| |
Douglas C. Engelbart, "Special Considerations of the Individual
as a User, Generator, and Retriever of Information" 1
American Documentation, April, 1961, pp. 121-125
DOUGLAS C. ENGELBART
Stanford Research Institute, Menlo Park, California
Received 23 December 1960
I would like to remove at the outset a possible erroneous expectation which
the title may have raised. I am not going to talk about what individuals
need or expect from your documentation systems. The purpose of this paper
is to bring out a basic sort of relatedness between your discipline and
a separate discipline which is needed relative to the individual's information-handling
problems. I want to invite you to consider turning some of your talents
and activities toward these problems. I should also hope to benefit from
reactions and correspondence from some of you, and perhaps to discover some
among you who share with me a basic conviction that more research attention
should be focused upon the individual.
About Your Systems
To help relate our two areas of interest, I will make a few observations
about your systems and your discipline.
I view your systems as being designed to provide a special kind of service
to certain kinds of people. When an individual composes a message to society
that passes the editorial filters and becomes published, your system provides
a means for distributing and storing copies of this in a way intended to
make it available to other individuals who may have need for the message.
The dominant challenge of your discipline involves the problem of looking
back in time to see what has been contributed by others that will be of
benefit to the individual of today. But you know that to help meet this
challenge tomorrow, the messages of today must be stored with care.
I like to think that the objective behind the design of these systems, behind
your discipline, is to increase the effectiveness of the individuals at
the terminals of your systems. The intellectual labors of the individual
who generates a document are made more effective by your seeing to it that
his contribution becomes a "visible" part of society's growing
structure of knowledge. The individuals at the user terminals of your systems
are made more effective by the ready availability of pertinent documents
from the past. Your documentation systems provide an important means for
cooperation among many individuals who are distributed through space and
time, and the true measure of the value of your systems is the degree to
which their presence serves to increase the effectiveness of the individuals
at the terminals.
I wish also to comment upon the effects within your discipline that followed
from the development of an automatic-information-handling technology. There
appear to be mixed feelings among you as to the degree to which your future
systems will be affected by this new technology. But no one denies that
this technology does offer radical new possibilities throughout your systems.
It seems to me that these possibilities at the very least require a complete
re-examination of the entire bases of your systems, and that virtually none
of the previously established techniques or procedures is invulnerable to
change of a relatively radical nature. Without reevaluation, what has evolved
in a human-limited information-handling environment cannot safely be assumed
as necessary in the new machine-augmented environment.
About Individuals
Let us consider why individuals might justify special concern on the part
of information-handling-system researchers. For this purpose, there is something
which can be gleaned directly from your personal experience. Most of you
are oriented toward solution of a documentation problem and, therefore,
are yourselves problem-oriented people. Pose this question to yourselves,
then: "Who makes the actual effort that results in each step of progress
in your field?" In the final analysis, it is always an individual who
marshals the arguments, generates the hypotheses, provides the drive, etc.,
upon which each forward step is dependent. It is the problem-oriented individual
who is the basic module in our intellectual community. It is his effectiveness
which must be the ultimate concern of anyone wishing to see that community
perform better. Many disciplines, yours among them, are contributing toward
the increase of this effect-
1. Paper presented at the Annual Meeting of the American
Documentation Institute, Berkeley, California, October 23-27, 1960.
121
tiveness, but it seems to me to be quite important to give a little more
direct attention to the problem-oriented individual.
To consider what has been done for the individual's effectiveness by past
efforts of our learned men, I might point out that the definite and formal
techniques and procedures provided us by social heritage mostly involve
specialized and idealized aspects of the workload and needs of the individual.
There apparently never has been an over-all or "system" approach
to the problem of assisting the individual in being effective in his over-all
problem-solving role. But even if there had been a previous unified discipline
concerned with the over-all effectiveness of the individual in his real-life
problem-solving environment, that environment today is changing rapidly
enough to justify a thorough re-examination of the whole structure of that
discipline. Let us turn our attention to the question of what limits the
effectiveness of an individual in his professional environment. If we were
to give as serious consideration to redesigning his personal system of techniques,
procedures, and artifacts as you information specialists are giving to your
larger systems, there would be the same type of just)justification for re-examining
the entire present system. For example, cannot automatic information-handling
technology provide just as drastic an alteration in basic possibilities
within an individual's private operating system as it seems to do within
your documentation systems?
I prefer not to discuss here the general problem of "doing something
for the individual from an over-all point of view." My intention is
to tell you that, among the several areas of the individual's needs which
are not getting professional research attention, there is one particular
area of need for which you people would seem to have the right sort of talents
to make some contributions. I should hope that some of you might be persuaded
to turn your attention to this particular problem area. If none of you choose
to do so, my discussing the problem area with you still has possible value
because, if there are significant developments made in this problem area
in the coming years, the consequences are sure to affect the needs and possibilities
of your documentation systems.
As a model for us to discuss, let us take an individual that you might consider
to be typical of the users of your documentation systems. The area of his
need to which I refer above might be called "documentation for the
individual." I do not limit this area to the type of documentation
for the individual that is associated merely with his keeping a bibliographic
reference file such that he can track down any article he ever read. I speak
of a more general possibility. The individual usually works with much smaller
packets of information than is represented by the average paper or document.
His real working stuff is composed of unit facts, concepts, considerations,
etc., that he gleans from many sources. An important source, of course,
is the "cannibalizing" of the documents which your systems might
provide him, but a good deal of this "raw material" is obtained
by personal observation, reflection, synthesis, etc., as well as direct
communication with colleagues. Our individual works in a closed domain,
within which a great many such small packets of raw material fly about,
being extracted, originated, compiled, transformed, integrated, compared,
lost, forgotten, thrown away, etc. As a result of all of this activity upon
his raw material, our individual occasionally constructs messages which
he sends to the world outside his personal, closed domain, and these messages
to his colleagues and to the world at large represent essentially his entire
professional contribution.
One important role played by documentation (on your scale) is to provide
the means whereby, in the over-all scene, a perspective can be obtained
of the society, of its history, and its intellectual structure. To me this
is a basic utility--when one gains the necessary perspective, he then can
see where to add his contribution. But in the end, such perspective must
be gained by an individual, and supplying him with all the necessary documents
bearing all of the necessary information does not in itself satisfy the
need for perspective. The particular perspective needed by a contributor
is almost always unique, one that he must construct for himself from available
or yet-to-be-found facts and concepts. He gains the picture only by laboriously
fitting together new possibilities, generating or ferreting out new facts
or concepts, and repeating this process until a view emerges that is compatible
with the existing structure and with his desires.
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, preferably a medium that can provide him
with spatial patterns to associate with the ordering, e.g., an ordered list
of possible courses of action. Beyond a certain number and complexity of
interrelationships, he cannot depend upon spatial-pattern help alone and
seeks other more abstract associations and linkages. A way to store, retrieve,
and manipulate the information within our individual's private domain, with
informationpacket sizes that match his actual needs (i.e., separate concepts,
facts, considerations, etc.), could go far toward increasing the effectiveness
of his mental capabilities to the level needed for the extended and complex
problems that are the pressing ones of our day.
A system of the nature I am proposing would have many of the characteristics
of the large systems that you deal with. The time constants from storage
to retrieval would be shorter, the package size for information would be
smaller, the number of different people involved as generators, retrievers,
and users would be much smaller, the subject coverage would be smaller,
and the total amount of information would be smaller--but the relative variations
in time constants, nature of information packets, type of need to satisfy,
etc., are still large enough to require a respectable amount of sophistication
in developing a good solution to this problem. Because of
122
the uniformly smaller scale in the different aspects of this personal
system, and because of the current fashionable appeal in my home territory
of the trend known as microelectronics, I call this personal documentation
problem area "microdocumentation. "
An Example of a Simple Microdocumentation System
I have been experimenting in an extracurricular sort of way for the past
six years with a crude bibliographic and note-filing system which can qualify
as a microdocumentation system, and which can serve as an example of the
sort of thing I have been talking about. I took time out of an intensive
research project to settle once and for all the pesky problem of a good
personal bibliographic retrieval system. I had never heard of Calvin Mooers,
but with some significant hints from the writings of those who obviously
had, I managed to re-invent what was essentially a Zatocoding system. This
gave me plenty of power for one aspect of my research--keeping track of
the documents which became involved in my study--but the very first day
back on the initial research job, analyzing some of these documents and
piling up notes, made me realize that a bibliographic system was only a
partial solution.
Within an hour, a note-taking system was born and put into use that can
coordinate nicely with almost any bibliographic system; in fact this note
system has proven considerably more useful to me than has my bibliographic
system. I write my notes on edge-punched cards (essentially the size of
IBM cards, with about 100 peripheral holes). These cards include all the
different concepts, facts, considerations, personal ideas, etc., that I
come upon or generate during the working day and that seem at all useful.
I put only as much on a single card as seems to comprise a separate usable
information item or concept--for instance, it frequently happens that a
single paragraph in a reference article will provide facts, suggestions,
concepts, or stimuli for personal ideas that obviously relate to different
aspects of expected future work. Such a paragraph would result in several
notecards, each of which might contain only a sentence or two. I use cards
very freely, because the storage and retrieval is such that superfluous
notes do not really get in the way very much.
