Thomas L. McKendree
This paper was included in Systems Engineering in the Workplace, the Proceedings of the Third Annual International
Symposium of the National Council on Systems Engineering [Now the International Council on Systems Engineering], held in 1993, and is placed on the web by the Molecular Manufacturing Shortcut Group for educational and individual use only. If "10-7" looks the same as "10-7", then your browser does not support superscripts, and some of the following numbers may be confusing.
Copyright 1993. All rights reserved.
This paper introduces the concept of metatechnologies--technologies which have
substantial effects on and through other technologies. It then discusses
molecular nanotechnology, a major impending metatechnology that systems
engineers need to understand. What they need to know, including a survey of
minimum performance characteristics, is reviewed. Finally, several educational
strategies which systems engineers might use to educate themselves about
molecular nanotechnology are discussed.
Table of Contents
A metatechnology is a technology which not only provides direct capabilities,
but one that also affects other technologies to dramatically improve system
performance. Digital electronics is the most obvious example of a
Molecular nanotechnology is an impending metatechnology which should launch
equally vast transformations early in the next century. It thus is a subject
that systems engineers should begin studying.
All technologies are in essence tricks that work to get things done in the real
world. For example, a transistor can screen out the carrier wave from
amplitude modulated to recover the original signal.
A metatechnology is also a trick, such as arranging feedback on transistors so
they latch at high or low voltages, depending on the pattern of input signals,
but a trick that is so flexible and powerful that it can be applied to a
tremendous variety of products, substantially changing whole categories of
systems. Digital electronics is the most evident example of a metatechnology,
due to the tremendous changes it is currently causing. These changes have been
continual because roughly every seven years computers become ten times more
cost effective (figure 1), repeatedly changing what are the most cost effective
arrangements of organizations, products and systems. Other meta-technologies
have also been developed, however, such as writing, the steam engine, and the
Writing was the basis for books. The steam engine was the basis for
self-powered trains. Digital electronics was the basis for inexpensive
computers and video games.
Writing supported standard written laws, and long-term, written contracts. The
steam engine supported ships that needed no wind. Digital electronics supports
inexpensive computers, video games, digital sound like compact discs, and
digital communications like ISDN.
Writing transforms product support with user manuals and maintenance guides.
The steam engine transformed production with powered machinery and transformed
distribution with powered transportation. Digital electronics transforms the
life- cycle of products in many ways. Computer Aided Manufacturing provides
more flexible and responsive production with shorter cycle-times. Computer
Aided Logistics Support (CALS) provides integrated product information between
suppliers and the government. Built in Test (BIT) supports easier maintenance
with fault detection and isolation.
1. Continuous tremendous growth in computing power has created a
metatechnology. Historical data from (Moravec, 1988).
The manufacturing process is a key part of the product life cycle, so any
technology that dramatically improves manufacturing is a metatechnology. The
assembly line is such a process metatechnology. While not itself incorporated
within products, it dramatically lowers the costs of producing products.
Even products which are not built by assembly lines, such as site-built homes,
are affected by this metatechnology. Tools used are much less expensive, and
thus much more available, because of assembly line production. Parts are
cheaper (and thus a wider array of much better parts are broadly affordable)
because they are made by assembly line. Such significant secondary and
tertiary effects are the mark of a metatechnology.
Writing allowed scientific data to build up. The massive production of
steam-driven factories required product/process integration. Digital
electronics transformed systems engineering with new tools like automated
requirements traceability, functional simulations, CAD, and linking CAD to
Molecular Nanotechnology: Thorough, inexpensive control of the structure of
matter based on molecule-by-molecule control of the products and byproducts;
the products and processes of molecular manufacturing.
--(Drexler et. al., 1991a)
In other words, molecular nanotechnology is the general ability to design and
build down to individual atoms. Its eventual development has already been
foreseen (Feynman, 1961).
