Crosslink design first shown
Daira introduced the Crosslink design that improves on ebb and flow, and the NU5 retrospective schedule will be revisited in four weeks.
Agenda summary 8
ZCAP Minor Grants vote notice
Zcash Foundation · Jack Gavigan
- There are 20 Minor Grants applications totaling over $225,000, but only $75,000 is available to allocate.
- ZCAP participants were urged to vote via the Helios voting email.
ECC update: Zcashd 5.7.0 and mobile SDK
Electric Coin Company · Kris, Zooko
- zcashd 5.7.0 was recently released. 5.6.1 is scheduled to halt at EOS on October 11, so it must be upgraded within about 6 days to keep the network stable. The window is short this time because the release was late.
- Zcash SDK 2.0.0 and hotfix 2.0.1 came out for both iOS and Android. The Android build for Zashi beta testers will be distributed within a few days, and the latest iOS build came out that day.
- The Zashi beta admits people on the waitlist in stages. To take part, contact Kris Nuttycombe or Zooko.
Zebra update
Zcash Foundation · Teor
- Testing and code cleanup for the Fast Spendability RPC are being wrapped up, and it will be included in the next release. Enabling the getblocktemplate RPC in the production build is waiting on a bug fix to be merged.
- The team also plans to offer the progress display interface in the production build via a setting. During preparation for the next release, a sync performance problem was found in pre-testing and is being fixed, and a release is expected within 1-2 weeks.
FROST community call results and status
Zcash Foundation · Conrado
- On the recorded FROST community call, the problems wallets and others must solve to actually deploy FROST were discussed, and the next steps will be planned from that feedback.
- The crate has been audited and works, but as a distributed protocol it needs a means of passing messages. The team is looking for the simplest deployment method. All audit findings were addressed.
- The team is giving the auditors feedback on whether the impact description of each finding is accurate, and the final report is due soon. After finishing the refactoring, it plans to release a release candidate, collect API feedback for a month or two, and then release the stable 1.0.0.
Sustainability Fund ZIP update
Shielded Labs / Equilibrium · Mark Henderson, Jason McGee
- Three granular ZIPs that build on each other were written: establishing the ZSF, the transaction fee split, and issuance curve smoothing reflecting ZSF deposits.
- Str4d explained that their intent was for the ZIP to describe how issuance transitions from the existing to the new one rather than a specific activation height. Daira said to put a placeholder in the height spot and add the rationale for the choice.
- Teor proposed verifying in code whether issuance actually increases or decreases at the chosen activation point. The current ZIP has issuance about 0.5% lower than the existing schedule over four years.
- Conrado pointed out that if it is activated before the halving without adjusting the curve, Dev Fund recipients would receive less than expected. Conrado therefore said activation after the next halving should be considered.
Trailing Finality and Crosslink design progress
Electric Coin Company · Nate, Daira
- Daira is working on a design called Crosslink that improves on Snap-and-Chat from the ebb and flow paper. Nate thought it could address some or all of the security concerns raised at Zcon4.
- The paper does not address that it must be possible to spend finalized transactions that are not in the best chain. Daira's design fixes this and aims for no security regression compared with the best-chain protocol alone, even if the BFT protocol is broken.
- Nate plans to post a Crosslink update, even if unfinished, within 1-2 weeks. Jack Gavigan advised writing down the design's goals and rationale first.
- Daira explained that the 100-block rollback limit in current zcashd and Zebra works by having the node shut down rather than rejecting the rollback, so it is not the same kind of finality.
NU5 retrospective schedule and network upgrade planning
Daira, Nate, Jack Gavigan, Str4d, Teor, Kris
- ECC said its emergency mode is nearing its end, and that once SDK distribution support is done, the NU5 retrospective will be possible in about a month. It was decided to discuss again in the late slot in four weeks.
- Str4d calculated that to activate the smoothing ZIP in time for the next halving, node deployment must happen by mid-next-year at the latest, allowing for EOS halt margin, giving up to about 6-8 months from now.
- Teor distinguished upgrades with fixed deadlines like the halving from those whose schedule can slip, and suggested narrowing down to only what must meet a deadline. Teor pointed out that an excessive complexity budget has caused schedule delays and mistakes.
- Daira thought the v6 transaction format, explicit fees, ZSA, and ZSF smoothing and deposits are about the maximum one upgrade can hold. Introducing v6 before ZSA activation has the risks of a format change and parser definition issues.
- Kris mentioned the idea of separating memos from outputs, and Daira mentioned shielded transaction aggregation (zcash/zcash issue 4946). The consensus was that the latter is hard for NU6.
Miners' long-term incentives and the fee split
Shielded Labs / Equilibrium · Mark Henderson, Teor, Daira, Jack Gavigan
- Mark Henderson announced that the ratio for splitting fees between miners and the ZSF deposit was changed from 50/50 to 60/40.
- Teor raised a security concern that if fees or the subsidy shrink too much, miners gain an incentive to fork or roll back the chain. It could come up first if the deposit fee approaches the block subsidy.
- Daira expected it to become a problem after two or more halvings, when the miner subsidy has dropped substantially. Teor asked that the ZIP state that it is fine for now and that it be revisited if economic conditions change.
- Mark said they will state in the ZIP that it is fine in the short term and the conditions for revisiting later.
- Jack Gavigan said that during the ZIP-1014 discussion miners showed no preference in block signaling, and that participation may be low this time too.
This summary is based on the original meeting notes; check them for exact details.