UPD 05/21/2019

A quick update on the upcoming Shelley testnets and how the rollout will work.

Two code bases are converging on Shelley, the Rust client and the Haskell client, unified by a shared collection of formal specifications.

The Haskell team works the slow scientific way, spec then reference code then refinements, and has now moved to demo driven development, while the Rust team runs a classic agile sprint backlog.

The Rust client is almost ready for a testnet release, and version one is a self node setup: a YAML generator parameterizes your own Shelley, bootstraps your own Genesis block and runs a multi node blockchain in a box with one node or a hundred.

Code written against the self node carries over unchanged, which is why the Haskell Cardano wallet backend is being implemented against the Rust node before the network stack lands.

The Cardano stake pool best practices Telegram group has about 1290 members, and surveys are going out to gauge technical acumen and willingness to run test scenarios through June.

Version one proves the configurations work across different hardware and operating systems, then level two moves everyone onto a single unified public testnet with gossiping and block downloads from real peers.

The third milestone is the incentive system with stake rewards paid in test ada, and alongside it run the Kudelski security audit and integration of the Haskell wallet backend with the Rust client and the Daedalus GUI.

The Haskell demos build cumulatively toward their own testnet around July, and a decision will be needed on how to close the delta with the Rust client, whose network protocol design took a year but is very elegant.

Emurgo will help with testing and Yoroi interoperability for the staking components, plus a blockchain explorer to visualize what is unique in the ecosystem.

Vincent and Tim will write a blog post on what makes the Rust client special, including a crypto core abstracted for trusted hardware like SGX, while the Haskell side shines on upgradeability after rewriting Cardano three times.

There are two multi-sig designs, a Plutus contract and native protocol level support already implemented on the Rust side, and multi-sig is a prerequisite for the Shelley launch.

Ledger Nano X support is coming with Vacuum Labs, ideally multi-sig on hardware devices for enterprise fund management, and Jormungandr has new C bindings built for exchanges.

June brings the self node plus videos on the command line interface, building transactions by hand and reporting bugs on GitHub, part of moving to an open source project instead of a top down company product.

July more likely than not brings network enhancements and the jump from a self hosted node to a single unified testnet, followed by the candidate reward scheme designed collaboratively with community feedback.

Join the Cardano stake pool work group on Telegram, fill out the survey, and a focus group will get DevOps tutorials on building from source and running configurations that try to break the software.

A later update stops the node being self parameterized and connects it to the real testnet with faucets, so you can self host or delegate to a stake pool, and then comes candidate one of the reward scheme plus hardening from the audit findings.

The self node update ships in late May or early June 2019, the Haskell wallet backend connects to the Rust node around June 7, and a Haskell testnet is expected late June or early July 2019 before deciding how aggressively to close the delta.

Modularizing the Haskell wallet backend so it could talk to the Rust code produced a universal UTXO wallet that could be rebranded for a Bitcoin or Dash node, stateless, concise and with a much lower line count than competitors.

Mainnet gets one minor update to the OBFT deployment in June and a hard fork date, then a choice between going straight to Shelley or first replacing the hostile old cardano-sl code with a complete Haskell rewrite of Byron.

Cardano Docs is being completely redone with technical writers also documenting the Rust client, and the developer channel and stake pool task force will be the first readers giving feedback.

Plutus is evolving fast with hard work on Haskell to JavaScript and Haskell to WebAssembly, the language is close to final, papers are going to peer review and a Marlowe hackathon in Mongolia drew 66 people.

Two Udemy courses on Marlowe and Plutus have more than 2000 students each, a Plutus book comes in July, and an O'Reilly book may follow in Q4 or Q1.

Several of the 23 women trained in Ethiopia have been hired as Plutus beta testers and QA, feeding into a push on developer experience that reuses decades of Haskell tooling and best practices going back to the 1980s.

Manuel Chakravarty leads a very strong Plutus team, and although Plutus could be integrated at any time it stays a post Shelley activity because decentralization comes before smart contracts.

Nothing stops you writing dApps today: play with Plutus Playground and Meadow, use the tutorials, examples and courses, and every Haskell course from now on adds a month of Plutus training.

Aiming for more regular communication because the testnets will move quickly, with a whirlwind of messaging, telegram and email before things are ready for mainnet.