[openpgp] Fingerprint Workfactor Hardening.

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

I had been somewhat concerned that this might be affected by

https://www.google.com/patents/US7929689

However looking through the patent, it appears that the inventive step
Microsoft is claiming is the use of a salt to make this process more
efficient rather than the process itself.

This does not put the mechanism in the clear, we still need to go through
the Microsoft lawyers to be safe. But it is a lot easier to get a company
to agree that a scheme doesn't infringe than permit open use of a valid
claim.


I am working on the doc right now. Note this is an update of the UDF doc
that has the purpose of pinging the MSFT lawyers.

Compressed Presentation

Fingerprint compression permits the use of shorter fingerprint presentation
without a reduction in the attacker work factor by requiring the
fingerprint value to match a particular pattern.


UDF fingerprints MUST use compression if possible. A compressed fingerprint
uses a version identifier that specifies the form of compression used as
follows:


96 No compression

97 First 25 bits are zeros

98 First 40 bits are zeros

99 First 50 bits are zeros

100 First 55 bits are zeros


Thus the fingerprint that would be represented in uncompressed form as
MAAAA-AAWIY-LTMFTG-CZTRO is instead represented as MBWIY-LTMFTG-CZTRO.