The 1989 master's thesis on how software might help a person think a whole problem through — answered in three moves: establish, organize, represent.
This thesis is concerned with the conceptual development of occupational applications – computer software designed to encourage the use of procedures that help invent, develop, control, and analyze ill-structured business decisions. Occupational applications are built upon a common framework called EOR (establish, organize, and represent). The EOR methodology is a synthesis of practical methods synthesized from architecture, scientific method, industrial design, engineering, social sciences, and problem solving. EOR is employed throughout the entire life cycle of a problem-solving task, from conception to communication.
As a primary target of EOR, occupational applications are conceptualized to meet the demand for two main opportunities:
Information technology must evolve to help solve ill-structured decisions in today's organizations. Problems faced by individuals in organizations are continuing to become more complex. There appears to be an opportunity to create tools which aid managers with non-programmable decisions and also add greater observability to their decisions. A "process push" can occur in which an explicit problem solving methodology (EOR) can be embodied in computer technologies at a reasonable expense in order to add overtness to decision making. What are the characteristics of a computer software tool that will aid business managers in resolving unstructured problems?
With the growing quantity of specialists or knowledge workers, it is logical to expect that these workers will need sophisticated computerized tools, like occupational applications, to model their entire domains of expertise. What will be the characteristics of specialists who work in the coming organizations and what information tools will they need?
The development of occupational applications is founded in two primary design philosophies – qualitative research and the concept of augmentation. Occupational applications should fulfill verbal and intuitive problem solving needs. Occupational applications should support, not replace, the user's unique intellectual capabilities. Analysis for the foundation methodology, EOR, was completed by a literature review and interdisciplinary interviews conducted on the campus of Auburn University. The reader should note that although occupational applications are well defined, the research focus is EOR.
Qualitative methodology provides tools for understanding the particular, the individual, and the unique as well as perceptions, meanings, and social realities. Qualitative data "are a source of well-grounded, rich descriptions and explanations of processes occurring in local contexts" (Miles & Huberman, 1984). Miles & Huberman describe qualitative analysis as three concurrent activities: data reduction, data display, and conclusion drawing. Business managers use similar methods to process diverse decision inputs — interviews, observations, meetings, letters, past experiences. Most of these inputs are verbal, not mathematical, in nature. Computer-supported qualitative methodology aims to improve a person's decision making by enhancing his procedures.
EOR and occupational applications are constructed to fulfill the concept of augmentation. Through interaction, the system should "augment" rather than replace human intelligence. Bounded rationality refers to the limitations present when there are several alternative choices available at one time; EOR aims to extend the bounds and eliminate some constraints by providing an environment for problem solving. In 1963, Douglas Engelbart published a paper called "A Conceptual Framework for the Augmentation of Man's Intellect":
By augmenting man's intellect, we mean increasing the capability of a man to approach a complex problem situation, gain comprehension to suit his particular needs, and to derive solutions to problems… We refer to a way of life in an integrated domain where hunches, cut and try, intangibles, and the human feel for a situation usefully coexist with powerful concepts, streamlined terminology and notation, sophisticated methods, and high powered electronic aids. (Engelbart, 1963)
Augmentation, not automated machines to replace thinking, is stressed because of the unique abilities of the brain and the attributes of the computer. As Francis Crick observed, a computer processes information at a rapid pulse rate and serially, while the brain is slower but handles information on millions of channels in parallel; a computer can do long arithmetic a human cannot, yet "human beings can recognize patterns in ways no contemporary computer can begin to approach" (Crick, 1981). The EOR philosophy supports the synergistic roles achieved by computer and mind.
The importance of research that stresses intellectual augmentation resounds in many ways. Herbert Simon: "there are no more promising or important targets for basic scientific research than understanding how human minds, with and without the help of computers, solve problems and make decisions effectively" (Simon, 1987). K. Hooper (1985) notes that the computer's dynamic, interactive qualities suit it to serve at a high level, as a helper in developing ideas — "a setting for thoughtfulness, and an arena in which to communicate." McPherson (1984) adds that there is "a dialogue between the analytical and the holistic parts of the brain during the design process," and the computer can take part by producing pictures, with all the information color-coded. Yet the field has historically posed a challenge: many aids are too complex to use, and many depend on the user's ability to decompose the whole problem into constituent elements — difficult where decomposition substantially alters the original problem (Wickens, 1984).
Occupational applications are highly specialized software programs crafted to suit the qualitative analysis needs of a specialist. They emphasize successful methods relative to many different professions. Within the business community this is significant because of the current emphasis on quantitative analysis: counting is important, but too much emphasis has been placed on it. A specialist's professional impact is directly related to the effectiveness of his problem solving techniques — his ability to master qualitative in addition to quantitative techniques. Specialists use occupational applications during a session, the activity of on-line problem solving directed towards the solution of a task.
