RE: Project Star Net

Tony Belding (
Mon, 27 Apr 1998 17:08:18 -0600

Anders Sandberg <> wrote:

AS> Why just one? I assume the energy needed for interstellar
AS> communication would be relatively manageable if it is properly
AS> collimated (no need for Forward's wonderful megascale lasers needed
AS> for fast interstellar solar sailing), so even if there are scale
AS> factors making it economic to have large transmittors there is likely
AS> no reason to have just one. The same goes for receivers.

What about signal interference? I assume we have limited EM bandwidth to work
with. You probably can't directionally distinguish signals from two stations
at interstellar distances. So, I reckon the most efficient way is to have one
directional transmitter and receiver for each adjacent node your station is
linked to.

AS> Bandwidth, error correction and receipts will be important
AS> here. Assume a human mind takes around 10^15 bytes - that will require
AS> a quite large bandwidth.

Indeed so. But I'm not sure how much EM bandwidth we have to work with, in
practical terms. This could be important to know, in terms of predicting what
sort of capacity the network will have. Potential investors would want to
know what sort of return to expect on their investment, for example.

AS> I have started writing up a nanotech seed for this purpose, I have a
AS> (not very finished or well planned) draft at

Thanks, I will check that out!

AS> Maybe we could develop the draft together, hopefully with useful
AS> input from the list?

That sounds good to me.

TB>> * What types of star systems should be candidates for
TB>> establishing nodes?

AS> Why not every system? There are likely asteroids everywhere, and extra
AS> capacity is likely a good idea.

There may be practical obstacles. I would not jump to the assumption that
asteroids are everywhere. Our own solar system resembled a demolition derby
in its early history, but that might not be typical at all.

As many as 80% of all stars are part of multiple star systems. The orbital
mechanics that are possible in these systems are rather mysterious. Nobody
knows if those systems can contain asteroids or planets at all -- and even if
they do, the orbits might be so eccentric that establishing a station and
keeping it running would be impractical.

Finally, a large majority of stars (I forget the exact statistic) are brown
dwarfs that may represent dark, cold, energy-starved planetary systems where
operations would be limited, and hardly anyone would want to visit anyhow.
These are really what I was thinking about when I asked the question. You've
got all these zillions of "dark stars" that are hard to even detect and don't
seem like they have much to offer.

AS> My communications technology knowledge is limited, but I would guess
AS> we want short wavelengths for maximum bandwidth (drawback: requires
AS> more energy to send than long wavelengths). Lasers sound fine to me.

Lasers? Can you distinguish the signal from the glare of the star that your
transmitter is orbiting? You said something about signal/noise ratios, but
the star itself could be the biggest source of noise around.

AS> I think the Internet model works fine: make it easy to connect, you
AS> only pay for your node in the network which both gives you access and
AS> acts as a part others can use.

I don't understand at all what you're saying here. Pay for your node? What
does that mean? The nodes are self-constructing and self-maintaining!

I considered an idea that each node would manage its own economic system by
issuing virtual "tokens" of some sort. These would be issued regularly --
say, once a month -- and would be valid during that time period. Sort of a
"use it or lose it" situation. Anybody could sign up to recieve one token per
month. Investors would automatically get extra tokens. Then the tokens can
be traded among different people, and used to bid for transmittor access.

(NOTE: It only makes sense to bid for transmission, not incoming data. With a
time lag of years between stations, it's not too practical to request
something from another node. You couldn't have anything like WWW, or at least
not within reasonable time-lag limits.)

If the station is located in a star system where material resources are scant,
it could issue access tokens to buy what it needs from the inhabitants.

So, if you were a non-investor in an otherwise un-inhabited star system, you
could sign up for your one "free" token, and that would be enough to buy
passage to somewhere else. After all, nobody is bidding against you for
access. On the other hand, if you were in a populated star system with lots
of network usage, your one token wouldn't be enough to ride. You could either
sell it to somebody, or you could buy enough tokens from other people to get a

AS> Note that if you use reproducing probes, you only need to pay for the
AS> original probe, and then it will expand on its own giving a huge
AS> return of investment. A good start might be an interplanetary net.

I don't think "return on investment" can have the usual meaning in this
instance. I mean, it doesn't matter how many nodes are created, the investor
can only benefit from the one in the star system where he's currently located.

An interplanetary net would be fundamentally different, since it could exist
within the bounds of a single economy. You can't really have an interstellar
economy for the same reason you can't have an interstellar government: the
distances are simply too great. There's no interchange of resources.

AS> Look at the net: an overloaded router tells the other routers that it
AS> is busy, and they route signals elsewhere.

That's only helpful if you're making a transmission across long distances.
I'm assuming, or guessing, that most people faxing themselves on the network
would want to go directly to an adjacent node. Then re-routing can't help.
Also, if a node has any kind of failure, it can take adjacent nodes /years/ to
find out about it, and they are transmitting data all that time!

AS> If travel is done by sending a copy of the traveller to the destination,
AS> keeping one copy inactive, and then waiting for a receipt (if received,
AS> delete the inactive copy, otherwise awaken it and tell it "sorry, access
AS> denied"), there is no big problem with occasional blocks.

That would be really rough, wouldn't it? You were going to visit Alpha
Centauri, then you wake up and find NINE YEARS have passed and you're still
here, because the A. Centauri station refused to accept you for some reason.
Obviously, this is a type of failure we would like to avoid as much as

AS> BTW, are there any standards yet for domain names in space?

Not that I'm aware of. A whole lot of new domain names are supposed to come
into service real soon now, so maybe there's something.

   Tony Belding