Crosslink ZIPs

[shielded-nate]

👋 I'd like to get a placeholder ZIP for Crosslink.

What else needs to happen here before an incomplete stub can be merged?

• Copy the zip-template.md for crosslink; change some headers with provisional crosslink info.

[nuttycom]

👋 I'd like to get a placeholder ZIP for Crosslink.

What else needs to happen here before an incomplete stub can be merged?

• Copy the zip-template.md for crosslink; change some headers with provisional crosslink info.

Please follow the ZIP 0 process for draft ZIPs, it's very clearly specified: https://zips.z.cash/zip-0000#zip-workflow

[daira]

What else needs to happen here before an incomplete stub can be merged?

Stubs are only created to reserve a ZIP number. In the case of Crosslink, I don't think it's clear yet how many ZIPs will be needed. There will probably need to be at least one for the BFT protocol, one to specify either a new block header version or how equivalent information is encoded in the coinbase transaction, one to specify the finality gadget, etc. We don't merge ZIP PRs that are just outlines without any substantive content.

[shielded-nate]

Merged from shielded-nate#5 , and quoting that summary:

---

This is the beginning of structuring Shielded Labs's Crosslink design specification into ZIPs. This version is mostly generated by copilot synthesizing a few sources, and is full of misleading, incorrect, or missing pieces. However, it begins a basic structure with an "overview ZIP" and a "construction ZIP" (which is intended to have similar scope to the TFL book "Crosslink 2 construction").

More ZIPs expected, perhaps for:

• BFT protocol
• Crosslink-specific network transport protocol (see below)
• Changes to PoW state and consensus
• Roster mechanics & ledger state
• Activation logic

Auxilliary / Related work in our development of the Crosslink Feature Net is decoupled from Crosslink proper, and these auxilliary ZIPs could be adopted or rejected independently of the Crosslink design (although rejection would require re-engineering the solutions they provide in the Shielded Labs implementation):

• A new peer transport protocol (aka "Shielded Transport Pigeon")
• A new PoW block syncing algorithm

[shielded-nate]

cc @daira , or other write-access accounts:

• This isn't worthy of a thorough editorial review due to known flaws / TODOs. OTOH, if you have high level early feedback, that can be helpful.
• I thought you may also be able to tickle CI just to catch any automated stuff.

Stay tuned. (I will be more explicit when I request dedicated editorial review effort in the coming weeks.)

[shielded-nate]

Merged "sub-PR" shielded-nate#7 after a superficial group review/discussion and next steps planning. This sub-PR adds a "Crosslink ledger state" ZIP, and the current text is just a hand-wavy copilot first-draft.

The intended scope yet to do includes ledger state extensions to Zcash Mainnet, including:

• A definition / design principle of "ledger state" as an append-only mapping of { pow_hash -> ledger_state },
• The specification of objectively-verifiable finality,
• New verification state components: the roster and bonds,
• Abstract roster modification actions and rules aka "PoS actions",
• Concrete PoS action specification as transaction format and semantic changes,
• Issuance distribution changes for PoS incentive design.

Related: We opened a Crosslink-specific ZIPs issue tracker here where we'll track fine-grained / early maturity issues around these draft-PR Crosslink ZIPs to reduce clutter/noise in this issue tracker until a later more mature stage.

ZIP Editor Review request:

• No explicit effort requested yet.
• Helpful "light" review effort could be providing feedback on the overall arrangement / structure of the ZIPs, this PR development process, opening / commenting on issues in https://github.com/shielded-nate/zips/issues (use this freely if you would like, including for Q&A, suggestions, etc...).
• Not helpful / inefficient review: feedback on the actual text or concepts, since they are mushy LLM slop that needs to be refined / rewritten.