Foresight Update 3
A publication of the Foresight Institute
Nanotechnology concepts continue to be introduced to diverse
audiences. At the Artificial Life Workshop, held at Los Alamos,
NM, nanotechnology was compared to the "artificial
life" theme. Parallels and differences were pointed out
between artificial replicators and living things, with the
differences predominating. We hope these issues will be included
in the Workshop proceedings and will announce when they're
A four-part mini-symposium was given at MIT in January. It was
well-attended: some attendees stood for hours to hear the
introductory presentation. The group especially enjoyed the
extensive discussion of the possible social consequences of
Nanotechnology systems were by far the smallest systems discussed
at the IEEE Micro Robots and Teleoperators Workshop, a meeting
focusing on the smallest mechanical parts we can fabricate today.
The next largest parts discussed were in the multi-micron range.
IBM has requested several talks on nanotechnology: first at the
Almaden Research Center, with an audience including top STM
researchers; then at the Watson Research Center, which featured
nanotechnology in its Physical Sciences Colloquium and found its
auditorium overflowing from the chairs to the stairs. Soon IBM
Santa Teresa will have a presentation as part of its Advanced
Education Seminar Series.
Chemical engineers heard about nanotechnology at a recent
Molecular Thermodynamics Seminar at UC Berkeley. At Johnson
Controls in Milwaukee, sensor experts considered long-term
Computer-oriented audiences continue to show a high level of
interest: Apple Computer employees, upon hearing an introductory
talk, immediately generated some extreme (but physically
reasonable) applications which are too wild-sounding to describe
in this staid publication. Other recent audiences include SRI,
the well-known think tank in Menlo Park, CA, and leaders of the
personal computer industry at Esther Dyson's PC Forum, where
nanotechnology was part of the closing talk. The organizers gave
each speaker at the Forum a copy of Engines of Creation
Your suggestions are welcomed on books for this
Tools for Thought, Howard Rheingold, Simon
& Schuster, 1985, $9.95 paperback. Fun book covering past,
present, and future computer tools for augmenting human mind.
Includes hypertext (Nelson, Engelbart, Xanadu), ARPAnet,
"epistemological entrepreneurs," future of network
culture. Photos, personality profiles. For general reader.
Solid Clues: Quantum Physics, Molecular Biology, and the
Future of Science, Gerald Feinberg, Simon &
Schuster, 1985, $17.95 hardcover, $8.95 paperback. Feinberg, a
Columbia physics professor and FI Advisor, covers a broad range
of topics in science, focusing on physics and biology, including
careful thinking on where science is going. Rarely does a
scientist discuss the future of science: a special treat.
Glossary. Accessible (but challenging in parts) to non-technical
Odyssey: Pepsi to Apple ... a journey of
adventure, ideas, and the future, John Sculley, Harper & Row,
1987, $21.95 hardcover. A "business" book for those who
don't read business books. The story of a top Pepsi executive
(read about a strange but amusing corporate domain one hopes
never to enter) and how he escaped to Apple Computer to make a
difference in the real world. Epilogue on Sculley's dream, the
Knowledge Navigator, which sounds like hypertext and hypermedia
publishing with some AI.
Computer Lib/Dream Machines, Ted Nelson,
Microsoft Press, 1987, $18.95 paperback. Newly revised version of
the classic work which revolutionized the way we see computers,
by the man who is widely regarded as the father of hypertext.
The Tomorrow Makers, Grant Fjermedal,
Microsoft Press, 1988, $8.95. Previously reviewed popular book on
robotics, AI, and nanotechnology, now in paperback.
The Ecology of Computation, ed. Bernardo
Huberman, Elsevier Science Publishers, 1988, $80. Open-systems
perspective on advanced computing. Includes a set of three
papers on agoric market-based computation. For the computer
Proceedings of the IEEE Micro Robots and Teleoperators
Workshop: an Investigation of Micromechanical
Structures, Actuators and Sensors, held Nov. 6-11, 1987, Hyannis,
MA. 28 papers on topics such as "gnat robots,"
micromotors, and "Nanomachinery: Atomically Precise Gears
and Bearings." Available through technical libraries or call
the IEEE in NYC.
