of terminal. We permitted, a variety of people from the scientific, engineering, and programming staff to use the system under controlled but realistic conditions. While I do not intend to go into the results of our study here, they were quite interesting and have had significant influence on our current course of action.
The ALPINE system was used to gain early experience with a graphical input-output terminal. The problem we chose to implement was a simplified version of one of long standing interest to NWL, the determination of drag function for free fall weapons. The experiment showed the feasibility of the approach for solving this type of problem, but more important here showed the power of the FORTRAN oriented graphical language available on this system. This language, along with our experience with graphical output devices, forms the basis for the user oriented graphical language embedded in the DISPLAYTRAN system.
With this background established, I would like to spend the remainder of the time in describing three user oriented systems that are being developed as parts of this project. We call them AAPl, OLDAS, and DISPLAYTRAN. Each attempts to adapt the digital computer to the user for a particular class of problems using the 2250 as a terminal.
To summarize, let us briefly review the experiment and results. We set out to devise economically feasible means of increasing the throughput of the scientist at NWL through the use of conversational computing and graphics. To achieve this goal, we elected to first experiment with available systems to direct or design of suitable tool for experimentation that could be developed for the S/360. Our experiments with QUIKTRAN were encouraging, and influenced the development of AAPl as a local offering - the QUIKTRAN system is still operational. Our experiments with ALPINE Graphic FORTRAN were encouraging also, and in conjunction with QUIKTRAN results, led directly to the development of OLDAS and DISPLAYTRAN for further experimentation in an operational environment. We believe that the experiments have been successful, both in demonstrating that the general techniques are useful and in clearly identifying a useful product for continued, productive experimentation.
in [ACM/IEEE] Proceedings of the SHARE Design Automation Project Annual ACM IEEE Design Automation Conference 1965 view details