---
tags: arthur, tezos

title: Arthur talk notes
---

# Arthur talk notes

###### tags: `tezos`

## Tezos introduction
- Security First
- Functional Programming mindset
- Better allows for formal verification
- Immutability and pure functions help avoid bugs
- Proof of Stake
- Token ownership disincentivizes malicious baking
- On-chain governance
- Fast rollout of protocol features
- Reduces risks of contentious forks

## Structure of a Dapp
- "**D**ecentralized **App**"
- Fuzzy definition
- App based on a public blockchain
- User sovereignty
- Limited or no "admin" write access to important state
- Architecture outline
- UI: usually runs in browser
- Middleware (optional)
- Blockchain as Single Source of Truth
- On-chain smart contract(s)

### Smart Contracts
- What is a smart contract
- Technological enforcement of business rules
- Vending machine analogy
- Can remove soda, but only as much as you paid for
- Can remove coins, but only as much as you got in change
- On a public blockchain:
- Decentralized
- Autonomous
- Complexity restrictions
- Due to distributed consensus
- Similar power to a '70s microcomputer
- O(1) to low O(log(n)) complexity requirement with regards to state vector size
- Cannot be updated (immutable)

### Smart Contracts on Tezos
- Michelson language
- Native language of Tezos' smart contract VM
- Not ideal for humans to read although technically human-readable.
- Human readable, assembler-like syntax
- Stack based
- Strongly typed, with ML-inspired type system
- Functional Programming features (lambdas, currying)
- High level languages
- Several smart contract languages are being developed, the best will be adopted and will survive
- SmartPy
- Python library with optional lexer for syntactic sugar
- Metaprogramming: Python code generates Michelson code
- Python is familiar to many programmers
- Several convenience features
- Unit testing framework built in
- Lazily loaded code and other built-in optimizations
- Criticism: Python is a suboptimal fit for Tezos
- Procedural vs. functional mindset
- Weak typing vs. strong typing mindset
- Ligo
- Strongly typed functional language, based on ML
- Compiled
- Multiple syntaxes for familiarity, above the same language backend
- CameLigo based on OCaml
- JsLigo based on Javascript
- PascaLigo, ReasonLigo
- Culturally a good fit for Tezos, rooted in OCaml culture (especially CameLigo)
- Clear mapping between high level code and compiler output
- Criticism: Fewer convenience features than SmartPy
- Unit testing needs to be done by hand / scripted
- The Ligo compiler provides a `run-fuction` primitive
- Unfriendly compiler error messages

### Middleware / Indexer
- Provides speed and convenience for the UI
- Blockchain is the Single Source of Truth
- Blockchain queries are slow
- Blockchain queries are inconvenient and restricted
- Big, generalist, public indexers
- They keep track of the entire chain state
- They have knowledge of the common token types and can display much information about them
- Main examples
- Better Call Dev
- TzKT
- DipDup is a layer on top of TzKT which provides value-added services
- Centrally run instances
- https://better-call.dev/ and https://tzkt.io/
- Limited resources and high user demand
- Historical reliability issues under load
- Docker images are available to run in-house
- Indexing the whole chain is resource intensive
- Application-specific middleware
- DApps may have their own custom middleware
- Developed as a traditional server-side backend
- Usually contains domain-specific logic
- Uses the blockchain as a data access layer
- Focused on the DApp's own smart contract(s)

### Front-end
- Taquito
- Javascript library for interacting with Tezos node
- Provides smart contract abstraction
- Written in typescript and major release version same as tezos node
- Wallets/Beacon
- Connect to Tezos nodes
- Key management
- Locally stored keypairs or HD wallet
- Hardware wallet integration (recommended)
- OAuth-derived private keys (easy onboarding)

### PyTezos
- Python library for interacting with Tezos nodes
- Makes contract interaction easy
- One can mint an NFT in 5 lines of code (example to come)
- Potential uses
- Power user interface through Python REPL
- Accessing on-chain contract state
- Generating, signing and submitting chain operations
- Automated integration testing of smart contracts
- Blockchain connectivity for DApp-specific custom middleware

### Flextesa
- Locally run blockchain for developer testing
- Docker images are available for quick deployment