Molecular nanotechnology is a distinct concept from most current use of the
word "nanotechnology." The term usually refers to existing and near-term
capaiblities to etch mechanisms, such as transistors and gears, to sub-micron
accuracies. A 0.5 u wide wire is also a 500 nm wide wire; it is flashier to
call this "nanotechnology." While of vast current utility, it does not in any
way represent the ability to place individual atoms exactly where desired, or
to make and break specified atomic bonds. Molecular nanotechnology is more
like scaling up chemistry to larger products than scaling down
The essense of molecular nanotechnology is designing and building artifacts to
molecular specification. The net result is that molecules can be designed to
be components. A Buckytube could be a cable or pipe. An atom can be a gear
Components become useful once they are combined into systems. While a number
of molecular nanotechnology products have been suggested (Drexler et. al.,
1991a), and it has been shown that an incredibly wide variety of systems should
be feasible (Drexler, 1992a), only two system designs have thus far been
presented in detail in the literature. The two system designs are for a
mechanical molecular computer and for a molecular manufacturing system that
both consists of, and can build systems of, molecular components.
A key element of molecular nanotechnology must be such molecular manufacturing.
This is the ordinary means by which objects specified to molecular precision
would be made.
The argument that molecular nanotechnology is something that comes once one has
molecular nanotechnology is circular, until one explains how a first molecular
manufacturing system can be built. One approach is to use ribosomes reading
DNA to fabricate proteins which self-assemble into the first molecular
manufacturing system. Designing the necessary proteins, and determining the
DNA sequence that yields them, are very hard design problems. This approach
has the advantage that the required manufacturing capability is available
today. Other approaches are easier in design, such as those which use scanning
molecular manipulators, but require developing new intermediate manufacturing
The systems engineering profession needs to address this subject. Molecular
nanotechnology offers the prospect of performance significantly better than
current capabilities. This will dramatically affect what systems we can design
in the future, and how systems we are now designing will fare in that future.
Furthermore, the world is developing tremendous demand for some understanding
of molecular nanotech-nology, and we systems engineers are particularly
well-suited to acquire that understanding. The subject is technical, and very
interdisciplinary; we are trained to address technical issues across many
disciplines. The technology will require specifying and developing systems of
unprecedented complexity and part counts; dealing with complexity is the focus
of our job. The key policy issues are tangled by non-linear systems of
powerful potential feed-back loops; we are among the worlds best at
understanding such complex systems. Both duty and future profits call. To
answer that call effectively, however, we first have to understand the
A Barrier to Understanding
Virtually every claim this section will make about molecular
nanotechnology appears to be preposterous and wrong. A crucial reason for
studying this field is that it violates, badly, most people's intuition about
feasible technology. The claims are both significant, and hard to evaluate.
All the numbers cited, however, are the results of careful calculations using
Fundamentally, what systems engineers need to know is that molecular
nanotechnology, by designing & building to atomic precision, will have as a
minimum the following surprising capabilities:
(Drexler, 1992a) demonstrates that it will be possible to position reactive
molecules in a programmable way with ~0.1 nm precision, allowing
mechanosynthesis. Such mechanosynthesis could run at >106
operations/(device* second). Mechanosynthetic assembly of 1 kg objects will be
possible in <104 seconds.
The implication is that it will be possible to build extremely rapidly and
accurately. This dramatically reduces the manufacturing portion of cycle time
(period from first order to delivery of a product). Building a 1 kg object
that quickly with mechanosynthesis will require a system containing
~1017 assembly devices and the necessary support structures.
The primary costs in manufacturing with molecular nanotechnology are the costs
of raw materials (which can be unpurified bulk chemicals) and energy (which can
be electricity off a power grid, or the chemical energy of a feed stock).
Thus, the best cost estimates for manufacturing are at most only a few dollars
per kg (Drexler et. al, 1991). This cost estimate is widely applicable across
many products, including those which require molecular nanotechnology to
To a first order, manufacturing cost depends on the amount of mechanosynthesis
performed, and does not increase when increasing the number of internal
interfaces for a mass. Thus, manufacturing costs will vary almost directly
with mass, and will be very insensitive to part counts.
