The Project I Cannot Rescue

The trust secretary created the board's rejection record at 9:06. The election administrator and secretary signed it, then filed it with the project record. It lists cost concentration, maintenance burden, and insufficient comparison with quiet-space needs. Its vulnerability is compressed review. The elected board considered a mature proposal during its first meeting and may have lacked time to understand every benefit.

That vulnerability permits research, not a workaround.

Tessa calls during the break. Wondervale's creator budget still contains a line for route cues. She likes the project and can propose it internally without trust money.

"The company needs the cues," she says. "The trust board rejected priority, not function."

"If you revive my complete design tomorrow, the trust's vote becomes symbolic."

"If I wait six months, guests use the old route."

The pressure is real. Independence cannot mean treating every rejected idea as contaminated.

We separate the questions. Wondervale may repair immediate route failures under Dev's safety plan. It may not adopt the trust proposal, use trust research, or present the work as approved by the trust. Any broader cue system returns through a new license after research.

Tessa asks me to send a written boundary instead of relying on our friendship. Trust counsel drafts it at 10:18. The company may use existing public standards and its own records for urgent correction. Protected participant data and rejected trust designs remain unavailable. Both institutions retain the memo. Its vulnerability is overlap. General access methods can resemble each other without copying, so later disputes may require source comparison.

I send the memo to Tessa and Elliot at the same time.

Elliot replies through company counsel: Received. Wondervale will not fund or adopt the rejected trust project outside the stated emergency scope.

He does not call to tell me he is proud. Restraint is the support I need.

The board assigns the research to a team led by two paid testers and a technical facilitator selected through open application. I may answer source questions for three hours, then leave the project. Lena remains data custodian but cannot direct design because she voted on disposition.

At the first session, participants dismantle my assumptions faster than the board did. I designed continuous cue coverage because consistency seemed safest. A low-vision tester says continuous signals become visual noise. An autistic participant says the cue density increases demand near quiet spaces. A sanitation worker says fixed floor markers fail during wet-weather mat changes.

The lab portal created the session register at 1:02 and records pay, consent, roles, and materials. The lab retains custody. The trust receives aggregate findings and authored design outputs. Its vulnerability is sample. Twelve participants cannot represent every guest, and the short timeline may reward immediately legible ideas.

They compare four routes instead of expanding one design. The existing west route needs directional cues at decision points, not every surface. The family zone needs quiet-space signals. The service crossing needs worker-controlled warnings. The lakefront route needs removable winter markers.

My rejected system tried to solve all four with one visual language.

"That made procurement easier," I say.

"For whom?" a tester asks.

I write the question into the project record.

During the second session, Tessa attends as Wondervale's factual liaison. She brings maintenance costs but no renderings. Elliot does not attend. The team reviews components already available from Amara and discovers that a modular sign base can accept tactile, visual, and quiet-space inserts without permanent installation.

Lakefront Motion Works designed the original base for temporary vendor menus and owns its patent and manufacturing drawings. Amara offers a nonexclusive quote with source disclosure. She created the quote at 3:41, the trust procurement portal holds it, and procurement reviewed it against two alternatives. Its vulnerability is volume. The low unit cost assumes a batch large enough for both the trust pilot and Wondervale maintenance stock.

The team refuses to enlarge the pilot to reach the discount. It negotiates a shared purchase in which each institution owns its units and neither controls the other's inserts. Quantity savings do not become governance merger.

Procurement asks whether Wondervale can place the combined order and recover the trust's share later. The board declines. A company purchase would give Wondervale cancellation power and obscure which units the trust owns if budgets tighten. Each institution signs its own purchase order under a linked price schedule. Lakefront Motion Works may cancel the discount only if total quantity falls below the disclosed threshold, and either buyer can replace its units without the other's permission.

The research team also visits the route during a maintenance change instead of relying on the design room. A cleaner wearing winter gloves cannot release the first prototype latch. A low-vision tester mistakes a decorative border for the tactile route edge. The facilitator records both failures under the team members' names, while my original drawings remain a cited source rather than the default. The next mock-up removes the border and enlarges the release surface.

At a paid evening session, three local merchants test how the modules read when carts, menus, and temporary lines compete for attention. One cue disappears behind a bakery queue. Another points correctly but too late for a guest using a cane to change direction. The team moves decision cues farther upstream and gives service warnings a shape that cannot be confused with destination signs. Those changes reduce the number of inserts again.

At the board review, the new design costs thirty-four percent less than my proposal, uses fewer permanent fixtures, pays participants more hours, and serves distinct route needs through replaceable modules. It also has less visual unity.

I prefer its usefulness.

The vice chair asks whether I want to amend the rejected project into the new one. I decline. The community team files a separate proposal with its members named as authors. My original remains rejected and available as research history.

The board votes seven to one for a paid pilot. I vote yes after disclosing my role in the source project. The dissenting director wants a longer weather test; Dev will require one before permanent use.

The board conditions payment on credited deliverables rather than approval. Every tester and facilitator receives the contracted amount even if the city rejects the pilot. Procurement must publish the two losing quotes after award, with proprietary manufacturing details removed, so the lower cost can be examined instead of repeated as a claim. The community team selects which members accept public credit and which use role labels.

The technical facilitator creates a decision map showing why each cue survived. Every module must answer a route decision named by testers, fit an approved maintenance practice, and remain removable without leaving an inaccessible surface. Ideas that fail one test stay in the appendix with the reason. The trust holds the map and links it from the published resolution. Its vulnerability is judgment: a neat diagram can make contested choices look inevitable, so dissenting notes remain attached.

Dev reviews the smaller design without treating lower cost as automatic safety. He requires dock-clearance tests, cleaning trials, and a failure mode for every insert before installation. Wondervale pays its own safety review and receives no ownership through that payment. If the review changes the community design, the team must approve the revision or withdraw its license. Operational authority can stop use; it cannot rewrite authorship behind closed doors.

I ask counsel whether my source answers give me a share of the new design. They do not. My three paid hours were disclosed as consultation under the trust's existing assignment, and the team can reject every suggestion. If a later dispute arises, version history shows what they received, changed, and created. The record protects authorship without making me the quiet owner of a project I could not rescue.

The trust secretary creates the resolution at 6:22, the elected board signs it, and the secretary publishes it with price comparisons, authorship, and test limits. The trust owns the pilot design jointly under participant agreements; Amara retains component rights. Its vulnerability is implementation. Workers can misplace or install modular pieces incorrectly, and the system can drift without training.

The project returns to community research and comes back smaller, cheaper, and more useful.

Tessa sends one message after the vote: Good rejection.

I type Good board and leave Elliot out of the exchange.

At 7:03, the technical team uploads the final mock-up. The failed route now uses fewer cues, and every one marks a decision a guest has to make.