What this shows
A real build lineage. Trammel is the current company label built on earlier verification work, not a venture invented after the fact.
What improved
The project got better every time it cut scanner drift, bolt-hole sprawl, and broad survey-replacement language, then kept only the strongest dated milestones.
Where it stands
The problem, founder fit, and V1 wedge are clearer than before. The active head path is now the flatter four-arm centering concept, and the commercial side is no longer just theory: structured cross-role discovery is now being used to test severity, spend, dissatisfaction, and whether a trusted fix looks worth paying for.
March 20, 2026
The commercial wedge tightened and discovery got more disciplined
The strongest current reading is no longer "replace survey." It is to help shops close the gap around repeat survey loops, reduce some fit-up survey cases, and let correction happen with better live guidance before final survey is called or recalled. The same date also brought a more disciplined discovery posture instead of just more internal theory.
What got clearer
- Founder field evidence now points to survey cost at about $1000 per spool in at least some workflows.
- Survey turnaround can stretch to 24+ hours in some cases.
- Fit-up survey sometimes happens pre-weld because the shop wants the first post-weld shot to land cleanly.
- The floor-level pain is not just math. It is repeated setup, handling, uncertainty, and stress.
- Founder field logic also suggests survey-fabrication competence is scarce enough that labour fit may be part of the pain, not just geometry.
What changed in product and discovery reading
- The strongest near-term value may be live correction guidance, not just a static readout.
- The first trustable offer may be service before product.
- Training now looks like part of the offer, not a side note.
- The buying logic is now being compared against adjacent premium shop tools instead of generic startup categories.
- The active first-print head has been simplified into a flatter four-arm flange-OD centering concept instead of the older deep-body mechanism.
- The discovery package has been tightened into a short intro page plus role-based manager, foreman, fitter, and QA/QC questionnaires aimed at severity, spend, dissatisfaction, and whether a trusted fix looks worth paying for.
What this still does not prove
- It does not prove willingness to pay from named buyers.
- It does not prove the exact trust threshold QA/QC will require.
- It does not prove whether the first durable model is service, product, or both.
- It does not prove survey acceptance after correction yet.
- The hardest total technical step still is not necessarily the first commercially critical proof step.
- It does not yet prove the labour-scarcity reading outside founder experience.
March 14, 2026
Current production work and design locks
The project shifted into a cleaner production-work posture. Current head research was checked against still-current parts, and the center-first design language was tightened.
What moved forward
- CM5-class compute still looks believable for a per-head embedded Linux path.
- Precision inclinometer candidates in the SCL3300 class still look worth evaluating for plane-related sensing.
- The center-first wording is now cleaner: face seating defines the plane, and controlled jaw convergence recovers the center.
- The competition view is cleaner too: adjacent categories exist, but direct confirmation-tool matches are still much less obvious.
What this did not prove
- It did not lock the final head electronics stack.
- It did not prove the final communications path.
- It did not prove delivered accuracy.
What still matters most
- Seating repeatability
- Centering repeatability
- Rigidity
- Shared transform stability
March 13, 2026
Trammel became the active name and the public package was cleaned up
The naming loop was closed for production purposes. Trammel became the active outward-facing name, and the public material was cleaned up under that one label.
What changed
- Trammel became the active working name going forward.
- Older naming branches were retired in favor of one outward-facing name.
Why it mattered
- The outward-facing story now speaks with one name instead of a split label structure.
- The review site, diary, and proposal material became easier to read and share.
What this marked
- This was the point where the project clearly moved out of naming churn and into production-facing cleanup.
March 12, 2026
The review package replaced the loose attachment flow
The project stopped relying on scattered attachments, rough file drops, and weak promo-style material. A static technical review became the main discussion layer.
What changed
- A static review page became the main public discussion layer.
- The work shifted toward readable diagrams, a cleaner diary, and a more inspectable package.
Why it mattered
- The project became easier to review seriously.
- The outside-facing material stopped underselling the venture.
What it did not mean
- This was a packaging improvement, not a claim that the hardware itself was solved.