Skip to content

docs(feasibility): standing feasibility assessment + seam-class/evidence-tier regularity matrix - #130

Merged
arena-ai-coding-agent[bot] merged 1 commit into
mainfrom
arena/01a0d84a-iseriser
Sep 25, 2026
Merged

arena-ai-coding-agent[bot] merged 1 commit into
mainfrom
arena/01a0d84a-iseriser

Conversation

@arena-ai-coding-agent

Copy link
Copy Markdown
Contributor

What

A recon-level feasibility assessment of the -iser family, written as a standing policy document rather than a one-off review:

  • docs/FEASIBILITY.adoc — the assessment itself: what the project is trying to achieve (product / thesis / estate, assessed separately), the overall verdict, the seam classification, evidence tiers, the language bar, per-language verdicts, and the recommended shape.
  • docs/feasibility-matrix.toml — machine-readable form: all 29 -isers classified by seam class (6 injector / 14 service / 9 scaffolder), evidence tier (T0–T3), and language level (L0–L3, estate languages), each with a grounded note.
  • README.adoc / docs/ATLAS.adoc — pointers to the assessment, and a re-scoped statement of the family promise.

Why

The headline promise ("superpowers … without you ever learning those languages", a "formally verified" ABI for 28 runtimes, aspects that "compose … without knowing about each other") is infeasible as stated for structural reasons — semantic guarantees do not cross C ABIs, manifests conserve rather than abolish the learning curve, runtimes each assume they own the process, and a 28-language proof programme is CompCert-scale per language.

The re-scoped programme — verified binding generators + foreign-runtime services + pattern scaffolders, each claiming only what crosses its seam — is feasible, has deep prior art, and is already partly demonstrated in-estate: chapeliser's Provable lane is green on main (Idris2 proofs check, Zig FFI tests, golden codegen, generated Chapel compiled and run by chpl in CI), and abi-emit-manifest/abi-verify is a genuinely reusable drift gate.

The regularity introduced (the point of the PR)

  • Seam classification — injector / service / scaffolder; an -iser may claim only what its seam carries. Class is a property of the semantics, not of ambition.
  • Evidence tiers T0–T3 — claims may not outrun what CI machine-checks; chapeliser's Provable lane is the family template.
  • Language bar L0–L3 — an -iser over an estate language below L2 (ABI-stable) is labelled experimental; the -iser's spec-forcing role is named, not hidden.
  • Claim hygiene — precise wording rules for "formally verified", "without learning", "provably safe ethical constraints", etc.
  • One-flagship concentration — chapeliser + futharkiser + one estate-language flagship to Tier 3; the rest to maintenance.

Notable per-language verdicts (focused set)

-iser Verdict One-line core issue
a2mliser feasible, HIGH canonicalisation of TOML/YAML is the hard core; key distribution is the product problem
k9iser feasible, HIGH needs a real grammar (build on Nickel?); mined contracts must be schemas/ranges, not frozen literals
eclexiaiser feasible, MEDIUM typed units cross the seam; runtime budgets can only guard/report, never prove
chapeliser feasible, HIGH flagship; green end-to-end Provable lane; ≥2-node multilocale is the frontier
futharkiser feasible, HIGH easiest true T2→T3 win; needs a CI lane that installs the futhark compiler
Idris2-as-ABI feasible if right-scoped invest in seam-contract authority + drift gate; cross-runtime marshalling proofs are infeasible

Notes for reviewers

  • No code changes; docs + one TOML matrix only.
  • The matrix is deliberately opinionated — it is meant to be argued with, then maintained as the standing record.
  • Evidence used: iseriser HEAD 6397682, both satellites, sibling -iser repos (sizes/CI/stars), chapeliser's green Provable lane, and the language repos (eclexia, a2ml, phronesis, …).

…vidence-tier regularity matrix

Recon-level feasibility assessment of the -iser family (docs/FEASIBILITY.adoc)
with a machine-readable per-iser classification (docs/feasibility-matrix.toml):

- seam classes (injector / service / scaffolder) — an -iser may claim only
  what crosses its C-ABI seam
- evidence tiers T0-T3 (chapeliser's Provable lane as the family template)
- language bar L0-L3 for the estate languages (a2ml, k9, eclexia, ...)
- per-language verdicts for the focused set: a2mliser, k9iser, eclexiaiser,
  chapeliser, futharkiser, and the Idris2-as-ABI bet in both its scoping
  options
- README/ATLAS pointers to the assessment

Co-authored-by: arena-agent <297053741+arena-agent@users.noreply.github.com>
@coderabbitai

coderabbitai Bot commented Sep 25, 2026

Copy link
Copy Markdown
Contributor

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Advanced

Run ID: 98db81c7-b3ca-492e-a3e3-b16a65e9c9f2

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@arena-ai-coding-agent
arena-ai-coding-agent Bot merged commit b007d33 into main Sep 25, 2026
5 checks passed
@arena-ai-coding-agent
arena-ai-coding-agent Bot deleted the arena/01a0d84a-iseriser branch September 25, 2026 17:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant