Part 2: Engineering for money
Part 1 explained how value moves on a blockchain. This part explains why moving it reliably is hard, and how the engineering world solves each difficulty. None of these ideas is specific to crypto: they are how banks, payment companies and databases have always kept money correct. On a blockchain they matter even more, because nothing can be undone.
| # | Lesson | The problem it solves |
|---|---|---|
| 10 | When networks fail | A request timed out. Did the payment happen or not? |
| 11 | Idempotency | How do you retry safely, when retrying might pay twice? |
| 12 | Persistence and crash recovery | The process died halfway through a payment. Now what? |
| 13 | Concurrency: locks, leases and fencing | Ten workers share the work. How do they not collide? |
| 14 | Trust: one server's word is not proof | A server says the payment failed. Is it telling the truth? |
| 15 | Secrets and key custody | Where do private keys live, and how do secrets leak? |
Each lesson ends by showing how crypto-aio applies the idea. After lesson 15, the Developer tour puts them all together, inside the library.