My bibliographic system gives a serial accession number to each reference
source from which I take notes. When I pick up a clean card to take a new
note, I write at the top the reference number and the page number, and then
begin writing the note. When I am done I underline a few significant words,
and sometimes write down a few relevant descriptors (not really necessary)
so that when I go to this card later its basic subject content will be immediately
visible. One field of the card is used to notch in the coded serial number
(or at least the two least significant digits). but most of the holes are
reserved for direct-code notching of descriptors. Cards are coded, stored,
and searched in particular notedecks, and each notedeck has its own set
of descriptors. It has always been easy to divide my fields of concentrated
interest into small enough areas that it is easy to get good retrieval specificity
from a small enough number of descriptors such that one hole could be assigned
to each for rapid encoding and searching. The descriptors are written on
a master code card; when 1 have completed writing a given notecard I place
the master card over it and very rapidly scan the descriptors and mark the
holes on the notecard that should be notched. Beyond writing the note, the
total extra time involved in choosing and marking descriptors, in marking
reference-source accession-number code, and in the later notching of a given
notecard is well under a minute. I feel that I have always gotten very large
returns from the investment of these extra minutes.
With this system I do not have to keep my notecards in order. Also, I can
very rapidly extract all of the notes from a notedeck that are relevant
to any of a large number of different aspects of my work. Useless notes
mostly stay in the deck, but can easily be weeded out if they become any
kind of a nuisance. There is now a place for me to store the various miscellaneous
ideas, possibilities, suggestions, facts, etc., that are always popping
up in the day's activity. When I am immersed in one phase of my work and
something stimulates a useful thought about another phase, I can quickly
write it down, code it, and return to what I was doing, confident that that
recent thought is integrated effectively into my activity.
I find it very productive to browse through a notedeck, integrating what
I find and what is stimulated in my mind, and synthesizing new concepts,
considerations, and ideas on new notecards. I make a practice of putting
the date of writing on each notecard, so that the chronological ordering
of the fragments of a conceptual structure that is in any stage of evolution
is readily apparent.
When searching for note items, a card that falls out on descriptor selection
and that proves particularly interesting gives me a reference source number.
I can needle-select all other notes in the deck from this same reference
to retrieve further relevant information. A good proportion of the need
to retrieve original reference documents that have once been "digested"
(i.e., that have any notecards in one or more notedecks) is met very easily
by a little needlework on one or more appropriate notedecks, and following
up on a few accession numbers from the selected cards. This reduces considerably
the demands upon my bibliographic filing system. When writing a paper and
using direct information from some notecard, the footnote acknowledgment
is captured merely by writing down the reference accession number, and letting
a secretary fill out the full referencing later.
There are many things about this system that could be improved, but it offers
considerable value to me as it is, and my purpose in describing it here
has been
123
to provide a crude but real-life example of the sort of thing I mean
when I speak of need and possibilities for microdocumentation. To continue
this "backyard" experimentation on a little more effective level,
I am bringing in a second party--my microdocumentalist if you wish--to help
me develop and operate an improved system. It is clear to me, however, that
development of really good microdocumentation systems will take much more
research than this effort represents.
About the Future
A microdocumentation system involving edge-notched cards, wherein the
user does his own clerical work, has already proven worthwhile in my own
case, and it may very well prove quite beneficial to a large class of individuals.
I feel, however, that clerical help with the routine parts of the system
operation will prove worthwhile and will expand the possibilities for designing
a truly useful system. Further, the development of some simple special equipment
to facilitate transcription, coding, and selection of the cards could produce
even greater system improvements. I feel sure that in today's economy, with
today's capability in equipment design and production, we can devise economically
justifiable microdocumentation systems that would considerably improve the
ability of the individual to collect and integrate the information necessary
for gaining perspective and finding solutions in complex problem areas.
I feel even more sure that the future will see some very radical changes
in the techniques and procedures with which our professional individuals
pursue their daily work. It is not hard to see a number of ways in which
automatic information handling equipment can be utilized in a microdocumentation
system. Just as in "macrodocumentation," the nature of the possibilities
offered by this equipment is such as to justify a complete re-examination
of essentially all of the techniques and procedures which we have inherited
and which we consider "the natural way to go about our intellectual
activity." Other areas of the individual's intellectual activity, besides
his microdocumentation system, are going to experience a similar and coordinated
kind of radical change. For instance, the detailed manipulations of symbolic
information which accompany our more sophisticated thought processes, and
which are now done manually, will be receiving attention similar to that
given to the area of microdocumentation. Any dynamic man-computer relationships
developed for aiding these sophisticated thought processes will involve
techniques and procedures which will be profitably coordinated with those
of the microdocumentation system.