$5 per kg is the high end of a very uncertain range. Once molecular
nanotechnology is applied to reducing the costs of molecular nanotechnology, by
providing durable and very low cost solar cells, for example, costs might fall
as low as a few cents per kg of product (Drexler et. al, 1991).
Note, this does not guarantee that total product costs will be low. Other
costs, such as taxes, liability, compliance with regulations, advertising,
general and administrative overhead, etc., could easily swamp out manufacturing
costs, if allowed. Design also has the potential be a large cost, but there
are reasons to suspect that the computers which molecular nanotechnology makes
possible will be able to automate a great deal of the design process.
The manufacturing capabilities of molecular nanotechnology are enough to
declare it a metatechnology. The dramatic decrease in manufacturing costs
alone will be more significant, in percentage terms, than the invention of the
assembly line and interchangeable parts. This manufacturing capability will be
applicable to high-cost low-mass items, such as computer circuitry, advanced
composite structures, pharmaceuticals, and fully integrated aerospace systems
ready for use.
Manufacturing does not represent the limit of molecular manufacturing as a
metatechnology, however, since molecular manufacturing will also include
products with the following characteristics:
The following numbers are based on a design in (Drexler, 1992) for a mechanical
computer using molecular components. The logic gates are ~10-8
u3 (~10-26m3). They switch in ~0.1 ns,
and dissipate <10-21 J. The design was selected to be amenable
to analysis. It is nearly certain that better designs are feasible, even
though the performance of such designs has not been calculated. In particular,
electronic computers designed and built to molecular precision should be
Computational power should be available at 1010 MIPS per Watt, (MIPS
means "Millions of Instructions Per Second," a computer operating rate). The
availability of power and cooling would likely be the limit to most practical
computation. (Drexler, 1992a) illustrates cooling a ~105 Watt,
cubic cm system, at an equilibrium temperature of 300 K, and thus the
capability to provide compact, 1015 MIPS parallel computing systems.
A very conservative investment cost for purchasing computers should be
$2x10-17 per MIP. This estimate comes from applying the $5/kg cost
to a cm3 of diamond, and calling the total 1015 MIPS. A
1000 MIPS unitary CPU should take up less than 0.1 u3 (Drexler,
1991b), and would thus cost less than $2x10-21 per MIP.
Since a 1000 MIPS CPU should fit within a cube less than half a micron on a
side, and need less than ~10-7 Watts, such computers could be
accommodated throughout a wide variety of objects. Their low cost should make
this affordable. For examples, (Drexler et. al., 1991a) suggested embedding
intelligence within structural materials, paint, clothing & furniture.
For macroscopic products, the low cost of computers suggests using prodigious
numbers. Furthermore, many products (e.g., intelligent paint) require
distributed computers. Thus, macroscopic products of molecular nanotechnology
would likely us massively parallel, massively distributed computing
For significant computer systems (roughly greater than ~104 MIPS),
the mechanical computer design begins to require parallelism to accommodate the
communication delays. Electronic computers built though molecular
nanotechnology are not well characterized, but probably would allow faster
computation before requiring parallel CPUs.
Small autonomous objects, such as cell repair devices (Drexler, 1986), might
well have unitary computers.
Mechanochemical power conversion should be feasible at >109
W/m3. Electromechanical power conversion should be feasible at
>1015 W/m3. Power conversion efficiencies could be
> 99%. (Drexler, 1991a). Cooling difficulties prohibit such power
conversion densities for anything as large a m3.
Molecular nanotechnology can include large structural elements made of
engineered, fibrous diamondoid material. Such structures should have >5 x
1010 N/M2 tensile strength, and a larger compressive
strength. The result is a strength-to-density ratio 80 times better than
Structural material could be suffused with actuators on nanoscale flexible
joints. The material would lose strength, although it should still be possible
to make structural components the strength of steel. The result would be
powerful, flexible materials. This would allow, for example, variable geometry
laminar flow wings.
Molecular nanotechnology should provide affordable and robust systems for
sorting out of the environment previously dumped waste. It should also be able
to easily neutralize toxic chemicals by reordering their molecular
Molecular nanotechnology does not offer any direct means of eliminating the
radioactivity of radioactive waste, although (Drexler, 1991a) suggests at least
one approach for addressing radioactive waste that molecular nanotechnology
could make cost-effective.
