The Example in 12.4.4 seems to be for KERI genus/protocol --AAABAA and may have one error.

[daidoji]

Discussed in the KERI dev call today 4/4/2024.

The example in section 12.4.4 is confusing I'll quote it here for ease of reference

-FAB     # Trans Indexed Sig Groups Count Code 1 following group
E_T2_p83_gRSuAYvGhqV3S0JzYEF2dIa-OCPLbIhBO7Y    # trans prefix of signer for sigs
-EAB0AAAAAAAAAAAAAAAAAAAAAAB    # sequence number of est event of signer's public keys for sigs
EwmQtlcszNoEIDfqD-Zih3N6o5B3humRKvBBln2juTEM    # digest of est event of signer's public keys for sigs
-AAD     # Controller Indexed Sigs Count Code 3 following sigs
AA5267UlFg1jHee4Dauht77SzGl8WUC_0oimYG5If3SdIOSzWM8Qs9SFajAilQcozXJVnbkY5stG_K4NbKdNB4AQ # sig 0
ABBgeqntZW3Gu4HL0h3odYz6LaZ_SMfmITL-Btoq_7OZFe3L16jmOe49Ur108wH7mnBaq2E_0U0N0c5vgrJtDpAQ # sig 1
ACTD7NDX93ZGTkZBBuSeSGsAQ7u0hngpNTZTK_Um7rUZGnLRNJvo5oOnnC1J2iBQHuxoq8PyjdT3BHS2LiPrs2Cg # sig 2

• From the call it appears that -EAB on line 3 may have been included in error (from the call it appears that the semantics of -FAB are to lead me to know that three primitives will follow or multiples of 3 if the AB count was higher)
• This seems to use the count codes from --AAABAA keri genus/protocol.
  • Like I'd expect the -FAB based on the KERI(--AAACAA) or universal code tables in the current spec to be a CESR native message top-level fixed field signable up to 4,095 quadlets/triplets. However, in this example and by looking at keripy -FAB was probably what the comment says it is.
  • Similarly with -AAD I expected Generic pipeline group up to 4,095 quadlets/triplets as in the current table but in keripy it looks like it is an index group with 3 signatures.

It may be useful to implementers to fix this example and provide a few more examples to fully demonstrate the scheme.

[pfeairheller]

This is part of the larger issue of making all examples cryptographically verifiable.

[johnandersen777]

@pfeairheller @SmithSamuelM Could you please direct me to the best place to start on how to help make the examples cryptographically verifiable?

[SmithSamuelM]

Basically to make all the examples cryptographically verifiable requires writing code. As we discussed in the working group meeting on more than one occasion, the action item is to write a set of unit tests one for each spec example that generates the cryptographically verifiable versions of each example. So far this effort has yet to be resourced.