This page is for anyone seasoned or a beginner who is deciding whether to bring their time here.
Legible designs and builds recomputable instruments and the applications around them, where the record is visible and declared. That can be an extension to Habitat or to the Reader, a new lens, a library, or a component built for one client's problem.
Everything made here is published and cited. Every engagement is independent, runs for a stated term, and the residual is sized to it. What counts as a citable object, and where a project's term ends, are worked out with you rather than handed to you.
Habitat reads a body of text or records as acts, observes the geometry they make, and reports what it observes. It trains on nothing and holds no opinion.
$ python3 verify.py
agent-trace v3_fsum PASS
hmda v3_fsum PASS
$ echo $?
0
All 8 cases ship a standalone verifier. numpy and the standard library, no network, between 228 and 433 lines. Measured across 4 builds, 2 architectures, 3 operating systems: one digest everywhere.
A client brings a body of work and a question about it. You design how it becomes legible, build it, and it publishes with its own checks printed beside it. The checks are the standard, so work is judged by what it shows and not by who made it.
Anything you make that clears the threshold: observable in the record, recomputable by a stranger, and citable as a published object.
Your name and your derivation ship with what you make. You take a stated share of the engagement you deliver, and a residual when what you made is used again. Legible sits inside a charity, so the return is compensation and not equity.
What you register is pinned to the version you registered under. The numbers will move. They bind what comes after and never touch what you already hold. The full sheet is at /residual/, and it prints.