The foundation of occupational applications and qualitative processing is a software-based methodology called EOR – an acronym which stands for establish, organize, and represent. EOR is a generalized system which provides questions, cues, and visual forms to assist the user in producing a definite piece of work – a task. A task is a relatively independent concept, problem, opportunity, concern or challenge; it is complex because of its numerous variables and many possible solution combinations.
EOR starts with language – with words. Within EOR, a unique collection of words is the symbolic expression of a task, and these words gain "meaning" due to the way they are codified via structured grammar. EOR has three parts to guide the user with the discovery of task attributes and relationships: tutor, events, and a visual model of the task called a form.
A survey of "problem solving" concepts is a necessary preliminary for EOR, because techniques from design, the scientific method, engineering, and problem solving should be combined into one all-inclusive methodology. The definitions and characteristics across disciplines appear to be common; all support EOR's method of establishing, expanding, isolating, and integrating.
The definitions are many and varied — design, creative thinking, decision-making, task resolution. A sampling: problem solving consists of objectives, controlled variables, uncontrolled variables (the environment), and the relationships among them (Ackoff, 1978); formal problems have givens, operations, and goals (Wickelgren, 1974); "the way in which problems are represented has much to do with the quality of the solutions that are found" (Simon, 1987); "a problem is the difference between what one has and what one wants" (De Bono, 1973). The recurring conclusion: the power of the problem solver is determined by the effectiveness of its problem solving techniques (Newell, 1969) — which is why techniques from the literature can be embedded in the Tutor dialogue to channel responses for the best results.
Simon (1977) identifies four principal phases of decision making: intelligence (finding occasions for a decision), design (finding possible courses of action), choice (choosing among them), and review (evaluating past choices). Decisions are programmed to the extent they are repetitive and routine, and non-programmed to the extent they are novel, unstructured, and consequential. Non-programmable decisions — exercised through experience, insight, and intuition — can be improved by training in the generalized operations of decision-making (the military "estimate of the situation" is one such checklist) — "but aids like this don't go far enough." As Mintzberg (1973) observed, the manager "must know how to step back from his data so that he can see… a set of broad, powerful models that provide simple but accurate pictures of various aspects of his reality."
Zand (1981) frames policy formulation as a special case of managerial problem solving, with an orderly sequence: scan status, generate goals, define problems, design alternatives, predict consequences, evaluate alternatives, review balance, select policy, implement. "Quantitative" and "qualitative" tasks sit at opposite ends of a continuum, with many variations along it.
Designing is "a hybrid activity of art, science, and mathematics meant to initiate change in man-made things" (J. Jones, 1981). Pile (1979) analyzes the design process as: identify and clearly state the problem; marshal the bearing information (research); take the creative step (holding the problem in mind while calling up assembled and general knowledge); test each proposal in concrete form; evaluate; implement. Architectural design is offered as the exemplar of ill-structured problem solving: Simon notes that an architect begins with general specifications, then "the emerging conception provides continual feedback that reminds the architect of additional considerations." "Most corporate strategy problems and governmental policy problems are at least as ill structured as problems of architectural or engineering design."
Thinking is often treated as a black-box mystery: inputs and outputs are recognized, but little is taught about the processing. "The best thinkers employ both visual and language expressions in correct spatial-temporal dynamics; human processing is basically the same for everyone, although some more frequently exercise one mode or another." One purpose of this research is to determine which processing methods can be synthesized and shared across occupations — fusing the problem-solving methods of design, business, the sciences, and engineering into a single task solution: the act of fully identifying all the possible parts, structures, and if-then interactions of a task.
Deprivation leads to the recognition of a need, which prompts an EOR session to define (establish), organize, and represent the task — moving from the most abstract part of problem solving (the need) to the most tangible (physical implementation and follow-up). How well and efficiently a task is resolved is affected by the individual (the knowledge worker), the task itself (its attributes and attribute relationships), and organizational imperatives (homeostasis, growth, or response to threat). [Figures 1–2 in the original — the problem-solving sequence and the variables influencing task resolution.]
↑ ContentsSuccessful thinking is directly related to a conducive environment. "A good work environment is like a cocoon in which a metamorphosis takes place." The environment is physiological and psychological — lighting, setting, noise, seating, social and personal factors — elements outside the EOR system that can nonetheless be isolated and controlled. The relationship between mind, body, and computer forms an open system of concentric rings representing decreasing levels of abstraction.
Forward thinking is "a steadfast, perceptory ability in which omnidirectional awareness is accompanied by an alert attitude to take full advantage of the moment." It is intelligent attention: focused yet enveloping all surrounding activity without distraction. "No thought or action can ever occur at any other time than right now." It takes two forms at the ends of a continuum — inward (as in meditation) and outward (as in athletics) — distinguished by setting and the phenomenon of motion.
However heterogeneous, individuals possess similar mental attributes — they can tolerate ambiguity, speculate, set goals, tire. Hofstadter (1979) lists the essential abilities of intelligence: to respond flexibly; to take advantage of fortuitous circumstances; to make sense of ambiguous or contradictory messages; "to find similarities between situations despite differences which may separate them; to draw distinctions… despite similarities which may link them; to synthesize new concepts by taking old concepts and putting them together in new ways." Gardner (1983) supports several different intelligences — domains of potential competence — over a single general-purpose mechanism. EOR's interface should have the potential to develop different intelligences; visual intelligence, which most business managers lack, "can be developed with systematic effort," and within EOR the verbal and visual intelligences are directly linked.
Short-term memory is working memory. Engelbart (1961) states that "no human being can hold very many concepts in his head at one time. If he is dealing with more than a few, he must have some way to store and order these in some external medium." Short-term memory has very limited capacity and is easily overloaded; relief comes with closure, the completion of a task. EOR's adjunct storage captures information about a task that would otherwise be lost — once recorded, elements can be temporarily forgotten so the user can attend to other parts, and recalled when needed. As Vannevar Bush wrote in 1945:
He has built a civilization so complex that he needs to mechanize his records more fully if he is to push his experiment to its logical conclusion… His excursions may be more enjoyable if he can reacquire the privilege of forgetting the manifold things he does not need to have immediately at hand, with some assurance that he can find them again if they prove important. (Bush, 1945)
Moving from a task's conception to words, and from words to a visual form, are successive abstractions — each a simplification with its own rules. Hofstadter (1986) cautions that the brain does not manipulate symbols by an act of will: "we are not symbol manipulators; in fact, quite the contrary, we are manipulated by our symbols." Yet humans took "a significant step toward harnessing the biologically evolved mental capabilities" when they developed the means for externalizing symbol manipulation — "particularly in graphic representation, which supplements the individual's human intellectual capability — automated external symbol manipulation." To some extent, "only those things that we explicitly set to symbols are real."
↑ ContentsTutor is the dialogue subsystem of questions and cues that lead the user through each event; "the process resembles the Socratic method." Words (answers) are grouped by adverbial question: "how" and "why" questions describe infrastructure and causality; "what" discovers information about the task and its parts; "where" conveys spatial relationships; "when" describes the temporal nature; "which" and "how much" disclose attributes. The idea of a tutor "is patterned after mondo, an ancient Zen method of dialogue between a Buddhist monk and pupil; mondo leads the pupil to satori (enlightenment)."
Tutor's assistance is adjustable: "there is a facility to add questions to Tutor as well as regulate assistance by varying the tutor scale from 'full support' to 'no support.'" As familiarity grows, the user can change EOR from a how-to-think program into a system rearranged for his own domain — "in this way, the system provides creative accretion." Responses are stored in a database so the user "does not have to remember all the many facets of the task and can concentrate on forward progress." Tutor offers a Process-Purpose Monitor (a means-ends gauge, after Simon's General Problem Solver, measuring the difference between intent and current state) and its questions and cues.
Tutor induces alternating expansive and contractive thinking. "Determining when to change the alternating pattern of expansion and compression is the essential ingredient for successful problem solving." De Bono's "Six Thinking Hats" offers a parallel: white (facts), red (feelings), black (logical negative), yellow (positive/constructive), green (lateral thinking, change), blue (defining the problem and asking the right questions).
"There is no way for EOR to determine if the user has 'correctly' answered Tutor's prompts because there is no one correct answer or task solution." Questioning makes the parts of the task apparent by drawing attention to the subject matter. "Absolute certitude is not possible for any task. In fact, doubt is a healthy condition to maintain." Where two viewpoints explain the same phenomenon, apply Ockham's Razor — use the one with the fewest assumptions that explains the most.
Establish is the first event — the task is diagnosed and identified at its widest scope, "setting up the boundaries." Research from perception suggests holistic perception precedes atomistic division: "pattern seeking… is the first step of a two-step process: pattern, then analyze" (McKim, 1980). "The term holistic describes a mode of information processing in which the whole is perceived directly rather than as a consequence of the separate perceptual analysis of its constituent elements" (Wickens, 1984). Tutor techniques for establishment include: Stating Objectives, Strategy Switching (letting spontaneous thought influence planned thought), Brief Statement (selective ignorance vs. exhaustive review), Causal Relationships, and the Morphological Approach (Zwicky).
After establishing, expansive or "soft" thinking is facilitated — metaphorical, holistic, subjective, intuitive, deductive. Quantity, not quality, is the objective: "judgment is suspended during the generative stage of thinking in order to be applied during the selective stage" (De Bono). Expansion draws on metaphor, analogy, and imagination, and on a catalog of techniques: Osborn's Self-interrogation List (put to other uses? adapt? modify? magnify? minify? substitute? rearrange? reverse? combine?), Systematic Method – Synthesis (Jones), Systemic Testing, Synectics (Gordon — making the familiar strange through direct, personal, symbolic, and fantasy analogies), Innovation by Boundary Shifting, Imaginate, Thesaurus, Reversal, and Random Word Stimulation.
"In the idea of combinations lies the essence of genius and mediocrity." Because few ideas are truly novel and everyone has access to similar information, "the point of variation lies in how certain elements are combined." The earliest origin of expansion-through-combination is Ramon Llull:
On a desolate mountain in 1274, Ramon Llull… experienced a divine illumination in which God revealed to him the Great Art… Llull believed that in every branch of knowledge there are a small number of simple basic principles or categories that must be assumed without question. By exhausting all possible combinations of these categories we are able to explore all the knowledge that can be understood by our finite minds… In every case, the technique is the same: find the basic elements, and then combine them mechanically with themselves or with the elements of other figures. (Gardner, 1958)
"A complete enumeration of possibilities is carried out by a system function to develop a separate listing, one by one, of each potential combination." In one sense, "unique combinations are limited only to the user's ability to identify the extreme options; in a multi-attribute task, just one 'far-out' variable can have profound effects on the entire task structure." The number of combinations is limited by three means — links (compatibility pairings), weights, and assessment of validity — and by Functional Innovation (J. Jones).
Isolation promotes analytical, atomistic thought. "Reductionism is the belief that complex tasks can be reduced to simpler ones" — but going from smaller to larger, "the higher level is not just more of the same — there is additional information (properties) which tells how it operates," so "the key to isolation lies in retaining meaningful breakdowns." Pirsig's knife is the image: "analytical thought" that splits the world into parts, "finer and finer until it is what you want it to be." The isolation catalog includes Squeeze and Stretch, Classification of Design Information, Mental Photograph, Dictionary Definition, Systematic Search, Analysis – Basic Questions (Koberg), Systems Engineering (inputs/outputs, system boundary, sources and sinks), Hierarchical Groupings, Determination of Components (Alexander), Page's Cumulative Strategy, Attribute Listing, and the Matrix.
Integration retrieves, refines, and recombines the words catalogued from establishment, isolation, and expansion. Its key concept is centering — "the process of synthesizing different points of view," viewing opposing parts "as two variations of the same reality seen through different viewpoints." Resolution occurs when the model matches the required objectives. Techniques: Interaction Matrix and Interaction Net (J. Jones), AIDA – Analysis of Interconnected Decision Areas, and System Transformation.
"A task solution is recognized when the user feels confident and correct." On the way, the problem solver experiences a hedonic response — "a feeling of absolute certitude accompanying the inspiration" (Gordon). A guide to the shape and number of elements: "the less we understand something, the more variables we require to explain it" (Ackoff); elegance of the solution = multiplicity of variables ÷ simplicity of solution (Gordon). Evaluation methods follow: Systematic Design – Evaluation, Selecting Criteria (operationism, fail-safe direction), Ranking and Weighing, and Ideas-Objectives Comparison.
Follow-up evaluates success or accounts for failed elements: were the relationships correctly structured? were all the parts identified? was isolation specific enough? It documents why assumptions were incorrect and updates the task model — feedback that EOR can store so that "rules can be developed, and learning is facilitated."
[Page 69 of the original manuscript is absent from the source.]Representation provides the visual depiction of a task. During most sessions the user works on several forms at once, each a feasible solution; the system can replace or merge any part, and the dimension of time is communicated by animated change in the form. Forms provide visual data retrieval — isolating parts and relationships, then panning out to view the whole. "Views of a form can be seen from north, south, east, and west… like rotating a globe on its axis to look all the way around" — echoing De Bono's counsel to "deliberately rotate attention over all parts of the problem, especially those which do not seem to merit it."
"By thinking in pictures, systems designers and analysts are providing additional communication channels between the analytical and holistic halves of the brain" (McPherson, 1984). The thrust of the chapter: "Visual and verbal models of thinking and communicating are complementary; one is not higher than the other" (McKim, 1980). Verbal and mathematical symbols are strung together linearly, enforcing certain thinking operations; "visual imagery, by contrast, is holistic, spatial, and instantly capable of all sorts of unconventional transformations and juxtapositions." The thesis closes by inviting the comments of Douglas Engelbart — the author of the augmentation idea that is EOR's root.
↑ Contents