Co-publication protocol — one design-partner comparison, warts included

Stage: after stage 5 (a signed partner). Its output feeds stage 1 for everyone else: the strongest evidence we can show a new lab is a real comparison on a real partner's rig. Buyer page it lands on: validation docket.

With the partner's written permission, we co-publish one comparison from a design partnership: their frames, their reference measurement, our reconstruction — including where they disagree and what remains unvalidated.

Rule zero — written permission, scoped

No partner name, logo, frame, figure or quote appears anywhere public without a signed or emailed permission that names exactly what may be published. The design-partnership agreement (page 22, item 8) already says public naming needs written permission; this page is how we ask for it.

Subject: Permission request — co-publishing the <lab> comparison

Dear <Name>,

As the engagement closes, we'd like to co-publish the comparison from your
frames on our validation docket. Proposed scope:
  - Named as: <"Lab X, University Y" | anonymised as "a university DHM lab">
  - Shown: <N> frames, their reconstructions, your <reference instrument>
    measurement, and the comparison table
  - Including: every point where the two disagree, and a section listing what
    this comparison does NOT validate
  - Not shown: <anything you exclude>
You will see the full draft and can change or withdraw anything before it goes
live. Reply "approved as scoped" (or edit the scope) to confirm.

Artifact: the permission email filed with the operator, scope verbatim.

What gets published (all of it, or nothing)

1. The setup: geometry, wavelength, pitch, distance, reference instrument.
2. The frames (or a stated subset) and each result ID + evidence record, so a reader can recompute.
3. The reference measurement, as the partner supplied it.
4. The comparison: the same quantities side by side, with the method of comparison stated before the numbers.
5. Disagreements — every case where our reconstruction and their reference differ beyond their stated reference uncertainty, with our best explanation or "unexplained".
6. What remains unvalidated — other geometries, other targets, other rigs, repeatability across days, anything this one comparison cannot speak to.
7. Credit lines and the citation of the service.

Numbers in this document are [MEASURED] for that one rig only. Sales copy may link to it; it may not lift a figure out of it into an email, a slide or a buyer page (claims rule, page 00).

Review steps

A partner may withdraw at any step before step 6. After publication, a correction request from the partner is applied within 5 working days.

Artifact when done: the live docket URL, with the permission and final approval emails filed.