Skip to content

Docs/fleet v0.6.0 - #43

Merged
timothy-weston merged 5 commits into
mainfrom
docs/fleet-v0.6.0
Sep 11, 2026
Merged

timothy-weston merged 5 commits into
mainfrom
docs/fleet-v0.6.0

Conversation

@timothy-weston

Copy link
Copy Markdown
Contributor

No description provided.

@timothy-weston
timothy-weston marked this pull request as ready for review September 10, 2026 09:44
timothy-weston added a commit that referenced this pull request Sep 11, 2026
…t reconcile

The first classification named Quick Dispatch and was too narrow; writing down that it
was wrong, and on what evidence, is what stops the next person re-deriving the narrow
version. Also names the three absolute Send-to-Robot statements on the v0.6.0 branch
that need the same qualification once it lands.
timothy-weston added a commit that referenced this pull request Sep 11, 2026
Stacking the limitations layer on the v0.6.0 docs so the overlap between them is
resolved once, deliberately, instead of at merge time. #43 owns the supported
workflow and had already absorbed most of the corrections this branch catalogued;
what it superseded was withdrawn here, and its two remaining absolute claims about
Send to Robot were qualified.
Documents what v0.6.0 changed in how Fleet is operated: who does what, the
two standard workflows, that activation is not residency, the transient
dispatch contract and its two entry points, the refusals Fleet answers with,
and how a failed run is closed. Stacked on the pre-v0.6.0 cleanup, whose
corrected badge vocabulary and run-condition matrix this builds on rather
than restates.
First audit pass over the new material. The page claimed behaviour the
interface does not offer — a retry door removed before release, a pause
exempt from the E-Stop, a configurable action on losing the link, pause-owner
reasons never served to a client, and a location correction reaching a robot
that already holds the mission — attributed the run conditions and the E-Stop
release to the wrong roles, and called the product Fleet, which is our own
word rather than the one customers see.
9587dae expanded the index's feature teasers into capability documentation, and
every audit since went to the sub-pages instead — so the E-Stop authority rule
was stated three times on the index and nowhere on the page documenting the
button, the lease definition and the detection-type counts were maintained in
two places, and two claims survived only on the index. Each teaser now gives the
page's purpose, the reason to open it and the link; the facts sit with the
feature they belong to.
Second pass, on three things an operator would have found contradicted in
front of them: a moved saved location raises the per-mission "location
changed" badge rather than the low-opacity "needs review" note about the map
shifting under a mission, robot:estop is a site-scoped permission that the
tenant Auditor does not hold, and the resume control is disabled while the
E-Stop is active rather than always yours to press.
…-056 rules

ADR-056 makes an explicit Send the only thing that establishes residency, and
the page asserted that as fact. Production does not hold the line: two publish
paths compose from the armed set without narrowing to what the robot already
holds, so an Active mission that was never sent can arrive as a side effect,
and a dispatch re-sends held missions from their current saved state rather
than leaving them untouched. Send to Robot is now described as the deliberate
workflow rather than the only possibility, and a scheduled run is named as the
resident work it is; the customer-facing exception belongs to the
current-limitations branch.
@timothy-weston
timothy-weston merged commit dd799ce into main Sep 11, 2026
2 checks passed
@timothy-weston
timothy-weston deleted the docs/fleet-v0.6.0 branch September 11, 2026 10:31
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