Week four was always going to be the hard one.
Weeks one through three: EVM. Different platforms — Base, Scroll, Core DAO — different security profiles, different deploy quirks. Same language. Solidity, all the way down. Week four is the first time the contract language changes. The question isn't "which non-EVM chain is active?" (most of them are, more or less). It's whether I can actually learn one in a week, starting from Rust-curious-but-not-fluent.
The field
The brief started with five candidates. Three more landed when I started digging: Fuel Network (Sway), Algorand (Python), Radix (Scrypto). Altogether: Rust-based languages, a Python chain, a Move chain, a Haskell-inspired language, and one chain where Rust contracts were recently abandoned.
Two eliminations worth naming
Polkadot / ink!: ink! development and maintenance ended in January 2026. Funding ran out; v5 is the last release. Polkadot is pivoting toward PolkaVM and JAM (Join-Accumulate Machine), but neither is production-ready yet. If a tutorial still lists ink! as a viable target, it's pointing at a dead end. Good to know before you spend a day on toolchain setup.
Cardano / Aiken: Lowest score in the comparison — 34.5/60. Not because Aiken is bad tooling. It isn't. The issue is eUTxO. There is no mapping(address => Message[]). Messages would be UTxOs carrying datums, locked at a script address, constructed entirely off-chain. A one-week build on Cardano isn't about shipping features — it's a week-long wrestling match with a new mental model. Wrong scope for a sprint.
The scored finalists
| Chain | Language | Score | Verdict |
|---|---|---|---|
| Solana | Rust + Anchor | 50/60 | Deepest community, best frontend story, account model is the cliff |
| NEAR Protocol | Rust or TypeScript | 49/60 | JS contracts make it approachable; wallet deprecation is a gotcha |
| Algorand | Python (Puya/AVM) | 46/60 | Most learnable language, wallet ecosystem is too thin |
| Aptos | Move | 46/60 | Polished tooling, resource model shift smaller than expected |
| Fuel Network | Sway | 39/60 | Interesting architecture, young ecosystem, one wallet option |
| Polkadot / ink! | Rust | 35/60 | ⚠ Maintenance ended Jan 2026 — not a safe pick |
| Radix | Scrypto | 35/60 | Foundation transition in 2026 adds uncertainty |
| Cardano | Aiken | 34.5/60 | eUTxO model is a week unto itself |
Why Solana, and why not NEAR
One point between them. NEAR had something real going for it: JS contracts. With near-sdk-js, a Solidity developer writes a NEAR contract in TypeScript and deploys same-day. No Rust. I almost picked it on that basis alone. One caveat worth flagging: near-sdk-js has historically been seen as less production-ready than the Rust SDK — higher gas costs, fewer security audits at scale. For a one-week experiment that's manageable. For production, it matters.
What actually moved the needle was community depth. Not the kind that shows up in a score — the kind that matters at 11pm when something breaks and you need an answer now. Solana has a Helius blog post for the exact error. A StackOverflow thread. Someone in the Anchor Discord hit this three weeks ago and documented the fix. NEAR is active, just smaller. That gap hits at the worst moment.
The other NEAR problem: wallet fragmentation. The original wallet (wallet.near.org) is deprecated. The replacement is scattered across Meteor, MyNEAR, HERE Wallet. You run into that friction during connect-wallet testing — not a good time.
The account model — PDAs, explicit account declarations, rent — is the steepest part. It exists for a reason: Solana's runtime (Sealevel) processes non-conflicting transactions in parallel. Unlike the EVM's single-threaded execution, Sealevel needs to know upfront what each transaction will touch — that's why every account must be declared explicitly. The friction is the feature. Learnable-hard, not paradigm-shift hard. And it's the most transferable thing you take out of the week.
The fallback
If Solana hit a wall mid-build — toolchain drift, account model slower than expected — NEAR with JS contracts was the escape hatch. I wrote that down before starting. One-week builds need a named fallback, not a vague "we'll figure it out."
→ The live app is at https://proof-of-support.pages.dev
→ Scoring methodology for the series: How I'm Scoring the Chains
Top comments (0)