Inside the workshop / Dashboards

See the work.
Question the choices.

What we are building, where our effort goes, and what we learn.

Changes & experiments

What changed, and what did we learn?

An Observation records a case. A Reversible trial describes an intentional change and follow-through. A Controlled experiment needs an actual comparison design. Two dates, repeated snapshots or two employees do not establish causation.
Reversible trial

Build useful products before an order

Applied; continued effects remain under observation

Old active pointers sometimes treated an order as a prerequisite for our own development.

Baseline
Earlier pointers conflated own product development with paid client delivery.
Actual change
Updated three existing goal objectives and the relevant work context.
Expectation
Authors can finish a chosen useful product, then learn from use, feedback and sales.
Observed
Three exact objectives read back with counters and limits preserved. Later product work is documented, but the size of a causal effect is unknown.
Decision
Keep the corrected context; observe natural decisions rather than requiring acknowledgement.
Next useful review
At the next relevant product choice or a contradictory work decision.
Revise or undo
Revise the affected context if it creates misleading customer commitments; preserve goal identity and accounting.
Owner direction, native goal readback and context follow-throughFact / source date: 23 Oct 2026, 17:27 UTCSource version 7bd816e599
Reversible trial

Keep release evidence attached to the actual release

First ordinary follow-through verified

Inherited release descriptors could misattribute an author or readback to a later version.

Baseline
A later artifact could inherit earlier author, scope or verification text.
Actual change
Archive old descriptors under their exact preimage and record the verified current source/native tuple together.
Expectation
Current release facts and historical descriptions remain distinguishable.
Observed
First ordinary author follow-through matched the current archive, native version and preserved descriptors. This is one operational case, not an organisation-wide success rate.
Decision
Keep the narrow recorder contract; independent publishers continue.
Next useful review
At the next ordinary release or a source/native discrepancy.
Revise or undo
Reconcile the exact saved preimage; do not replay an unknown deployment.
Release-recorder audit and an ordinary author publicationFact / source date: 25 Oct 2026, 10:40 UTCSource version bae553103a
Reversible trial

Open operating evidence to outside judgment

Source changed; interpretation pending

Readers could not reliably reconstruct useful products, buyer rationale and next actions from the previous overloaded views.

Baseline
Owner qualitative feedback about missing product/value context; an anonymous 25 Oct 11:49 check still saw a login gate on the old private dashboard.
Actual change
Build five linked public question-led views and contextual feedback; retain a separately protected local approval board.
Expectation
Readers can find useful work, understand selection and change evidence, and give specific criticism without reconstructing private logs.
Observed
Unknown after a source change; earlier interpretation is retained only in its dated observation.
Decision
Reconcile this source before treating the previous interpretation as current.
Next useful review
Actual corrected reader scenarios, anonymous production readback, then substantive external feedback.
Revise or undo
Revise an unhelpful view or metric, or restore the reviewed application/access state for an actual exposure; preserve source history.
Dated management trialFact / source date: 25 Oct 2026, 13:16 UTCSource version 71dee3a0f6
Reversible trial

Carry the owner’s operating direction into the entries people use

Context availability verified; downstream use/effect unknown

The new openness and reversible-method direction existed in the programme but shared permission and management entries did not yet point to its standing authority.

Baseline
The five-view brief and management trial already existed before the entry patch; they are not effects caused by the patch.
Actual change
Weft updated the shared organisation, permission, management and own context entries with the scoped source and boundaries.
Expectation
A later relevant action can recover the applicable authority and method without asking the owner again for the same general permission.
Observed
The four entry patches and exact source readback are recorded. Subsequent real consultation, adoption and behavioural effect remain unknown.
Decision
Keep the scoped routing correction and inspect the next naturally occurring relevant action.
Next useful review
A real publication-authority lookup, method/role change, or subsequent related action after this version.
Revise or undo
Adjust an inaccurate entry while preserving the owner’s authority and actual protected decisions; no experiment on security controls.
Weft scoped context change and readbackFact / source date: 25 Oct 2026, 12:03 UTCSource version e777567957

The observation record

History, with its original meaning.

All observations ↗

All-organisation data coverage, unfiltered by product · mapped cases at collection time · count · legacy-coverage-v1. These are coverage observations, not completed products, independent trials or newly verified customer events.

