2023-07-29
Intersect, Repos, and Cardano Product Backlog
- A July 2023 whiteboard explaining how the Cardano codebase transitions from the Input Output repos to Intersect, the community members-based organization, over the following year.
- Details the mechanics: IO's tribes become community working groups feeding one Cardano product backlog, run on a quarterly PI system, with on-chain credentialing courses replacing gatekept access.
- Explains the split between CIP 1694 as the government that grants permission and Intersect as the execution vehicle, and how implementation-agnostic Agda specs enable polyglot alternative clients.
- Frames it as the hardest and most significant thing in the project's history, a self-healing decentralized structure meant to last centuries, contrasted with a US system that cannot fire its own leaders.
42 entries
Opens July 29th 2023 with a whiteboard on Intersect and how the codebase will transition from the Input Output repos to it.
Intersect is the name given to the Cardano members-based organization, an aggregation hub the community can rename, owned and run by its members.
Calls the roughly 636 GitHub repositories the heart of Cardano, from cardano-node and Daedalus to Plutus and the CLI, a staging ground for 15 to 16 companies to build the protocol together.
The IO repos will move to Intersect over the summer with a new governance structure, and that IO organizes work into tribes each covering a concept like smart contracts, core technology, or identity.
All the tribes feed one Cardano product backlog, called GCD for Getting Cardano Done plus continuity work, the daily better-faster-cheaper improvements.
Working groups at Intersect will take over the tribes' function, with IO tribe leaders joining and community groups from NFTs, DEXes, and Cardano DeFi joining, and the backlog moving over as a suggestion.
The quarterly PI system with pre-commitments, post-mortems, and a two-to-eight-week release cadence, and how credentialing lets the working groups fully decentralize the product function.
Distinguishes what CIP 1694 does, acting as the government of Cardano that grants permission for hard forks and authorizes CIPs, from Intersect, which is the execution vehicle that gets things done.
An interplay between the government and the working groups on a consolidated backlog, giving democratic consent to ratify a CIP and initiate a hard or soft fork.
It is easy for community members to join the not-for-profit MBO, with over 200 already joined and the organization growing.
Floats an idea to attract open-source contributors: four courses granting on-chain blockchain certificates, covering Haskell, Agda, open-source best practices, and a Mastering Cardano course.
The courses establish a technical baseline in extended UTXO, Plutus, and the wallet APIs so anyone can prove core-dev skills, self-certifying with no gatekeeping and a blockchain certificate at the end.
Warns that credentials are usually a gatekeeping function, who controls the spice controls the universe, so the goal is open content anyone can complete, with existing contributors grandfathered in by the working groups.
Sums up the structure as a consent engine where every ADA holder has a say, plus an execution engine anyone can join, knowledge transfer, and a hard-fought operating model doing reliable two-to-eight-week releases with public PI accountability.
Says it builds on the existing CIP process and universal backlog, learns from Hydra as a successful open-source program, and adds blockchain-based certificates.
Raises the open question of alternative clients like Rust or TypeScript and asks what the canonical Cardano actually is.
The major innovation that the CIP 1694 and Conway specifications were written in Agda rather than prose, making them implementation-agnostic blueprints you can do code extraction, testing, and certification from.
A polyglot future of many alternative clients conforming to standards so each has the same quality assurance as the canonical Haskell client, preserving the hard-fork combinator so government-approved upgrades actually happen.
Lays out the 2024 GCD agenda of end-to-end formal-spec coverage of all prior work into one set of universal blueprints, plus a certification pipeline so third-party clients certify against it.
It is also a chance to simplify the blueprints, for example retiring some Byron logic to reduce complexity and speed the Rust and TypeScript clients to market with no user consequences.
Stresses Intersect is built and run by the community, with IO seeding buckets and institutions with ideas and capital to bootstrap, but the community owning it as a custodian of knowledge.
The structure gives the government options, buckets to fund and execution vehicles that run for a transition period before the government votes to continue or pivot, all while preserving the 2,100-plus-day uptime.
Names the risk that bureaucracy could slow things down after years of speeding up quarter by quarter, and insists on many voices with legitimacy through a vote rather than trust-me-bro.
Defines minimum viable governance: run it for a period of six months to a year, then ask if Cardano is getting more decentralized, keeping its quality and security, including more voices, and gaining users and utility.
Says if the government cannot answer those questions well, you move from MVG version one to version two, adding checks and balances in a spiral of more inclusion, wisdom, and execution capability.
The structure handles primary and secondary projects, with Cardano at the top and Hydra, Prism, Plutus, and Marlowe beneath, just like Linux and the Hyperledger Foundation with Fabric, Besu, Indy, Aries, and Firefly.
The project structure gives DApp developers a place to plug their repos into a broader collaborative structure like Hyperledger, with Intersect mastering the management of these concerns.
As governance grows it can fund essential community projects from the treasury, building a self-sustaining, self-reinforcing decentralized ecosystem shaped by six years of talking to Apache and the Linux Foundation.
IO Engineering is a subsidiary of Input Output Group doing specialized Cardano engineering, alongside firms like Well-Typed, Duncan Coutts's company, plus TxPipe and Galois, all sharing one backlog.
Community groups like MLabs and DCSpark can join for what they care about, such as side chains rather than core technology, choosing their level of participation while working from implementation-agnostic specs.
Claims this will be the most decentralized and inclusive development process in the industry by the EDI, since no one is shut out for being disliked or a competitor, with self-certifying credentials and one visible backlog.
The CIP process is inclusive and managed by the Cardano Foundation, and that adding firms is easy so the treasury can fund specialized firms to accelerate the roadmap.
Lists three 2023 events as checkpoints, Rare Evo, CNFT Con in Las Vegas, and the Foundation's Summit in Dubai, and says MBO members will eventually pay fees across a free, paid, and merit tier.
A merit tier was a point of discontent with the Linux Foundation, arguing people doing great work should not pay for committee seats, so the government can fund alternative MBOs for engineering versus adoption and marketing.
Credentials do not require real-life identity: a rockstar developer can use a DID, a credential, and an alias, verifiable without geography, language, or identity, so privacy is never compromised.
Says if you disagree with all of it, that is the point of CIP 1694: propose an alternate vision, and if the government votes for it, the Apache 2 open-source code can be forked and funded by the treasury.
Asks who the product manager of Cardano is and answers that it is every ADA holder, calling this the most significant event in the project's history, one that would let durability be measured in centuries.
Contrasts it with a US system that cannot fire an unpopular president or 90-year-old senators because the operating model was built by a small club creating the fantasy of choice, framing his life's work as a system that heals itself.
Events like CNFT Con, where 3,000 people showed up, are public-accountability checkpoints, and that once work moves over it goes from a done deal to a suggestion that must be ratified and then executed by servant leaders.
Argues the collective skill of millions exceeds any Microsoft, Google, Apple, or Amazon, so it is not if Cardano is number one but when, and he is excited for the most open development process in crypto history.
Frames Intersect as an aggregation point to bring in good ideas from Polygon, Solana, Polkadot, Algorand, and Ethereum, to deliver economic identity to a billion people and a governance and financial operating system for nations without one.
Closes on the humility to accept an idea regardless of its origin, with checks and balances against groupthink and incumbency like the entrenched US power structure, and thanks everyone for listening.