Walk-through and help rewrite new protocol spec with researchers

[ch1bo]

Why

The formalism used in the paper is not capturing well the nature of on-chain verification within Cardano. For example, validators do not produce values, but merely check existing ones. Hence, the Hydra researchers are drafting up new specification in a slightly different formalism to more faithfully capture the checks to be done on-chain.

What

As they have created a first draft here (https://docs.google.com/document/d/11LlTTUuHu-3d-krq-O1h59kHm93Kb4cCQ6GhOcodFys/edit#), we now have a set of meetings to walk through both, the formalism and the actual specification, and review it together with them. This story captures the work on these sessions and the outcome being

• A specification of the on-chain verification part
• A list of any gaps already identified

[ch1bo]

Gaps so far

• SM cid checks - how would we be sure an NFT is indeed an ST? pack it in datum instead?
• ST in output value - seems not be checked right now
• T_max very big on close/contest (close, but only have deadline in 10y)

[ch1bo]

Things we do differently right now (on top of the ones above)

• Check committed outputs are reimbursed on v_head instead of v_commit
  • individual checks would "collapse" if two identical TxOuts were committed and only one needs to be reimbursed to pass
• Validator can assume first output is v_head output (simplification)
• Snapshot number and utxo hash need not be in the redeemer on close/contest
  • the output datum needs to include them and we can check using that
  • only the signature is enough
• Do not allow contest by closing party and only allow contest once
  • need to collect who contested in the datum and check it
  • needed? leads to upper bound of T_final when we do this 👇
• Deadline handling changed between paper and V2 -> T_final gets pushed out on each close/contest
  • this is easier to determine and scales with number of parties

[ch1bo]

@KtorZ @ffakenz I would say this subtask of #448 can be called done as we have a first draft of the on-chain spec + above list of currently identified gaps, which we also collect in the overall follow-up story #452