Coverage of nanotechnology continues to increase. In January
an excellent discussion of nanocomputers appeared in A. K. Dewdney's
"Computer Recreations" column of Scientific
American. It included a large color illustration of
nanomachines clearing fat deposits from a blood vessel. FI's
address was given for more information, and we've received over
300 requests so far, with more coming in.
Also in January, nanotechnology was featured in a television
special on supercomputers produced by Christian Science
Monitor Reports, shown in many U.S. cities on January
30-31. The January 28 San Jose Mercury News
discussed nanotechnology on page 2 in an article on scanning
tunneling microscopy; the article made the error--first made in a
Washington Post story and now common in articles
drawing on it--that mechanical nanocomputers will be a trillion
times faster than today's machines. Readers are asked to
help us make clear that they will be a trillion times more
compact, but not much faster.
On February 1 an article by Grant
Fjermedal (author of The Tomorrow Makers)
appeared in the Seattle Times in connection with a
lecture by Eric Drexler at the University of Washington. A brief
and fairly inaccurate article appeared in PC Week on
February 16. The nanotechnology article in Analog
mentioned in our last issue has been voted "best non-fiction
article" in the magazine's annual readership poll.
We've had many requests for investment advice drawing on the
prospect of nanotechnology. This is not our specialty, but we can
report on an investment book that does address the issue: Blood
in the Streets, by James Dale Davidson and
Sir William Rees-Mogg (Summit Books, 1987) devotes seven
pages to the topic, and more to technological change in general
and its effects on investment strategies.
Modeling on the Mac
Molecular Editor, Version 1.1
MicroChem, Version 2.0 [and 2.5]
[Note: Because of rapid progress
in molecular modeling software, the application-specific
information in this article is largely obsolete. For more recent
information, a "Universal
Molecular Modeling Software List" is available
on the Web as part of the NIH
Guide to Molecular Modeling.]
Molecular modeling software systems fall into two categories with
widely differing capabilities. Those in the first category enable
users to build and display three-dimensional models of molecular
structures; those in the second use information about molecular
geometries and energies to calculate molecular properties and
Molecular Editor, from Drexel University, is an
example from the first category. It can show a set of points in
three dimensional space as seen from any chosen angle,
representing atoms at those points as spheres (displayed as
shaded circles) and representing bonds between them as lines. It
can shade circles and lines to represent different atom and bond
types. The program, though, knows nothing about molecular
structures other than atomic sizes (which it uses to size its
circles). Users are free to draw in nonsensical structures; the
documentation notes that the program can be used to teach
generalized cartesian geometry.
Atoms can be positioned by dragging to a
location, by a move-to-origin command, or by entering
coordinates. Groups of atoms can be selected and then duplicated,
dragged, rotated, or reflected, enabling the user to build up
complex, regular structures with reasonable economy of effort
(see Figure 1). Since Molecular Editor can read and
write coordinates in text files, structures can be calculated
elsewhere and imported for display and manipulation. Like most
(all?) molecular modeling programs on the Macintosh, Molecular
Editor gives an intersecting-spheres representation of
bonded atoms: where two spheres should show a line of
intersection, one simply sees an overlap, with the circle
representing the closer sphere on top.
The program is generally well-executed. Its interface follows
Macintosh conventions closely, with no nasty surprises and broad
availability of the undo command. On-screen help is extensive,
and the documentation is reasonably complete.