Perhaps as significant, molecular nanotechnology is inherently much less
disruptive to the environment than making equivalent products with
technologies. It is very resource efficient. It offers high performance
products at low mass, thus reducing further needed resources. By controlling
the molecular structure of its outputs it need not release chemical toxins.
Finally, molecular nanotechnology could be powered entirely through solar
Both computers and molecular nanotechnology involve huge numbers of tiny parts
operating together ver rapidly as a complex system. Coordination and control
are key issues in each field. These similarities suggest systems engineering
will be similar between the two fields.
There are two important differences, however, between systems engineering for
current computers and systems engineering for future molecular nanotech-nology.
First, many molecular nanotechnology components will provide actuation
functions very different from information processing. Second, future molecular
nanotechnology systems will have as many as ten-billion times more parts than
current computers have transistors.
Estimating when a technology will arrive is difficult, since the question is
not subject to well-defined formulas. Nevertheless, one must try.
(Drexler, 1992b) has a guess:
Analysis and simulation based on existing scientific knowledge is enough to
show what molecular nanotechnology can do, but developing it will require the
construction of better molecular tools. The pace of development will depend
not on unpredictable breakthroughs, but on the magnitude and quality of a
focused development effort. The total development time is hard to predict, but
15 years would not be surprising.
For a sense of the uncertainty surrounding Dr. Drexler's estimate, examing
figure 2. The curve is based on a Markov model of transitions through required
intermediate technologies. It is more precise than is actually known.
Nevertheless, the curve's sweep appears reasonably accurate. There seems
virtually no chance in reality of developing molecular nanotechnology before
2000. It is difficult to estimate when it will be developed, but it is hard to
construct plausible scenarios which delays it much beyond 2020.
2. A probability distribution for when molecular nanotechnology may become
While current estimates of when molecular manufacturing will be developed must
carry significant uncertainties, it is clear that there is a very real and
large probability of being developed within the lifetimes of products now. As
new systems begin entering development, explicit consideration should be made
of how the potential for molecular nanotech-nology might affect the life of
Molecular nanotechnology offers many capabilities, many of which do not
correspond to current trends. It thus may violate many implicit assumptions.
3. Molecular Nanotechnology has potential for changing long-term trends in
computing power. Historical data from (Moravec, 1988).
As one example, remember that computers have been growing rapidly in capability
for decades. Figure 3 shows how the estimated performance of molecular
nanotechnology should allow better computers at tremendously lower prices,
decades sooner than would occur under the rate of growth the world has become
accustomed to. This alone could greatly change many long-term plans.
There are several ways systems engineers can learn about molecular
nanotechnology. The subject is of great significance, and will grow in
importance as the technology comes closer to realization. Molecular
nanotechnology deserves growing study. Therefore, it makes sense to start
A prospective student must first recognize the two big difficulties in studying
molecular nanotechnology. The first is that it is highly interdisciplinary.
Of all people, however, system engineers are particularly skilled at
integrating multiple technologies into a single perspective, and thus the
interdisciplinary nature of molecular nanotechnology should be less of a
problem for them. For systems engineers, initial exploration can initially
ignore the guts of why molecular nanotechnology will work, and instead look at
the implications on systems of already calculated performance parameters. This
approach makes starting easier.
The second biggest difficulty in learning about the subject is accepting
current understanding as potentially correct. Those first examining the field
need to suspend disbelief long enough to openly consider the possibility of
molecular nanotechnology. The current best understanding upsets most people's
assumptions. On the other hand, if reality is going to upset those assumptions
anyway, it is better to face the facts early, and adapt while there is still
Once one is really open to the possibility, and seriously consider molecular
nanotechnology, it is time to be critical, and make sure that the possibilities
stand up to evidence and criticism.
This is the classic strategy when learning about a new field. A problem is
that there is not much literature yet on molecular nanotechnology, and what
exists generally is not oriented towards a systems engineering audience.
