Working case
The declared interfaces
Green: passes · Dashed red: conflict · Dotted amber: unknown or unchecked. Select a line to inspect its rule.
A line means a declared relationship. It does not mean the two plans run at the same time.
Interface inspector
When the work happens
Solid bars are the viewed revision; outlines retain the original baseline. Shading marks a conflicting capacity rule’s time overlap. Drag a solid bar or use its left/right arrow keys to move one day at a time; duration stays fixed.
Day intervals include their start and exclude their end: a finish at day 11 can hand over to a start at day 11. Concurrent crew totals include all named plans, with no site-wide capacity test.
What would each repair change?
These are three stated alternatives, in a fixed order. Preview their effects on this revision before applying one; none is an optimised recommendation.
Change the local plans
“Permit required” records an assumption, not whether a permit has been obtained. Two “No” values agree without proving readiness.
Edit a local plan or a declared rule
Compare every local account with baseline
| Plan | Version | Days | Crew | Cost (£k) | Permit required? |
|---|
History and local decision
Worksheet decision
This is a label on this revision in this browser session, with no external approval or persistence. Later plan, rule or budget edits clear it. A cost ceiling is a comparison, not evidence of feasibility.
Construction, sources and limits
This worksheet evaluates one candidate against three kinds of declared constraints: a named pair’s crew allowance during overlap, a directed finish-to-start sequence, and equality of a shared assumption. An added QA plan introduces another pair’s capacity rule. Unknown values remain visible; passing listed checks does not establish complete project feasibility, operational capacity or safety.
Rosiak’s Sheaf Theory through Examples requires matching local sections before unique gluing (chapter 5, pp.148–149; chapter 10, pp.315–320). Sheafification is a separate construction on a presheaf and a specified coverage (pp.341–353). These ideas help distinguish agreement from an intervention that changes values or rules. The repairs here are explicit worksheet policies; they are not sheafification, a general search for compatible plans, or a monotone co-design construction. See the public source trail.
Essai 009 compares estimates of the same service; Essai 010 changes the scope and its declared interfaces. Here the question is how different kinds of plan constraints interact, and what changes when a proposed repair passes them.