Some Thoughts on Shelley and Cardano Features

Opens December 20th 2018 before Christmas to cover Shelley and the features expected in Q1 2019, starting with the just-released 1.4 client.

1.4 restores into a new data layer and compresses per-block files into an epoch package, an I/O-intensive process taking 30 minutes to hours on mechanical drives, though no blockchain re-download is needed.

Antivirus and firewall heuristic engines can quadruple or 10x the restoration time by analyzing the heavy I/O, advising users to temporarily disable antivirus during installation.

Storage shrinks only after a background purge over about a day, giving up to a 10x reduction depending on file system, and points users with issues to the forum and a coming dedicated Telegram channel.

Moves to Q1 and defines the Shelley release as three dimensions, testing, deployment, and social, built on new protocols for networking, Ouroboros Genesis, delegation, and incentive schemes tied to stake pools.

Prototypes exist for the network stack, Ouroboros BFT and Genesis, delegation, and the incentive scheme, backed by 50-plus pages of specifications, targeting broad-scale testing in the first half of Q1.

The work was de-risked by running two competing teams: a pragmatic Silicon-Valley-style Rust team moving fast and a larger specification-driven team moving more slowly.

Recalls a stake pool registration list in 2017 drew over 4000 interested operators, and plans to reopen it in January 2019 with the reinvigorated Cardano Foundation, aiming for close to 1000 pools on a testnet.

Migrating the current Byron system to Shelley, needing at least two client updates since Ouroboros Classic from 2017 differs greatly from the evolved Ouroboros Genesis.

Ouroboros BFT will be used as a halfway house forked in at the base level with a migration path to Genesis, BFT being the last code committed to the legacy repo and now in broad testing.

BFT is tested by running Rust clients against Haskell clients as two independent implementations verifying the spec, becoming version 1.5, followed by ledger, API, and network changes toward Shelley.

Reflects that 2018's troubles came from an undocumented, technical-debt-laden codebase whose original engineers had left IOHK, forcing a halfway-house cleanup across 1.1 to 1.4.

IOHK rewrote the wallet logic, refactored the core to be testable, and wrote two formal specs, one retroactive for the old system to validate legacy transactions and one for the new Shelley system.

The new Shelley system has no legacy concerns and provides a lingua franca of set theory and math, an implementation-agnostic spec any developer can read to reason about valid transactions, from extended UTXO to ledger rules to consensus.

1.5 with Ouroboros BFT will have a graceful upgrade path needing no restoration, followed by a decoupled wallet backend, Icarus-style addresses, and a gradual move to the v2 API.

Shelley's threshold is social too, needing enough people running stake pools, achieved by a gradual epoch-by-epoch winnowing of federated Ouroboros BFT control until 100 percent of slots come from stake pools, with Duncan Coutts explaining more.

The decoupled wallet backend lets them augment the wallet layer, adding manual transaction construction, offline signing, manual UTXO selection, and wallet scriptlets already possible in the Rust client.

A Q1 priority for the Daedalus team is a built-in terminal enabling an advanced user mode for manual transactions, offline signing, and running scripts, useful for power users, forensics, and exchanges.

Turns to the controversial delegation center design, noting that like an app store or search engine, first-page stake pool placement wins outsized attention, so the community must discuss fair display of delegation options.

Plans to socialize the design once SMASH is done in January via community groups like the Cardano Effect, acknowledging it moves from hardcore math into opinion and fairness.

With the new data layer and Icarus-style BIP-44-compliant addresses, Ledger hardware support is coming, with Emurgo engaging Vacuum Labs who analyzed the prior contractor's code and estimate five to six weeks of work from January.

Ledger support could reach Daedalus by February, Yoroi is adding treasury support, and users will transition from nondeterministic to Icarus addresses for faster half-size recoveries, interoperable between Yoroi and Daedalus.

Daedalus will get GUI screens for an offline center segregating hot and cold wallets, since it must now handle paper wallets and Ledger devices alongside monolithic hot HD wallets.

Cold staking lets a Ledger or paper wallet control stake for delegation and eventual voting on Cardano improvement proposals or treasury ballots without exposing private keys to a hot wallet.

Daedalus uses React Polymorph on Chromium and Node via Electron, a traditional web model enabling rapid prototyping that converges GUI designs in days to weeks.

Summarizes Shelley's three components again, testing, upgrade, and social, ending in a gradual epoch-by-epoch reduction of BFT nodes until the system wakes up fully decentralized.

Praises the new third-party Ledger partner and Yoroi's rapid evolution, teasing announcements from Sebastien, then pivots to usability.

Addresses usability, saying users entered Cardano expecting Daedalus to be a polished consumer product like Windows 10, when the 2017 release was very early beta software full of bugs and growing pains.

Reports 2018 improvements: the network stack 3 to 4x faster, Daedalus memory down 75 percent, disk from gigabytes to megabytes, better UTXO selection without fragmentation, and fast Icarus-style recoveries.

IOHK built a Chrome-based reference wallet handed to Infinito and Emurgo enabling hundreds of thousands of users, calling it a beta product with no catastrophic fund-loss or network-down bugs, on Windows, Mac, and now Linux.

2019 warrants a dedicated product experience officer to work backward from user experience to improve startup time, UI, UX, and multisig, a role heading into the HR queue.

Not everyone needs a full node, so 2019 will bring light and mobile client segregation once a resilient full-node backbone exists for the peer-to-peer switch.

Stresses that when Shelley ships Cardano will be fully decentralized, 50 to 100 times more than Ethereum, Bitcoin, and EOS, needing the right social structures now that the Foundation is back investing.

Closes with community development plans: a Cardano Ambassador Program, a hub program, the growing Cardano Effect, and hopes for Korean and Japanese copycat podcasts, wishing a Merry Christmas and Happy New Year.