[Blog Post] Peter Bakker on the VMS

[Torsten]

Is there a test that can distinguish meaningless and undecipherable? And if there isn't, what does that mean for us?

No, such a test doesn't exists. It is possible to use a meaningful text as cover text for secret message and it is also possible to use a text which itself is meaningless for the same purpose. There are only tests for estimating the amount of information a given unreadable text can carry.

For example, what do we do with a plaintext encoded by a one-time pad whose key is irrecoverably lost? I suppose it would be undecipherable, but not meaningless -- even though the meaning can never be recovered.

This depends on the cipher method used as well as on the key. The key must be truly random, as long as the plaintext, never be reused in whole or in part and must be kept completely secret. Since the key must be truly random as a result also the encrypted message would be truly random.

If we could distinguish whether it had once been meaningful even though we can never read it, what would that tell us? Would it affect how we think of the VM as an artifact (e.g., hoax vs. book of secret science)?

There might be indications that a text was meaningful. For instance a nearly random sequence of characters could indicate an encrypted text and if the text is not random we are able to analyze its structure.