lkuper / CSE232-2020-10

A graduate course on distributed systems
https://decomposition.al/CSE232-2020-10/
11 stars 1 forks source link

Discussion Questions for "Knowledge and Common Knowledge in a Distributed Environment" #14

Open versey-sherry opened 3 years ago

versey-sherry commented 3 years ago

@ksuhr1 @git-ram @maxwellrbradley @andavid0

maxwellrbradley commented 3 years ago

@lkuper Lindsey, the question responses are now ready for you to review.

lkuper commented 3 years ago

Can you leave the "Discussion summary for Q1", "Discussion summary for Q2", etc. headings on the page so we can tell where the question ends and the discussion begins?

Other than that, the main issue is that there is some confusion between "everyone knows that everyone knows P" ("E^2 knowledge" in terms of the paper) and "common knowledge". Common knowledge of P is when everyone knows that everyone knows P, but also everyone knows that everyone knows that everyone knows P, and so on, infinitely. It's true that E^2 knowledge is useful, but it's not the same as common knowledge. (Also, E^2 knowledge is achievable in a distributed system under the omission model, but common knowledge is not. Same for E^k for any finite k.)

Q1:

Q2: all looks good except the last part, which has some confusion between E^2 knowledge and common knowledge.

Q3: I'm not sure what you mean by "a dual form of" at the end. Halpern and Moses meant "dual" in the mathematical sense. It's a slippery concept to pin down, but it involves something like two different points of view of looking at the same object, where those two different points of view mirror one another. In class we talked about "exists" and "for all" being dual. More examples here: https://en.wikipedia.org/wiki/Duality_(mathematics)#Duality_in_logic_and_set_theory

git-ram commented 3 years ago

@lkuper We updated the answers according to the points you mentioned. Please let us know if anything is still unclear. Thanks