[Note: since publication of this paper, http://nano.xerox.com/nano has emerged as an excellent place to look for molecular nanotechnology literature on the web.]
The canonical author in the field is Dr. Eric Drexler. Since the subject is
very interdisciplinary, encompassing many parts, reading a book most helps
keeping parts in context, and holding the key ideas together.
(Drexler, 1992a) supports the numbers surveyed in the previous section. This
is the best source on the detailed guts of the technical arguments for why
molecular nanotechnology will work. This book is the choice for people who
want to see equations.
(Drexler et. al., 1991a), through the use of scenarios, is best current source
for a sense of how molecular nanotechnology could affect people's everyday
life. No one really knows, but this surveys some of the best guesses. The
book includes a very good discussion of eventual costs. It is scheduled for
rerelease as a paperback in the summer of 1993.
(Drexler, 1986) was the first major publication on molecular nanotechnology.
Many ideas first expressed in this book have since been more precisely worked
There have been two Foresight Conferences on Nanotechnology. Proceedings are
available for each. (Crandall and Lewis, 1992) is the proceedings of the first
conference, and (Institute of Physics, 1992) is the proceedings of the second
conference. Any bookstore should be able to order the former. According to
(Foresight, 1993), the latter is available by ordering Volume 2, numbers 3
& 4 (in the bound-together book, if still available), of the journal
Nanotechnology, from IOP Publishing Ltd, Customer Service, Techno House,
Redcliffe Way, Bristol BS1 6NX, England. The price is $134, or 69 Pounds,
payable by visa, check, or bank transfer.
Not many papers for systems engineers. A great deal of work on the path
towards developing molecular manufacturing in basic research journals, such as
Science. Key topics to look for include molecular modeling, scanning
probe manipulation, developing better models of proteins and how they fold, and
Also, the Journal of the British Interplanetary Society, October 1992
issue is devoted to molecular nanotechnology & Space.
Taking a course is nearly an ideal strategy for learning a new subject.
Unfortunately, only one course has been offered on the subject, at Stanford, in
1988. It would be very reasonable to launch a new courses, however, using
(Drexler, 1992a) as the basic text.
At a university, such a course should be lower graduate-level and might be
hosted by the chemistry department. It should be very interdisciplinary,
taking students, and perhaps faculty, from chemistry, condensed-matter physics,
molecular biology, materials science, mechanical engineering, computer science,
industrial engineering, and systems engineering.
Such a course could also work at a community college, acting as a venue where
non-academics could come together to jointly study (Drexler, 1992a) and
molecular nanotechnology. This requires being able to draw a sufficient number
of students to form the class.
This is a very viable learning strategy, because most people learn best by
doing. Designing systems is what many systems engineers do. Making the
attempt is sufficient to guide systems engineers to what is most important for
them to learn.
Molecular nanotechnology is still in the early concept exploration phase. Some
viable designs have been found, but they have been intended to be easily
designed examples. Attempts should now be made to look more for optimal
designs. Molecular nanotechnology will benefit because a growing body of
designs will allows the idea to be more precisely criticized and refined.
Furthermore, converging towards the optimal designs will provide an improving
estimate of what molecular nanotechnology will be able to accomplish.
Those who wish to use this approach, but who are uncertain of the substantial
performance numbers cited in (Drexler, 1992a), could apply substantial safety
factors to those numbers, and still would be likely to discover surprising and
interesting system designs.
Finally, those who follow this approach will help by providing needed papers
and articles on the subject.
This is another way systems engineers could learn by doing. In systems
engineering molecular manufacturing systems, one learns the key constraints on
what is feasible, understands how molecular nanotechnology could work, and
develops an understanding of the most important and difficult problems.
This also is an excellent approach to producing needed papers and articles on
the subject. The intended readership should be experts in the chemistry or
biotechnology fields that will eventually be creating a molecular manufacturing
Most of the work on molecular nanotechnology presupposes advanced molecular
manufacturing systems with nearly arbitrary ability to build to molecular
precision. Intermediate manufacturing systems, more advanced than current
capabilities, would be useful if developed. Chapter 16 of (Drexler, 1992a)
provides examples of such intermediate systems. Systems engineers might
develop alternate systems to trade against, or refine Drexler's designs.
