28 June 1961
Memo to: File, Program A.
From: D. C. Engelbart
Subject: THE CURRENT PICTURE OF PROGRAM DEVELOPMENT FOR "AUGMENTING HUMAN INTELLECTS."
This is sort of a "snapshot" of the state of our thinking. We find that our ideas relative to the Program launching are developing much faster of late---possibilities that "feel right" are fitting into place, and a picture of definite action is emerging that seems to fulfill very satisfactorily the various short and longrange needs that we previously have established (relative to such as longrange program objectives, shortrange useful developments, manageable problem size, appeal to sponsors, etc.).
This memo therefore represents a quick picture at something that is now in a state of transition. Most of the details that could fill out the memo are subject to change, so most of the memo has "that sketchy look." We welcome feedback comments and questions, and we hope to have more definite data in the near future.
We are concentrating nowadays on getting into shape a "launching plan" upon which to base a coordinated promotion effort. We hope that serious promotion work can begin within a few months. The program ultimately will contain research, development, and application activities, but we plan to begin only with research. This research program would involve four basic "components": (1) the effort toward developing the innovations in the individual's way of doing things that are expected to make him more effective (we call this our Synthesis function); (2) the effort toward developing measurement and analysis techniques by which we can study and evaluate the human activity that we wish to make more effective (our Analysis function); (3) the effort to develop testing facilities that provide support for, and common grounds for interaction between, Synthesis and Analysis (we are thinking of a computer and special testing laboratory); and (4) the forward and outward looking efforts that provide scouting, guiding, and coordination for the other effort (our search effort).
First, I might mention an old statement (Jan. 2, 1959) that still expresses very well the nature of our program. It says, "Project A, devoted to the development of intellectual methodology, optimally automated." This is essentially a program to develop new methodology for people to use in attacking complex problems. The methodology is used by a human, in conjunction with associated artifacts and language, within what we call an Individual Symbol Manipulation System (i.e. ISM System is composed of a human and the "augmentation means" that he has been trained to useand we break the latter down into Language, Artifacts, and Methodology). Therefore, we essentially have a systemdevelopment program, in which we intend to develop innovations in Language, Artifacts and Methodology such that the human's effectiveness in working on complex problems is much improved. We expect the innovations in the Artifacts (especially computers and displays) of this ISM System to make practical some radical and extremely effective innovations in the Language and Methodology aspects.
The question now is, how to organize and launch a program (Program A) that can do a good job of fostering these innovations and getting them adopted by people who have to solve problemsall of this to be consistent with SRI's charter, longrange plans, and, of course, the availability of sponsorship.
III. ROUGH PICTURE OF PROGRAM WE MIGHT TRY TO DEVELOP.
Eventual composition of Program A. We expect to have two primary types of activity, Research and D and A (Development and Application). It is not clear how these might be separated administratively, but it looks as if their support would be found generally from different sources. me Research activity would probably be government supported, and would be doing the more basic and longerranged studies and developments. The D and A activity would be oriented toward developing working "packages" to meet real life needs, utilizing the basic findings of the Research activity. D and A work would be sponsored by specific clients who have needs for specific help in automating methodology in some specific kinds of intellectual tasks. There would be considerable advantage in having both kinds of activity "under one roof" from at least two
points of view. First, there would be advantage to the workers in both groups from the cross stimulation, and secondly, sponsorship in both areas should be easier to get because of the presence of the other activity.
Within each activity, there will be the two functionally distinguishable aspects, analysis and synthesis. This being effectively an "engineering" job (to improve the performance of a system), the "synthesis" aspect will be dominant. We must remember, though, that special developments in the means and techniques for analysing the system will be a very strong secondary need. We also speak of a third aspect to our general activity, the aspect which includes the advanced scouting and guiding for the whole program, and which is somewhat separate from both the Research and D and A activities. We call this the Search aspect.
So, our Program would consist of two types of activity, Research and D and A, and within these would have the three principle functional aspects of Search, Analysis, and Synthesis. The question is, how to get on the air with all of this?
IV. HOW PROGRAM WILL PROBABLY GET STARTED.
While the Program will ultimately deal with the problems of automating the methodology of the general problem solver, it would seem best to start the Program in a restricted area. If this initial area of attention were chosen properly, it could lead nicely to the general work and at the same time provide the large advantage of giving us good initial experience with concepts, problems, and innovations basically related to those of the general problem, but here confined within an area of consideration that is small enough to allow good coverage by our initial working force.
