Skip to main content
  • Status: accepted · Date: 2026-07-17 · Deciders: Davor Runje
Refined by ADR-0023. Engineering is delegated via the engineering-delegation contract to a bound engineering backend, naming no specific tool. The original decision named one; that framing is superseded.

​
Context

The full research lifecycle includes both scientific work (claims, evidence, decisions, papers) and engineering (designing/building the code that produces evidence). Mature engineering workflows already handle the latter well (design → plan → implement).

​
Decision drivers

  • Don’t reinvent engineering workflows; keep honest-scholar’s identity sharp.
  • The scientific parts (rigor, positioning, decisions) are what’s missing.

​
Considered options

  1. Scientific scope only; delegate engineering via a contract to a bound engineering backend.
  2. One plugin covering both science and engineering.

​
Decision

Option 1. hypothesis-testing hands design/plan/implementation to the bound engineering backend via the engineering-delegation contract (resources/contracts/engineering.md) and stores the artifacts under the hypothesis folder.

​
Consequences

  • Clean separation; honest-scholar composes with an engineering backend rather than competing with one.
  • A dependency on some engineering backend for the engineering leg (acceptable; the consumer binds whatever it uses in .honest-scholar/config.yml).

​
Rejected alternatives

  • All-in-one plugin — duplicates mature engineering tools, bloats scope, blurs identity.
meta-spec §1 (non-goals), §2; sub-spec 1 §2.2; ADR-0023 (the engineering-delegation contract that refines this).