That our technology does not now offer us directly the means for implementing
advanced microdocumentation systems is not entirely because it would be
beyond present capability. Generally speaking, developmental efforts have
been applied toward other special applications, but the basic technology
could provide us with much that would be economically beneficial to such
as microdocumentation systems, if this sort of application became of interest
to product development people. The real bind in early development of some
truly exciting systems lies in the research needed in the organizational
and procedural aspects of the problem.
You might give thought to the way in which an individual, whose basic capabilities
have been "augmented" by a good microdocumentation system, would
have his needs for service from your systems affected. Perhaps more important
even, would be the new capability this individual would exhibit in ferreting
out needed information from any of your systems.
I have often made the unsubstantiated but serious challenge that the money
spent on the big retrieval and scientific communication problem would bring
greater returns to the same end by being spent on developing really good
systems for the individual. I think I could offer reasonable defense of
this position even if it were assumed that microdocumentation were to be
the only "individual" system involved--but my defenses would be
considerably stronger if we included other areas of aid to the individual.
The individual's systems present a much more manageable problem. I claim
that we would be able to solve the big problem much more easily and cheaply
if we took care of the "basic modules" first (individuals), and
that also we would get the huge bonus of providing significant improvements
in the effectiveness of the individuals in their direct working environment.
There are several good papers to which I would like to call your attention,
which present various associated and supporting views regarding special
aid for the individual. The most directly relevant is one that many of you
may already be aware of, a paper by Vannevar Bush (1),
whose arguments of fifteen years ago could hardly be bettered for today's
case. I. J. Good (2)
and J. C. R. Licklider (3)
present relevant and very stimulating thoughts, also.
Conclusions
Man-machine cooperation appears to be a big factor in the possibilities
for making over-all gains in the effectiveness of individuals. There are
trends already developing toward new means of communication such that a
human can better transmit signals to a computer; research on display devices
is active; and the programming researchers are making rapid gains in developing
good languages or signal codes so that an individual can easily give directions
to the computer to do his bidding. There is also a significant trend toward
computer and programming systems that allow time-sharing of large and powerful
computers on a demand basis by many users, which promises to reduce the
cost of sophisticated service to an individual toward an economically realizable
level.
124
Your field of "macrodocumentation" should not lose sight of
the pressures, from other computer applications, toward developments of
closer man-computer cooperation. For instance, it will not be up to you
alone to decide whether or not it will be stylish or practical to have the
individuals who use your systems learn special communication techniques,
special languages, and special procedures for working with machines. These
things are very likely to be brought into the picture independently. Remember
that most individuals spend a lot more time suffering because of the problems
in their own little domains than they do being aware of problems in the
"macrodocumentation" service. Giving consideration to this makes
it seem inevitable, for instance, that your documentation systems will go
in the direction of using cooperative man-computer efforts, and that we
will see the evolution of special formal languages to be used operationally
in your systems.
Admittedly, these thoughts relate to rather distant times, and consideration
of users who have been provided with streamlined personal information-handling
techniques is not necessary in the system problems that concern you today;
however, there may be some long-range worth in bringing up such considerations
at this stage. There are shorter-range possibilities stemming from this
type of argument that could affect the course of some of the systems under
development today --the individuals who are part of the operating structure
of your systems, specialists (e.g., indexers, search experts) whose activity
is more limited than the general problem solver, and for whom special training
and special equipment may earlier be practical, could very well be given
some of the same individualized types of techniques, procedures, and equipment
service which I predict our more independently operating problemsolving
individual will later be using.
This makes three ways in which my vision of future development may affect
you: you might expand the boundaries of your discipline to help develop
good general microdocumentation systems; you might develop specialized individual
systems of this nature to make better use of the human capabilities within
your own systems; and you might expect the needs and possibilities for your
future systems to be affected by the "augmentation" of the individuals
who are to be the system users. I have found in your literature of late
several references to the possibility of a new professional specialist emerging--an
"information scientist" who can join a research group and take
care of the search and analysis of pertinent literature. It might be that
this type of an individual could also be trained as a microdocumentation
specialist, too, so that he can do his own work better and so that he can
help his colleagues set up and operate their own personal systems.
Perhaps next time we shall meet at a conference of the American Microdocumentation
Institute?
References
1. BUSH, VANNEVAR. 1945. As we may think. Atlantic
Monthly, (July). Pp. 101-108.
2. GOOD, I. J. 1958. How much science can you have
at your fingertips? IBM Journal of Research and Development. 2: 282-288,
(October 4).
3. LICKLIDER, J. C. R. 1960. Man-computer symbiosis.
IRE Transactions on Human Factors in Electronics. HFE-1: 4-11, (March
1).
125
|
|
