Skip to content

Full per-file provenance/attribution audit for the 56-language corpus #4

Description

@squid-protocol

What's needed

SOURCES.md (added alongside this issue) records project-level attribution for the 10
paradigms explicitly named in README.md's "The Paradigms" section, but that's a first pass,
not a complete audit:

  • No exact commit or file path is pinned per paradigm, unlike gitgalaxy-raw-output's
    corpus/v1/manifest.json, which does this for its own corpus.
  • The other 44 language/format directories under data/ aren't itemized in SOURCES.md at
    all yet, even though many of them are referenced by name elsewhere (Apollo-11's AGC flight
    software, curl, and dozens more mentioned in README.md and in
    squid-protocol/gitgalaxy's tests/README.md).
  • The Perl side of "The Legacy Regex Trap" paradigm has no recorded source project.

Proposed fix

Walk data/, identify the upstream project (and ideally exact commit) each file or directory
was pulled from, and add a row to SOURCES.md for each. Where the original source can't be
determined with confidence, say so explicitly rather than guessing -- an honest "unknown"
is better than an asserted-but-wrong attribution.

Context

Follow-up from a README pass adding a "Contributing" section that asks new submissions to
include source + license links (see README.md and .github/PULL_REQUEST_TEMPLATE.md) --
the existing corpus should hold itself to the same standard being asked of new contributors.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions