Re: [openpgp] Mining protection in fingerprint schemes

[Phillip Hallam-Baker <[e-mail]>]

Fingerprints are used for two purposes.

1) To define a trust anchor.
2) To verify that a purported trust anchor is correct

Only the first of these requires the ability to enter the fingerprint
at a keyboard. The most efficient way that is compliant with existing
keyboards etc. is Base32. While not all keyboards have latin
characters, there is an effective limit on the number of keys, screen
real estate etc.

The second requires only the ability to compare the fingerprints. Here
the PGP word scheme shows a more efficient approach, using an alphabet
of effectively 256 glyphs that are phonetically distinct words.

But we need not stop at 256 code points. We could just as easily use
Base65536 or Base1048576 alphabets. Comparing pictures, words,
whatever.

Words are my favorite as they allow keys and/or fingerprints to be
encoded in messages. Lets say that our alphabet has a restriction that
the words have to be 4 letters or more.

A 128 bit key in base65536 requires 8 words, lets say these are boat
syringe lawnmower noise suitcase horse advice loaf. You can make up a
cryptic phrase to encode the key steganographically.

Fix the boat, use a syringe. The lawnmower noise is bad. etc.


The downside to word dictionaries is that they are language specific.
Images are better.

For the Mesh, I plan to eventually use vocabularies of curated images.
Get a bunch of 256 volunteers, give them the task of curating
collections of 256 picture themes each to find pictures that are
visually distinct.


I have more here:

https://datatracker.ietf.org/doc/draft-hallambaker-udf/