Cryptography - Feedback and local tweaks from UCB run-of-show incoperated

[nuke-web3]

• [ ] update https://github.com/Polkadot-Blockchain-Academy/pba-cryptography--master as well from student feeback and https://gist.github.com/naterarmstrong/a04cd8604dc4604ba852b6a329cddb77
• [ ] update the mod README with the schedule that actually happened for reference.

Feedback form results

TODO before UCB tag is applied

• [ ] https://github.com/Polkadot-Blockchain-Academy/pba-cryptography--master/issues/2
• [ ] Review https://github.com/Polkadot-Blockchain-Academy/pba-content/issues/604 for things we missed

[naterarmstrong]

Next time for the assignment:

• Get a handle on what the bar for mandatory should be
• Teach more in lecture on how to construct merkle proofs step-by-step, should make multiproofs easier for everyone
• Make some hidden tests required for passing. Make it clearer how optionals tie into passing, if would not otherwise pass with mandatory alone. Separate edge-casey hidden tests from correct-overall-implementation hidden tests when aggregating for grading.

[naterarmstrong]

For lectures:

• Possibly break out a whole lecture on cryptographic proofs, proving, trust, etc.
• This can talk about merkle proofs, inclusion proofs and non-inclusion proofs in both trees and tries, proofs of validity, interactive proofs of computation, non-repudiation (proving a message was sent), etc. Need lots of diagrams.

[naterarmstrong]

Summary of initial thoughts from feedback:

Seems like a group of students didn't actually struggle on the theory of some of the multiproofs stuff, but just aren't used to coding algorithmically like that. Also comments on how much the first mandatory exercises were about going into the docs I'll probably write that an instructor should go and do the first exercise of the first three sections in front of the class, sharing their screen Can do cargo doc --open, and a rough explanation of how to explore the code, how to read library docs, etc. Hopefully people already know that, but doing it in front of the class would let them at least know how they should be doing it, and hopefully struggle more on hard things than due to not knowing best practices Also lots of requests to cover merkle proofs more in depth. I think honestly a whole section on proofs would be pretty valuable. There are a lot of forms of cryptographic / computational proof in blockchain (proofs of validity, interactive proofs, fraud proofs, merkle proofs, even just non-repudiation), and I think they could be pulled out into a nice lecture after hash based data structures Also updating the slides for hash based data structures could do a lot. They are good illustrations, but not dynamic enough IMO. Going step by step for certain things would do a lot for student understanding, based on what I saw helping people on saturday But overall most people seemed to feel very positively about the module as a whole, and felt they know a ton more about the field:

Another thing is that some people would like it to be more hands-on in some way. There was pretty substantially differing opinions on how they would like it to be hands-on (some wanted math, some wanted application, some wanted coding), and I'm not sure how feasible doing much any of that would be. However, there was definitely the consensus that there was a bit of a gap between the lectures (high level theory) and assignment. I'm sure there's a way to bridge that a little better. Not sure exactly how at the moment though

[naterarmstrong]

For lectures again:

• If we have 3 days for cryptography again, talking about cryptography / privacy in society could be a very interesting lecture, although potentially off-topic or too political. Talk about relevant legislation in US/EU (very relevant if NDAA passes in US), end-to-end encryption in practice, anonymity and crime + anonymity and government, privacy and networking (tor), etc.

[nuke-web3]

related: https://github.com/Polkadot-Blockchain-Academy/pba-cryptography--master/issues/4

[naterarmstrong]

@NukeManDan I think everything necessary for opening slides should be good. There are still some ideas in this issue, but maybe they can be moved to a general future ideas issue, and we can close this one out.

The assignment-specific changes don't need to happen soon, so they can sit for a while.