Discussion:
Asymmetric cryptography without computers
(too old to reply)
Jon Roland
2014-08-14 15:05:52 UTC
Permalink
We happily use public-key crypto tools every day, but we use them on computers. So I pose this question to anyone on the forum who might be able to answer.

Suppose we want bright people equipped only with pencils, paper, and an abacus to manually generate a 2048-bit key pair and encrypt a 10KB message using one of the keys. How long might it take to do that? Hours, days, weeks, months, years?

Now suppose the sender does not provide the receiver with the complete key, but omits parts of it, those parts to be generated from data observed in a future event he knows will happen on a date certain, but which cannot be predicted, as a way to make sure the message is not decrypted before that future event.

The receiver would also not have a computer to do the decryption, especially not quantum computers to crack it even without the missing parts.

In effect, the message is a time capsule that can't be opened prematurely, even if future recipients get computers.

If you suspect I am researching for a science fiction story, you are correct.

Any help would be appreciated.

I should further explain. The enterprise is directed by a person from the future who knows the celestial events that provide the data to generate the missing parts of the key. Perhaps the celestial coordinates of the seven brightest objects surrounding a nova. So the method of generating the missing parts is set at the outset, and the events occur hundreds of years after the original message is encrypted.

Think about how a visitor from the future to a medieval past would set a chain of events in motion, but need to prompt future generations with messages to keep them on a track to deal with some distant future catastrophe that they would need the information to avoid. The visitor has only one chance to guide history despite any loss of faith or direction that successors of the original senders of the message might undergo. He can't just provide them all the direction at once, because after some generations it would no longer be credible, or attacked as witchcraft. (Nostradamus encoded his predictions for a reason, but not as usefully as he might have done.) So the visitor reinforces the guidance with several messages about lesser but important disasters and how to prepare for them, to be revealed just before they occur, but in time to prepare. (E.g., eruption of Yellowstone, tsunami caused by collapse of La Palma, etc.) Each revelation rebuilds confidence in the reliability of the next ones, and thus political support for the sacrifices that need to be made to prepare.

So I need to know whether such cryptography is feasible without computers.
Le Forgeron
2014-08-14 20:42:55 UTC
Permalink
Post by Jon Roland
We happily use public-key crypto tools every day, but we use them on computers. So I pose this question to anyone on the forum who might be able to answer.
Suppose we want bright people equipped only with pencils, paper, and an abacus to manually generate a 2048-bit key pair and encrypt a 10KB message using one of the keys. How long might it take to do that? Hours, days, weeks, months, years?
Now suppose the sender does not provide the receiver with the complete key, but omits parts of it, those parts to be generated from data observed in a future event he knows will happen on a date certain, but which cannot be predicted, as a way to make sure the message is not decrypted before that future event.
The receiver would also not have a computer to do the decryption, especially not quantum computers to crack it even without the missing parts.
In effect, the message is a time capsule that can't be opened prematurely, even if future recipients get computers.
If you suspect I am researching for a science fiction story, you are correct.
Any help would be appreciated.
I should further explain. The enterprise is directed by a person from the future who knows the celestial events that provide the data to generate the missing parts of the key. Perhaps the celestial coordinates of the seven brightest objects surrounding a nova. So the method of generating the missing parts is set at the outset, and the events occur hundreds of years after the original message is encrypted.
Think about how a visitor from the future to a medieval past would set a chain of events in motion, but need to prompt future generations with messages to keep them on a track to deal with some distant future catastrophe that they would need the information to avoid. The visitor has only one chance to guide history despite any loss of faith or direction that successors of the original senders of the message might undergo. He can't just provide them all the direction at once, because after some generations it would no longer be credible, or attacked as witchcraft. (Nostradamus encoded his predictions for a reason, but not as usefully as he might have done.) So the visitor reinforces the guidance with several messages about lesser but important disasters and how to prepare for them, to be revealed just before they occur, but in time to prepare. (E.g., eruption of Yellowstone, tsunami caused by collapse of La Palma, etc.) Each revelation rebuilds confidence in the reliability of the ne
xt ones, and thus political support for the sacrifices that need to be made to prepare.
Post by Jon Roland
So I need to know whether such cryptography is feasible without computers.
Assuming the astronomical events are the colour of the full moon (white
or red)... that's one bit, and the totality or partial occlusion of the
new moon... that's another bit, 2 weeks later. Of course, as seen from a
specific place (choose something which remains many eons, such as the
top of a mountain, or a nearby valley (with a monastery of eternel monks
devoted to observing and keeping the data).

Then the sender could simply encode its message with OTP (one time pad)
mechanism.

At 26 bits per years, it might be wise to have a short alphabet (5 bits
?), or use redirections to a sacred book of code (which contains many
opposite directions and sentences, and ends each section with part of
the next key/pattern (and holes to fill)).

There might be a need for an error-correction code (doable by hand), in
case a monk fails to assist to the event (due to a local war or so).
It means that the message might be purported a bit of time before the
end of the transmission, but it is only confirmed at its very end.
(which can lead to some tensions of prozelits when they find they were
wrong about the message "Wear a blue dress" and it turns out as "Prefer
rice to wheat for the next two year".)

It might be wise also to allow starting from clear data every century or
so, with some external events, in case of misbehaviour of previous monks
and powers that try to manipulate the data. (reset the decoder when
seeing a comet at the nearest position from the sun ?)
--
Email to the Reply-To address will not be read.
IQ of crossposters with FU: 100 / (number of groups)
IQ of crossposters without FU: 100 / (1 + number of groups)
IQ of multiposters: 100 / ( (number of groups) * (number of groups))
Loading...