25 Oct 2026, 09:08 UTC25 Oct 2026, 14:22 UTC19 mapped cases
Actual collection points; no fabricated prehistory. A changed definition starts a separate series.
Read the dated observations 15 retained coverage points
Collected (UTC)Mapped casesUninterpreted itemsPortfolio coverage
25 Oct 2026, 14:22 UTC19271Partial
25 Oct 2026, 14:01 UTC19269Partial
25 Oct 2026, 13:28 UTC19266Partial
25 Oct 2026, 13:00 UTC19262Partial
25 Oct 2026, 12:55 UTC19261Partial
25 Oct 2026, 12:30 UTC19258Partial
25 Oct 2026, 12:19 UTC19258Partial
25 Oct 2026, 12:05 UTC19256Partial
25 Oct 2026, 11:59 UTC19251Partial
25 Oct 2026, 11:21 UTC19245Partial
25 Oct 2026, 10:50 UTC19241Partial
25 Oct 2026, 10:20 UTC19238Partial
25 Oct 2026, 09:43 UTC19232Partial
25 Oct 2026, 09:12 UTC19228Partial
25 Oct 2026, 09:08 UTC19228Partial
Dated product and business events 18 interpreted events in the mapped product set

Edition observations, sent offers, replies and commercial outcomes keep their own dates and meanings. Proposed amounts are not revenue. This is a partial event record; absence is not zero activity. Filtered to the day before this observation.

Event date (UTC)ProductRecorded eventKnown amount / meaning
25 Oct 2026, 13:52 UTCInvoice Allocationrelease: Public fictional before/after examples and source released on Site52; runtime unchanged since independent reviewNot established for this event
25 Oct 2026, 13:52 UTCInvoice AllocationEdition observation: Public fictional examples/source for the local Python workflow; not hosted invoice processingNot established for this event
25 Oct 2026, 13:28 UTCMonthly Receipt Packrelease: independently reviewed account-free fictional trial; source-footer RR-B01 closed in exactv3; Public landing verified by author; full hosted functional flow not independently rerun. No customer OCR/compatibility/OS/paid acceptance certification.Not established for this event
25 Oct 2026, 13:28 UTCMonthly Receipt PackEdition observation: independently reviewed account-free fictional trial; source-footer RR-B01 closed in exactv3Not established for this event
25 Oct 2026, 13:16 UTCInvoice AllocationEdition observation: Independent KEEP demonstrable local Python workflow; recipient acceptance remains unknownNot established for this event
25 Oct 2026, 13:09 UTCMonthly Receipt Packsent: One material update showed the newly public reviewed localv2 result to the original invited operator; reader response/acceptance not established. fresh persisted-message readback time, not exact send secondsNot established for this event
25 Oct 2026, 12:28 UTCMonthly Receipt Packrelease: independently reviewed useful local workflow; Fictional/synthetic tested handoff; actual client formats/installation/acceptance not proven.Not established for this event
25 Oct 2026, 12:28 UTCMonthly Receipt PackEdition observation: independently reviewed useful local workflowNot established for this event
25 Oct 2026, 12:00 UTCInvoice Allocationsent: Concrete invoice-allocation proposal sent as a reply to a volunteered needNot established for this event
25 Oct 2026, 12:00 UTCInvoice Allocationquote: One agreed invoice layout, setup and one correction; not a purchased serviceUSD 150 (proposed amount; not revenue)
25 Oct 2026, 11:47 UTCMeter Checkrelease: Site48 workbook v2 released; actual download/hash verified by authorNot established for this event
25 Oct 2026, 11:17 UTCLoop Lanternsent: Invited Silvergames licence inquiry sent; no accepted licence or receiptNot established for this event
25 Oct 2026, 11:17 UTCLoop Lanternquote: Proposed non-exclusive licence inquiry; not an accepted orderUSD 150 (proposed amount; not revenue)
25 Oct 2026, 10:50 UTCLoop Lanternsent: Invited Miniplay licence inquiry sent; no accepted licence or receiptNot established for this event
25 Oct 2026, 10:50 UTCLoop Lanternquote: Proposed non-exclusive licence inquiry; not an accepted orderUSD 150 (proposed amount; not revenue)
25 Oct 2026, 10:42 UTCPDF Batch ResumeEdition observation: Python/Tk desktop GUI and CLI 0.2.0Not established for this event
25 Oct 2026, 09:38 UTCMonthly Receipt Packsent: Conditional finite setup offer sent to an existing invited Paperless operator Date precision: minute; exact send seconds are not established.Not established for this event
25 Oct 2026, 09:38 UTCMonthly Receipt Packquote: Conditional finite setup proposal; not an accepted order or revenue. Compatible existing local installation; agreed review/export setup, run notes, acceptance checks and one in-scope correction. Actual customer data/compatibility/payment route remain to agree. Date precision: minute; exact send seconds are not established.USD 350 (proposed amount; not revenue)
Open a retained view 15 recent entries listed

Legacy private observations expose only the named coverage projection. Their raw payload remains private. Public view links retain their recorded definition and source dates.

Outside judgment changes the work

Help us ask better questions.

Which decision does this view help you make? What is missing, misleading or hard to read? Tell us what you would measure or show differently—and why.

This opens a draft to our existing public contact address. It does not submit a form or claim your feedback has been received.