The chief drawbacks relate to bugs and speed. Various
manipulations are best performed by cutting molecules (or
molecular fragments) to the Macintosh clipboard and pasting them
into another window for modification. Unfortunately this process
sometimes fails to bring some atoms or bonds across (for
mysterious reasons), and pasting--especially into a window with
another molecule in place--can lead to the equally mysterious
formation of improper bonds, sometimes to nonexistent atoms in
never-never land outside the window. Such bugs, while
frustrating, were rare enough to be tolerable. The speed problems
are also tolerable, but the rotation and redisplay of a structure
containing a hundred atoms or so can take about a minute. There
is reason to think that better code would speed this up
In general, Molecular Editor is a professionally
executed piece of software, though not rock-solid or fast. It
generally does what it is supposed to do and costs remarkably
little. It is available for $33 (including domestic shipping)
from Kinko's Academic Courseware Exchange, 4141 State Street,
Santa Barbara, CA 93110; phone 800-235-6919 outside CA,
800-292-6640 inside CA, 805-967-0192 outside U.S. (Macintosh
512e, Plus, SE, or II required; 800K disk).
MicroChem 2.0, from Intersoft, is a more
ambitious package in the second category of molecular modeling
systems [comments inserted in brackets describe reported changes
in version 2.5]. Its organic modeling unit includes several
modules that communicate via text-file representations of
molecular structure (tables of atom types, coordinates, and
bonding relationships) [version 2.5 is reported to provide for
more convenient interchange of structures between modules]. The
input module enables a user to provide coordinates or crude
sketches to describe a molecular structure. The basic molecular
mechanics module takes a description of this sort and modifies
it, attaching hydrogen atoms to any atom with dangling bonds and
adjusting the lengths and angles of bonds in ways that take
account of their preferred geometries. Another molecular
mechanics module enables users to compare the energies of sets of
molecular conformations generated by rotating one part with
respect to another around a single bond (holding each part's
geometry fixed). Other modules (parts of separately sold
"modeling units") include a zeolite maker, a
macromolecule maker, and a group additive properties unit for use
in drug design.
In general, the implementation of these modules gives a sense of
programming resources being stretched too far [though
improvements reported in version 2.5 show that further resources
are being applied]. The input and display modules have functions
that overlap with those of Molecular Editor, but the
implementation is poor. The user interface for the input module
in particular is painful. All choices are made through menus
(with no command-key shortcuts [added in 2.5]) and a click causes
an immediate action, with no chance either to see what target is
selected beforehand or to undo the operation afterward. Clicking
squarely in the middle of an atom can cause an operation on a
non-overlapping atom remarkably far away [now said to be fixed].
This combination of circumstances causes real problems with
deletion. One first chooses "delete" from the menu,
then clicks on the atom to be deleted. All too often, something
else then immediately and irretrievably vanishes. Further, one
cannot select and move a single atom, much less a group. The
display module interface is awkward but less can go seriously
wrong here. Still, it fails in the basic task of reliably
stacking atom-images in back-to-front order to make foreground
atoms obscure background atoms, rather than vice-versa (see
Figure 2) [this, too, is said to be fixed].
Figure 2. A segment of the Figure 1 shaft structure, done in
MicroChem, showing first the skeleton alone, then the locations
of the hydrogens, and finally the results of drawing the atoms.
[The atom-stacking bug is reportedly fixed.]
The first of the two molecular mechanics modules takes a crude
input structure and adjusts it to minimize its energy. In doing
so, however, it ignores interactions between non-bonded atoms;
these include electrostatic forces and the van der Waals
repulsions that keep two atoms from occupying the same place at
the same time. It takes account only of forces resulting from the
covalent skeleton of the molecule--forces that are often
dominant, especially in conventional chemistry. Faced with bonds
in linear or branching molecular structures, it simply constructs
a geometry having the preferred bond lengths and angles. Bonds in
ring and cage structures, though, are generally distorted; this
module calculates the shapes of these looped substructures by an
iterative energy-minimization procedure that essentially treats
bonds as springs and lets them settle toward their equilibrium
The neglect of non-bonded interactions may be adequate for many
purposes (and it keeps the time required for energy minimization
within reasonable bounds) but the design of nanomachines
typically involves these interactions in a major way. Non-bonded
interactions are surface interactions, and surface interactions
are what gears, bearings, and most other mechanisms are about.