An advantage of such research is that results would be for nearer-term systems.
Thus, it would have earlier applicability, and be less susceptible to
This would be a major challenge, trying to develop synthesize Which
intermediate architectures are viable. What is a path through viable
A fundamental purpose of professional organizations is to help members learn
the knowledge and skills critical to their work. Thus, NCOSE should begin
providing support for members interested in molecular nanotechnology. As a
minimum, this could be a review of (Drexler, 1992a) in the NCOSE newsletter.
More substantial would be the creation of a special interest group within
NCOSE. The group could studying how molecular nanotechnology will impact
systems engineering. Such a focus would be of interest to systems engineers.
An alternate focus would be of more use to the customers of systems
engineering. That is to begin the systems engineering of molecular
nanotechnology. That is a proper job of our profession, and would put systems
engineering in the lead of what seems destined to become the most important
emerging field of technology. Molecular nanotechnology is still a wide-open
field, and thus a golden opportunity for ourselves and our profession.
At the 4th annual NCOSE symposium, there should be more papers, and hopefully
even a full session, on the subject of molecular nanotechnology and systems
Scenario development is an excellent tool to honestly consider fresh
alternatives, particularly those that differ from one's current expectations.
Good scenarios are stories which allow people to temporarily suspend disbelief
(Schwartz, 1991.) While the following are only brief sketches, not fully
worked out scenarios, they are worth considering. For each, ask "if this were
going to happen, what should my organization do? What should the project I am
working on do? What should I do?"
(Title from Drexler et. al., 1991a) Everything goes similar to how you
probably always expected. A general ability to design and build to atomic
precision is never developed.
Many development problems might prove much harder than currently estimated.
Perhaps development becomes caught in the quagmire of a giant bureaucracy that
co-opts potential alternate developers. Perhaps last-minute technical or
regulatory hang-ups delay full up test of the first full-up developmental
system. In any case, it takesover two decades to develop molecular
For consistency, this scenario requires that little work succeeds in
manipulating molecules with scanning probe tips, molecular biology progresses
slowly and progress be slow in molecular modeling. As a result,
pharmaceuticals would be developed less quickly than is now expected,.
A very long delay probably requires that within decade the growth in computer
capability stop. Otherwise the "growing computer" scenario brings the world
unstoppably close to molecular manufacturing.
In this scenario, ever more powerful tools for designing and building to
molecular precision are developed, such as better molecular modeling, even more
chemistry techniques, superior protein CAD systems, hybrid
atomic-scanning-probe/biomolecule-tip devices, and even inventions that
haven't been thought of in 1993. Ways are found to make variants to some
products which were originally conceived of assuming they would need mature
molecular nanotechnology to produce. The world might spend years with these
intermediate capabilities, producing growing amounts of increasingly broader
products. This smoothes the transition by providing limited capabilities
similar to mature molecular, but sooner. When capabilities that strictly meet
the 1993 definition of "molecular nanotechnology" are finally built, perhaps in
12-20 years, there is a great deal of hype, but the world takes no single
massive jump. The development is just another increase in competitive
This variant of progressive development goes as follows: Growing computer power
yields ever more powerful computational chemistry tools. Eventually this
provides robust molecular and protein modeling able to solve the reverse
protein folding problem (given an engineering specification for a protein, find
a DNA sequence that codes for an amino acid sequence that will fold-up into a
protein that meets the specification). After a few years of easily making
designer proteins for increasingly divergent types of requirements, it becomes
obvious that a determined development effort could produce molecular
nanotechnology, and what capabilities that would offer.
Over the subsequent few of years, with massive funding of parallel paths, the
right DNA sequences are developed. Combining that DNA with tools available
even in 1990, limited molecular manufacturing systems built of self-assembling
proteins are constructed. These in turn produce the more easily designed,
robust molecular manufacturing systems, which do not require proteins. At this
point mature molecular nanotechnology is real.
