A question about codes, hacking and cyphers

[drivin_me_nuts]

I had a conversation with someone recently regarding cyphers.

It went like this..

He said something along the lines of,

If you created a code based on transposing letters from a series of books, for all practical purposes, the code would be unbreakable. Particularly if the resultant letter had rules attached to it that drove the selection of another letter from another book, that then became the final letter of the message... say 3 iterations of different books to get to the encrypted letter.

Hopefully this makes sense.

So, if that was the cypher, would it be 'unbreakable'.

From someone who knows nothing about how cyphers work.

[stablemate]

drivin_me_nuts said:

So, if that was the cypher, would it be 'unbreakable'.

From someone who knows nothing about how cyphers work.

Not entirely. A very similar proposition to how the Enigma machines worked (not exactly). Letters transposed, and then the transposition itself altering how future iterations of the same letter transpose themselves.

It would be breakable with knowledge of which books were used, without knowing the transposition rules to start with. With vast amounts of computational power it would even be possible to work out which books were used.

Breakable yes, but far from easily.

[grumbledoak]

drivin_me_nuts said:

So, if that was the cypher, would it be 'unbreakable'.

This isn't that well described but I don't think so. The only truly unbreakable cipher is the One-Time Pad. In this the key must be random, as long as the message, and only ever be used once. Lookups and rules might sound complex and confusing but that type of encryption can be very easy for computers to 'crack'.

[21TonyK]

Uncrackable in the sense that it cannot be cracked in an acceptable time to the person cracking it then that's achievable. Completely uncrackable? No. Given the will, time and resources what can be done can be undone.

[CypherP]

I'm known to be quite difficult to crack, yes.

[Opel-GT]

https://en.wikipedia.org/wiki/One-time_pad

[Wacky Racer]

Off topic, but a few months ago I had a trip to Bletchley Park.

If you have never been, well worth making the effort, if you are in the Milton Keynes area.

[glazbagun]

Not totally related, and this might be teaching people to suck eggs, but this is the simplest video I've seen on how Public Key Cryptography works:

https://www.youtube.com/watch?v=YEBfamv-_do

[Vaud]

Wacky Racer said:

Off topic, but a few months ago I had a trip to Bletchley Park.

If you have never been, well worth making the effort, if you are in the Milton Keynes area.

Great museum. I went for the first time about 15 years ago and was shown around by one of the women who had worked in the huts. Great stories.

[schmalex]

I work in High Grade encryption. Fortunately, not in the mathematics side!

I had written a long reply, but then bottled posting it!

Simply put, yes it would be breakable with some of tools and techniques available today but would take long enough to deter all but the most committed.

Edited by schmalex on Wednesday 27th April 21:45

[bucksmanuk]

glazbagun said:

Not totally related, and this might be teaching people to suck eggs, but this is the simplest video I've seen on how Public Key Cryptography works:

https://www.youtube.com/watch?v=YEBfamv-_do

Thanks for that - fascinating

[rotarymazda]

drivin_me_nuts said:

If you created a code based on transposing letters from a series of books, for all practical purposes, the code would be unbreakable.

Only if you keep the method secret, your books secret and have enough books.

I worked on encryption products around 15 years ago, brute force cracking of billions of combinations per second is not hard.

Since keeping a method secret is hard, algorithms like AES/3DES/Eliptic Curves are publicly available but use mathematical operations that are very difficult to reverse.

[drivin_me_nuts]

Thanks for all the replies, it's a fascinating subject to read about and the museum is definitely on my list of places to visit when next up MK way.

Cheers,

$£^

[droopsnoot]

There's been a couple of interesting documentaries on BBC4 recently on the subject, I think one of them (the one about Gordon Welchman) was on again last week. Well worth having a look out for them when they repeat them.

[Vaud]

For those wanting a gentle introduction to the topic:

Simon Singh's The Code Book is very well written in my opinion:

http://simonsingh.net/books/the-code-book/

For masochists - who need help sleeping at night, David Kahn's The Codebreakers – The Story of Secret Writing (Revised edition) is exceptional but not light reading:

https://www.amazon.co.uk/Codebreakers-Comprehensiv...

Masochist, because as you fall asleep reading it, the sheer weight of the book hitting you in the face will cause significant injury.

[ATG]

Simple transposition of one letter for another using something like a one time pad is unbreakable regardless of the computational effort thrown at it so long as you only encode a small amount of text. A short message doesn't give the code breaker enough data to play with. No statistical patterns will emerge. They may well manage to decode your message but they'll also come up with many alternative decodings and they'll have no way of discerning which one is correct.

[Vaud]

ATG said:

Simple transposition of one letter for another using something like a one time pad is unbreakable regardless of the computational effort thrown at it so long as you only encode a small amount of text. A short message doesn't give the code breaker enough data to play with. No statistical patterns will emerge. They may well manage to decode your message but they'll also come up with many alternative decodings and they'll have no way of discerning which one is correct.

Correct. A (true) one time pad used correctly is unbreakable for straightforward attack.

But they are buggers to distribute, manage and use - much like keys were prior to Public Key Cryptography.