Open questions
The questions a careful reader arrives with and this project cannot answer yet: the technical ones, the two that are not technical, and the ones about scope.
Technical
Four questions that can be answered with a measurement, once there is something to measure.
- How large should the quorum be, and who is in it?
- It is the component that decides whether an attestation means anything, and the only one that holds a key and carries a running cost. Verification itself is settled at 6.4 ms per proof. Size, membership, slashing and who pays the gas are not, and they are the critical path now that the wrapper is gone.
- How long does Layer 1 proving take in a browser?
- Halo 2 proving in a browser is settled: the shielding path does it in 16.5 seconds. What is not settled is this circuit's own number. Scaling its 2.10 s single-threaded floor by the 4.3× penalty measured on the same kind of arithmetic suggests about nine seconds. That is still an extrapolation, and the payload it would add is unmeasured.
- How many confirmations before an anchor is posted?
- Zcash settles in probability, so an anchor can be reorganised out from under an attestation that has already been written. How long an attester waits is a real parameter. The shielding path already picks 10 blocks for its own anchor, which is a starting point rather than an answer.
- Memo or public input for account binding?
- The circuit takes the account as a public input today. A 512-byte encrypted memo can carry it instead, and that path asks no integration of any wallet, because it works in every Zcash wallet that already exists.
The harder two
Neither of these is a cryptography problem, which is what makes them the harder two.
- Who reads an attestation first?
- A card's value is not in the plastic, it is in being accepted. Visa's hard problem was never the card, it was recruiting merchants. This is separate work rather than a side effect of building the card.
- How does this pay for itself?
- The privacy graveyard is full of projects that died on monetisation rather than on cryptography. It has to be answered from the first week, not the last.
Scope
How far this reaches, and what it costs to keep reaching.
- What share of migrated ZEC does this actually serve?
- Serving Ironwood only means the addressable base is whatever has crossed the turnstile. Migration is mandatory and one-way, so the figure only rises. The figure itself has not been checked, and it is the first thing anyone will ask.
- Who runs the attesters, and who pays their gas?
- The only component with a recurring cost rather than a one-time one, which connects it directly to the revenue question above.