From the Proceedings of the 20th Annual Technical Symposium of the Washington, D.C., Chapter of ACM, Crisis in Computing: Innovation in a Constrained Environment (June 18, 1981: College Park, Maryland) vi-viii. Jim Pottmyer was program chair for this symposium.
© Copyright 1981 Association for Computing Machinery, Inc. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Association for Computing Machinery. To copy otherwise, or to republish, requires a fee and/or specific permission.
Return to: Jim Pottmyer's home page | Jim's other writings
Technical innovation has three parts. These are:
Accomplishing the first two -- concepts and illustration of utility -- results in an "invention." Invention is not synonymous with innovation. Innovation depends upon prior invention, but it also requires investment of resources to an economic end.
There is a popular notion that invention occurs as an inventor sits alone in thought, then suddenly says "eureka" as a light bulb appears in a comic-strip thought balloon overhead. Fame and fortune follow. The creative idea, alone, is not an invention.
Creative ideas are abundant. More scarce is insight for economic potential. And truly rare are the dedication and means to solve enough of the engineering problems to demonstrate the utility of the idea.
Once there is invention, innovation waits upon favorable economic circumstance for the investment to bring it into use.
Today's digital computer is not a result of a single "aha." The following terse chronology [1, 2] of ideas and inventions shows this. It is left as an exercise for the reader to trace the idea of the computer further back into ancient commentaries.
Use of the abacus spread. A mechanical contraption, a mere artifact, assisted the mind in one of the seven liberal arts.
Raymond
Lull (1235-1315 A.D.), in his Ars Magna et Ultima,
described a finite state machine (to be constructed of concentric,
adjustable circular sheets or rings). By proper manipulation of
the device, Lull believed it possible to draw every kind
of logical conclusion. Lull's idea is more than just
a curiosity since it seems to have influenced Leibniz over three
centuries later.
Gottfried Wilhelm Leibniz further developed ideas of a combinatorial method as a basis for all science. And he founded symbolic logic. (In particular, he used artificial symbols for logical constants as well as variables. The idea of a completely formalistic logic, however, was developed by Boole in the 19th century.)
Besides philosophical ideas, Leibniz demonstrated the utility of a computing device. His numeral wheel calculator attracted notice, and he had two prototypes constructed. Although rotary registers had been made earlier by Pascal and Schickard, the Leibniz stepped cylinder design was the one used in the first reliable rotary calculating machine constructed a mere century later.
Hahn constructed a reliable rotary calculating machine. Muller published an idea for a difference engine having automatic sequence control.
Once a dependable calculating machine existed, it still took investment to refine the design and begin manufacture. Thomas de Colmar set up the "arithmometer" industry, constructing his first machine in 1822. Throughout the rest of the 19th century, inventions were plentiful, leading to explosive innovation in calculating machines.
Charles Babbage's construction of difference engines and his plans for the "analytical engine" (assisted by Lady Lovelace) showed the practicality of automatic sequence control.
Shortly before World War II, Howard Aiken pulled together -- into a unified design promising economic utility -- the major components of the modern digital computer, e.g., arithmetic computation, storage of computed values, and sequence control based upon computed or stored values.
The War gave dispensation from ordinary delays in justifying investment (as far as ballistic calculations were concerned). And Aiken's Mark I completed the "innovation" of the digital computer.
To summarize, a millennium elapsed from the invention of the abacus to the Mark I. It took over six centuries to realize Lull's idea of a general logical machine. The Mark I followed Leibniz' idea of symbolic logic and his prototype of a numeral wheel calculator by two and a half centuries. And the Mark I came one century after the manufacture of calculators, and after Babbage built difference engines with automatic sequence control.
Innovation requires MORE THAN a single idea --more than one "light bulb" overhead!
Over the last thirty years, the U.S. computer industry exhibited exceptional innovation. As phenomenal improvements were made in the performance and price of computers, it proved easy to justify investing in replacements for computers only recently acquired. With improved hardware price performance leading the way, many opportunities existed to try out novel ideas.
Tax considerations have favored investment in computers and office machines, as well as in transportation and production equipment containing embedded computers. The investment tax credit favors investment in most equipment relative to investment in durable structures. And recently, rising interest rates have reduced the present value of tax depreciation write-offs more for structures than equipment. [3]
Given the success story of the computer industry, is there a crisis in computing regarding innovation?
Semiconductor chips are so cheap that, if free, computer systems' prices would be almost unchanged. Further improvement in chip technology is possible. But changes can better price performance only by reducing the number of components that have to be connected and by simplifying the engineering problems of their interconnection. There is no room to make components much cheaper.
Any dramatic future improvement in chip technology requires a consensus regarding the protocols to permit very large, capable modules to be designed with minimum parts for quantity production.
So long as the payroll or inventory program had to be reimplemented again and again on evolving computers, there were lots of opportunities to try out new ideas. A lot of research has been funded from operational budgets.