We think that we have isolated an initial area of consideration which has promise of being extremely good for the purposefrom a number of points of view. Essentially, it involves studying possibilities for "augmenting" a special type of worker. His intellectual activity is really quite general, but embraces a range of subject matter that is much more manageable by us than would be the general case. There is considerable demand for this kind of worker,
and more growing, so any success we have in devising ways to help him will be attractive to sponsors, whether or not we can generalize upon these 'ways' to help the general problem solvers. Also, excitingly enough, his augmented capability (the worker we are thinking of studying) will be very useful for us, too. The "worker" about whom we speak is a computer programmer, operating in a special way that we will elaborate upon at a later time, after we work out more details.
Within this initially restricted area of concern, we can set up the full range of Research activity, and indeed, can expect to do D and A work, too. The Analysis activity can begin with initial concentration upon one type of worker, and can gain much valuable experience before it is faced with the whole range of activities of the general problem solver. There still must be worked out the means to study a subject, evaluate data within a meaningful analytical framework, and thus evaluate a subject's performance in specific or general aspects. Also, we will need to analyze the subject's Methodology for specific areas where improvement can be sought.
The Synthesis activity can begin giving fairly general coverage to the limited area of consideration, and should gain a lot of valuable experience in redesigning a person's intellectual methodology. The Search activity, in its guidance and coordination of Analysis and Synthesis, will also benefit from a more restricted area of activity. Also, its scouting into the "future trails of activity" will have a much restricted frontier to operate upon, which will be a distinct advantage. (It is assumed that this frontier can grow at some steady and controlled rate, until the area covered is quite general.)
An important part of the Program from the outset will be the experimental facilities that allow us to test the products of Synthesis-- i.e. to apply the techniques developed by Analysis upon the innovations developed by Synthesis. We expect to develop a simulation facility that will include a general-purpose computer and special, real-time matching equipment. We want test subjects to be able to work in a physical environment that is meaningfully controlled in a rather complete sense, including Artifacts that function for him in the way prescribed by Synthesis. While the test subjects are working in this environment
our Analysis people should be able to get very complete observational data upon their behavior. The initial facilities should be designed with an eye for their expansion and generalization, as well as for keeping up with new Artifact innovations.
To summarize how the Program will probably be structured at launching time: Initially only Research activity (D and A to come along when appropriate later). Pick a restricted area of study that can grow gracefully into the general area, and set up Search, Synthesis, and Analysis activities in good strength to operate initially in this restricted area. Set up a good, flexible test facility, involving a computer, special test environment, and appropriate matching equipment.
V. NEAR-FUTURE WORK FOR THE SRI, PROJECT-A GROUP.
Primary interest is to get organized for launching the Program. This means working out more details in several areas. The largest initial costs will be involved in setting up the test facility, and so we shall expect to provide more details for this in the initial plans and estimates than for most of the other parts. We are currently assessing our needs, and surveying the available computers and display systems. We also need to work out in more detail the nature of our attack upon our restricted area, so that the initial work of the Analysis and Synthesis people can better be planned, and so that we can make reasonable prediction of the specifications they would like applied to the test facilities. We hope to end up, within a couple of months, with a fairly detailed outline that is suitable to base a proposal upon.
About sponsorship. We choose now to plan the initial phases of the Program as if we were to have onesource sponsorship from someone interested in the basic objectives of the Program. We shall not constrain ourselves particularly to keep the plans modest. This will provide the ideal conditions under which to start, and if it sells we shall be very pleased. If we can't sell it in this form, we shall at least know that we couldn't get things just exactly as we wanted to, and we will also have learned from our approaches elsewhere pretty much how to structure something that will sell.
About possible cooperation with Stanford. It is highly likely that, if we were to get a program of this sort going, with facilities as mentioned,
there would be a lot of advantage to both SRI and Stanford to get faculty and students from different departments at Stanford to participate. From initial talks with Bill Linvill, it is quite evident that he would like to cooperate, and that he would expect some of his systemstheory graduate students to take a very active interest in some of our basic problems. We shall explore these possibilities a little further, to see if advantage is to be gained by including plans for such cooperation in initial proposals. It might be pointed out that there are more than enough good problems in this area of Research to go around, and that the D and A activity that should arise later could become a big and continuing activity that Stanford is not particularly geared to participate in. Therefore, getting early mutual organization with Stanford in the Research area would be very beneficial to the progress of the Research, and if that goes better, the later D and A activity would be all the more fertile.
D. C. Engelbart