Grant Application Ideas

[ehildenb]

We can apply for grants to do Rust development to several organizations:

• Web3 Grants Program: https://github.com/w3f/Grants-Program. Better suited for projects where a PoC doesn't exist.
• Substrate Builders Program: https://substrate.io/ecosystem/substrate-builders-program/. PoC does exist, time to build production ready tool.

Some ideas for grants:

• MIRI track (~ 2-3 months, not similar to any other work we've done, maybe something better suited for another firm):
  • Update MIRI to handle Rust compiled to Wasm.
  • Test on several real Polkadot smart contracts.
  • Run against entire corpus of deployed contracts on Polkadot.
• KWasm track (~ 3-6 months, similar to Foundry work):
  • Port web interface of summarizer to KWasm (internally, also port summarizer algorithm, not deliverable of grant to keep it closed).
  • Make property-testing tool in Rust (perhaps using some existing format for property testing), similar to Foundry.
  • Enable fuzzing over Rust property tests using LLVM backend for KWasm.
  • Enable verification of Rust property tests using Haskell backend for KWasm.
• KMIR track (~ 4-8 months, very similar to KPlutus engagement)
  • Develop the KMIR semantics.
  • Verify a bunch of small Rust algorithms in KMIR semantics to battle-test it.
  • Verify (by hand) some simple Rust smart contracts (ERC20 or equivalent).
  • Make property-testing tool in Rust (perhaps using some existing format for property testing), similar to Foundry.
  • Enable fuzzing over Rust property tests using LLVM backend for KMIR.
  • Enable verification of Rust property tests using Haskell backend for KMIR.
• Compiler validation track (~ 4-8 months, mostly research type engagement):
  • Define bisimulation relation between KMIR and KWasm.
  • Develop tooling for verifying bisimulation relation of programs going through the Rust compiler.

[yanliu18]

Here are a few more ideas coming from recent auditing experience:

• Similar to MIRI track, but extending it to include some other tools such as Kani model checker for developers to develop safer smart contracts/ chain applications:

  • Extend these tools to WASM
  • Compare these tools regarding soundness, performance etc.
  • Build a benchmark for testing future tools.
  • Create tutorials for using these tools
  • Could be an intern project.

• High-precision financial computation: The motivation is to provide best practices for financial product programs

  • To study the common coding patterns regarding crypto computation, in particular AMMs, Lending protocols etc.
  • Compare the precision and rounding error between these patterns with regard to precision, rounding error, performance
  • Recommend the best practices in Substrate and Rust space
  • Providing toolings to plot a comparison chat with regard to the chosen implementation vs. the best practice.

• Differential testing: For a young ecosystem, its infrastructure is not stable. The constant changes could break existing applications and invalidates those already audited systems. Differential fuzzing could be a remedy for infra upgrades.

  • Could possibly be a cargo tool that compliments the existing regression tests.
  • Extend the existing protest framework/ cargo-fuzz with differential fuzzing.
  • I am not sure though, how much benefit developers would get with this tooling support providing this project could be quite a big effort.