Now there is excess capacity for the "mundane" applications. And we have learned to do the payroll or inventory almost right (after the ten thousandth try). Increasingly, for new ideas to be prototyped, they must be justified on their own potential for future return.
But our tools for prototyping are found to be primitive.
Frequently, the only means to build prototype software cheaply are to eliminate specification, configuration management, and quality assurance during development. The worst thing that can then happen is that the prototype work! Users perceive a choice of using a tool today or waiting another two years for the tool to be redeveloped properly. The choice too often is to press a prototype into service and pay again and again through the years trying to maintain an unreliable system.
Many papers and presentations for this symposium, in several of the sessions, address languages and other tools needed for proper prototyping.
Studies of the problems encountered by people in using computers have typically been the ones requiring the most realistic (and thus most expensive) prototypes. A better descriptive theory of user-computer interfaces offers exit from the quagmire of having to build a "completed" system to find out if it works. Several presentations at this symposium address human factors.
Existing investment increasingly influences how much novelty is tolerable in new systems. Lately, announcements of new systems have failed to fulfill our dreams of bold new technical thrusts. Price performance is key. New systems are expected to offer migration paths so that no one will be forced into converting existing applications any more than necessary.
There is a large inventory of operating programs with which we are reluctant to fiddle.
This reluctance to "touch" working programs also affects attitudes towards standards. In these attitudes, government interests are frequently different from those of much of the private sector.
The government must, of course, be above suspicion of favoritism in procurement, and it has a special role in fostering competition. Hence, government computer users cannot usually take the easy route of sticking to the migration path offered by the vendor of currently installed equipment. Government acquisition of computers is a lengthy process, hindered by the lack of interconnection standards. The private sector has more opportunity to stick with one vendor through several system upgrades, resulting in less short-term priority to some types of standards.
Further, if an ambiguity was overlooked in an adopted standard, someone on the vendor's migration path does not want the ambiguity resolved in a way that will force him into conversion. Someone else, wishing maximum competition in the marketplace, desires the ambiguity resolved -- one way or another -- in the interest of long-term benefit even if short-term costs are higher.
Two panel sessions at this symposium specifically address standards issues -- from both government and nongovemment perspectives.
In the early 1920s, N.D. Kondratieff developed the idea of long cycles (40 to 60 years) in the world economy, from studying wholesale price indices, interest yields, wages, and other available statistics since the closing of the 18th century.
We have now come a full cycle since 1930.
The time is at hand for major restructuring of the economy and society. Industrial production capacity is utilized at only 80 per cent. Although the overall unemployment rate is about 7-1/2 per cent, skilled workers are scarce in some sectors of the economy.4 (Computer programmers, for example, are scarce.) Large parts of the country's "physical plant," constructed over the last expansionary portion of the long cycle (e.g., the suburbs, the air traffic system, the interstate highways), are far from worn out but use energy less efficiently than now seems prudent.
Investment patterns in the year 2000 will be quite different than they are today. But it isn't easy to retrain Detroit auto workers to be Sun Belt computer programmers. It is difficult to build new energy efficient equipment and structures when there is a lot of serviceable life in those inefficient ones we already own.
Usually, in periods of depression, a lot of bad debt is written off. Traditional ways to do this are through default or through inflation. However, most of the debt today cannot be written off through default as it could in the l930s. Mortgage debt on one- to four-family homes, largely owed by individuals, stood at $956G at the end of 1980, compared at that time with the national debt of $937G and consumer installment credit of $313G. [4] Defaults of any magnitude on home mortgages and consumer installment purchases would overwhelm the legal system. The national debt, home mortgage debt, and consumer installment credit account for more than half the debt outstanding in nonfinancial sectors.
So, if "bad debt" is to be written down, we are left with inflation. But inflation makes the raising of capital for innovation difficult.
Fortunately, the introduction of the computer gives us unprecedented current information about the economy. And we need no longer be completely at the mercy of "natural forces" of the long wave. But, although the idea of economic regulation oriented to a 50-year cycle is of great antiquity,[5] we are inexperienced in encouraging technological investment and research in the bad years of the cycle.
The 9-digit ZIP code is resisted, and there are outcries against dropping price marking of grocery store items that can be electronically scanned. This suggests that the public does not easily accept everything which we, as technicians, offer. Before there can be large-scale investment in wiring (or fibering) America with local communications loops to permit "work-at-home," "shop-at-home," and other promised innovations, attitudes will have to change. A human generation may be needed. And during this generation, we must do a better job of serving the public with our computers and respecting people's privacy.
The primacy of the United States in computer technology has recently come into question. One session at this symposium is specifically devoted to international trade questions. Don Sackman notes on page 11 of these proceedings: "The handwriting is on the wall -- and it is Kangi!"
Despite the problems noted here, the tone of this symposium is
upbeat. Innovation is not dying -- judging from the papers
and presentations given here. Through our symposium, we hope to
anticipate and avoid problems caused by the forces which are slowing
our explosive innovation.
Return to: Jim Pottmyer's home page | Jim's other writings