The second molecular mechanics model can calculate the energy of
non-bonded interactions, but only during the rotation of one part
of a molecule with respect to another about a bond, while
treating the rest of the molecule's covalent framework as rigid.
This has its uses, but is inadequate for the design of typical
molecular mechanical devices.
And again, problems appear. The first molecular mechanics module
first minimizes the energy of looped substructures, then
constructs any non-looped structures that need to be added using
a simple geometric algorithm. Figure 3 shows a molecule of cubane
in two views; the module found the correct form for the looped
structure (a regular cube), then constructed irregular, crooked
hydrogens around it.
Figure 3. Two
views of MicroChem's version of the cubane structure, showing
irregular hydrogen placement.
To be useful for the design of simple nanomechanisms, the
molecular mechanics modules need to be combined and extended.
First, MicroChem needs options permitting energy
minimization and calculation using both bonded and non-bonded
interactions. Second (and more esoteric), it needs a way to let
users specify forces on atoms and constraints on their motion, to
allow (for example) calculation of energies for a bearing in a
variety of rotational positions. These more elaborate
calculations would inevitably be more sluggish, but overnight
runs are sometimes acceptable, and the greater abilities of the
Mac II and future machines will help. We should eventually get
systems adequate for amateur molecular-machine hacking, but in
the meantime, there is a need waiting for someone to fill it. [As
these bracketed comments show, however, MicroChem
itself is evolving at a fair pace.]
The MicroChem Organic Modeling Unit, Version 2.5, is
available from Intersoft, Inc., One Concourse Plaza, 4711 Golf
Rd., Suite 412, Skokie, IL 60076 (tel. 312-699-4143). Academic
license, $295, industrial license, $495. Special prices and
arrangements are available for multiple modeling units, multiple
licenses, high schools, and classroom use. (Macintosh Plus, SE,
or II required; preferably with a hard disk, but at least a
second 800K disk drive).
[Very Large] Space Station Design and Development,
May 4, Chicago Hilton, $148 prereg., $195 on-site. We've added
the "very large" to indicate that space colonies are
included; aimed at architects, May 3 evening reception for new
International Association for Space Architects. Contact
Guidelines, 800-634-7779 outside CA, 415-254-9393 in CA.
Third International Conference on Supercomputing,
May 15-20, Boston. Includes a presentation on nanotechnology on
May 20. Contact ISI, 813-866-2694.
Space Development Conference, May 27-30, Denver,
CO. Co-sponsored by FI, ask for FI discount. Includes one hour on
nanotechnology. Registration $60 through May 1. Contact Box
300572, Denver, CO 80218, 303-692-6788.
Gordon Research Conference on Immobilization and
Biotechnology, Aug. 8-12, Plymouth, NH. Includes
"Nanomachines and Molecular Assembly" on Aug. 9.
Preregistration $270-310. Contact 401-783-4011.
Directions and Implications of Advanced Computing,
Aug. 21, St. Paul, MN, $50. Plenary speaker is Doug Engelbart,
pioneer in hypertext. Sponsored by Computer Professionals for
Social Responsibility. Contact DIAC-88, CPSR/LA, PO Box 66038,
LA, CA 90066.
AI and Hypertext: Issues and Directions, Aug.
23, St. Paul, MN. Half-day workshop includes hypertext
publishing. Limited to 35, position papers due May 2. Sponsored
by AAAI. Contact Mark Bernstein, 617-782-9044.
Second Conference on Computer-Supported Cooperative Work,
Sept. 26-68, Portland, OR. Includes "technological,
sociological, organizational, cognitive, and task domain
perspectives." Sponsored by ACM. Contact Suzanne Sylvia,
Fourth International Symposium on Molecular Electronic
Devices, Oct. 1989, Baltimore/DC area. Watch this column
From Foresight Update 3, originally
published 30 April 1988.
Foresight thanks Dave Kilbridge for converting Update 3 to
html for this web page.