Re: Dyson shell redux

From: Robert J. Bradbury (
Date: Thu Mar 15 2001 - 12:10:03 MST

Jeff's idea is to my knowledge novel and quite provocative.

Now obviously since from Spike's comments we know solar
sailing is possible, one could in theory surround the
sun with a ballon of solar-sail thickness which should
be supported quite happily even without spinning it.

You probably need to run Buckytube cables around the
the thing to provide the strength necessary to support
micrometeorite punctures and resist the pressures resulting
from solar flares. Lots of calculations to do here to
study the feasibility I suspect.

I don't like Ander's application of all of the mass in the
solar system to this application because it doesn't seem
necessary. (I also get about twice as much material as he
does to work with in the solar system -- something we need
to resolve at some point...). His number is for a sphere
at 1 AU which doesn't have to apply to Jeff's proposal.

A principle you can try to work with is how do you
make the best use of the available energy with the
smallest amount of material possible. If you only
use a fraction of the available material for Jeff's
Dyson *SPHERE* Lite, then Ander's can use the rest of
it for building a Nutronomium-Brain (I forget what he
calls these).

Now the question is what kind of a good computer architecture
can you fit on top of a Dyson Shell Lite. The one that comes
to mind for me is a 2-D cellular automata. I.e. no 1-cm^3
rod-logic CPU's, instead you cover the surface with
very thin CPU's that only communicate with their nearest
neighbors. Now, what kind of calculation that is good for
I don't know (we really need to get some heavyweights
in computational architectures on this list). It seems
to me that a Dyson Sphere Lite is also a "memory" lite
architecture since you don't have many atoms around in which
to store your information. So Jeff's architecture weighs
in with a high computational capacity / memory ratio.
Perhaps this is useful when a rogue "un"-friendly
MBrain comes marauding through your solar system
(because a nearby gamma-ray burst it once experienced
flipped several bits in Eliezer Yudkowsky's finely-tuned
but fragile "friendliness" algorithm), and steals the
bulk of your computronium, leaving you with some matter dregs
that it wasn't worth the energy cost to alter course to harvest.

If you want to avoid heating up the sun, you are going to
have to simply take the hits caused by the UV photons and
the solar wind (i.e. you can't reflect them back to the sun).
Your nanobots are going to have to run around continually
taking apart and reassmbling the high-energy exposed surfaces
to get them back to 100% efficiency. But hey, that's life.
You may have an interesting problem that you have to continually
throw away the material you are accumulating from the solar
wind (otherwise you upset your delicate balance between
the gravitational attraction and the support the wind provides).
Or you could use it to "grow" your sphere (but doesn't this
become unbalanced at some point???).

There is one other way of handling a "sphere" architecture but
it requires a *lot* of material (probably more than we
have available). That is the method used by Anderson
and Pohl in "The Saga of the Cuckoo". They use momentum
transfer of large fluid volumes to provide the support
(you could also probably use particle beams). However
that to me seems to be a recipe for wasting a lot of
energy that could otherwise be used to power your
computronium. But in that "reality" the whole point of
the sphere was to provide a "world" in which sub-optimal
CHNO based life-forms could exist, so you don't have the
freedom that we do to consider things like Virtual Realities
based on solar-wind suspended solar-sails with micron thick

Lets try in the discussions to be careful with terms --
a "shell" would be lots of nested orbiting satellites
(like a Matrioshka Brain), a "sphere" would be a standard
fully contiguous surface. These ideas have been misunderstood
for too long and its time to set them right.

FYI: the complete list of Dyson Sphere/Shell Papers is at:
A retrospective on how things got so messed up is now at:
  (Sorry its not in hypertext yet, someday soon).

Jeff, there are detailed calculations that have been done
on some of the support issues by several Russian scientists
that I've had "roughly" translated. If you want to go into
this in more detail, send me a formal request for access
to the database if you can't click through to the papers.


This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:40 MDT