I'm following the advice of a user who answered a question on StackOverflow.
Page 15 section 4.A.4 of NIST call for proposal (the 2016 one) says:
For the purpose of estimating security strengths, it may be assumed that the attacker has access to signatures for no more than 2⁶⁴ chosen messages;
My original question was where I'd find NIST's requirements about the minimum number of signatures that can be generated with a digital signature scheme (let's say some algorithm).
I don't manage to get the meaning about a minimum number of possible signatures from this sentence, as @DannyNiu seems to understand it in his response (i.e. a signature scheme has to provide at least this lower bound number of signatures).
What I rather understand from that sentence is that it's just an assumption for the security evaluation, i.e. the evaluation is done assuming an attacker has less than 2⁶⁴ chosen plaintexts with their matching signatures. We ideally want signature schemes with a virtually unlimited amount of possible signatures.
Maybe someone could try to explain this sentence to me, in case the @DannyNiu is right, or otherwise confirm that I understand it correctly.
I should provide some background/context: This document is (among others) about so-called hash-bashed signatures. Current signature schemes can be used endlessly (i.e. you can create an infinite number of signatures). Hash-based signature schemes, on the other side, have a limited number of signatures that can be generated. So NIST should have a requirement like any proposed signature scheme SHALL be usable for at least x signatures.
2^64 would imho indeed be a reasonable value, but it might also be 2^50 or 2^80.
[For the purpose of ELU, 2⁶⁴ is merely an extremely large number, approximately 18.4 trillion.]