In this scenario, most of the key development issues in molecular
nanotechnology prove quite solvable, once directly confronted. There will be
growing interest in the field, and being easy to do it is developed quickly
This scenario includes a potential for blind-siding that needs to be
considered. It is consistent with what is known about molecular nanotechnology
that if development proves easy, then development may only need dozens of
researchers using fairly standard equipment. Thus, molecular nanotechnology
might be created in commercial secrecy. What would happen if in a decade a
competitor surprised you with capabilities near the performance numbers cited
earlier for molecular nanotechnology? If you could not survive such a
blind-siding, how are you insuring that you won't be thus surprised?
Blind-siding is almost the only scenario that leads to molecular nanotechnology
without a period of massively funding before full development. This is not
because developing molecular nanotechnology ever needs billions of dollars. It
is because once an organization with money honestly believes that molecular
nanotechnology is imminent, and thinks it know how to develop the technology,
the situation makes a great deal of sense for the organization to trade money
for time, even when a large amount of money provides only a small
foreshortening of the schedule.
Metatechnologies are those technologies which exert tremendous leverage on many
systems. The next metatechnology appears to be molecular nanotech-nology. The
field is in early concept exploration, and system engineers belong in the
Before systems engineers can lead, we have to learn, so we understand where to
go and how to get there. The biggest challenge is suspending disbelief long
enough to fairly evaluate the possibilities, supplying the necessary critical
rigor in such a interdisciplinary field.
British Interplanetary Society,
Journal of the British Interplanetary
Society, Vol. 45, No. 10, October, 1992.
Crandall, B.C. and Lewis, James, ed., Nanotechnology: Research and
Perspectives. MIT Press, Cambridge, 1992.
Drexler, K. Eric, Nanosystems:
Molecular Machinery, Manufacturing, and
Computation. John Wiley & Sons, New York, 1992a.
Drexler, K. Eric, written testimony to the US Senate Committee on Commerce,
Science, and Transportation's Subcommittee on Science, Technology and Space, on
"New Technologies for a Sustainable World," U.S. Government Printing Office,
ISBN 0-16-039898-3, 1992b.
Drexler, K. Eric, Peterson, Chris and Pergamit, Gayle, Unbounding the
Future: The Nanotechnology Revolution. William Morrow & Company, New
Drexler, K. Eric, Personal Communications, 1991b.
Drexler, K. Eric, "Engineering Parameters for Materials and Subsystems of
Aerospace Interest Produced by a Mature Molecular Manufacturing Technology."
Paper prepared for the National Space Society Annual Conference, Houston,
Drexler, K. Eric, Engines of Creation. Anchor Press/ Doubleday, New
Feynman, R.P., "There's Plenty of Room at the Bottom." Miniaturization, H.D.
Gilbert, ed., Reinhold, New York, 1961.
Foresight Institute, "Two Foresight Nanotechnology Conference Proceedings."
Foresight Update, No. 15, pp. 10-11, 1993.
Institute of Physics, ed., "Toward Molecular Control: Second Foresight
Conference on Molecular Nanotechnology." Nanotechnology, Vol. 2, No. 3
& 4, IOP Publishing Ltd, Bristol, 1992.
Moravec, Hans, Mind Children. Harvard University Press, Cambridge
Schwartz, Peter, The Art of the Long View. Doubleday, New York, 1991.
Mr. McKendree is a Ph.D. candidate at USC, in the Industrial & Systems Engineering department. He is studying Systems Architecture under Professor Eberhardt Rechtin, and his dissertation subject addresses the systems architecture of a molecular manufacturing system nanotechnology.)
Since 1986, Mr. McKendree has been a full time employee of McDonnell-Douglas Aerospace. He earned a Masters in Systems from USC in 1989, and attended the International Space University in 1990. In 1986, he was awarded a bachelors degree in Aeronautics and Astronautics, and a second bachelors degree in Mathematics, from MIT.
[Since the publication of this paper, Mr. McKendree has become a full time employee of Hughes in Fullerton, California. He can be reached at firstname.lastname@example.org]