| fuz_schema | correctness | speed | Checking every key of a strict object, where counting the keys was faster | A share of the time of every strict parse | declined |
|---|
| fuz_code | correctness | speed | A correctness gain is taken when it costs a few percent of lexing time | A few percent of lexing time, weighed against the gain only when an input would turn pathological | guidelines |
|---|
| pglet | correctness | speed | Float text written digit for digit as PostgreSQL writes it | A measured slowdown against the faster formatting | tradeoffs |
|---|
| fuz_forge | correctness | simplicity | A TypeScript server kept as the reference twin of the Rust one, held to it by shared tests | Two implementations to keep in step | overview |
|---|
| fuz_app | security | speed | A login for an unknown account still verifies a dummy password hash, so its timing matches a wrong password | A password hash verified on every login for an unknown account | security |
|---|
| zap | security | speed | Security controls are read from the live machine on every run, with no cache | Every plan reads them again instead of trusting a stored verdict | security |
|---|
| fuz_schema | worst case | size | An issue path that stays linear on deep input | The bytes a smaller path-copying form would have saved | declined |
|---|
| gro | speed | security | One realm, mutable intrinsics and the JIT on by default | Some hardening, which an opt-in tier listed as future work would give back | runtime design |
|---|
| tsv | speed | size | Profile-guided builds for the shipped CLI and native addon | Larger binaries | performance |
|---|
| blake3 | size | speed | A second build without SIMD, about 30% smaller, for where bundle size matters more | In that build, the two to three times SIMD gains on large inputs | install |
|---|
| fuz_schema | speed | memory | A lowercase conversion that always copies, where skipping the copy allocated less | An allocation per parse, since allocation only breaks ties | declined |
|---|
| fuz_schema | memory | speed | No lookup table held by each object schema, where one made some checks faster | The faster lookup on input out of the schema order | declined |
|---|
| tsv | speed | simplicity | A cleanup of a hot path is measured like any change, and refused when it costs time | Code that stays less tidy than it could be | performance |
|---|