Context
docs/TYPE-CONNECTIONS.adoc is the estate's coherence mechanism for the type-research families: shared glossary, boundary statements, connection obligations, ownership routing. A 2026-09-27 sweep of the type-set shows three drift risks the map does not yet cover, plus two naming hazards. This issue tracks the sweep; per-repo issues link back here.
Checked this date: nextgen-typing, occupancy-types, kategoria, typell, absolute-zero, januskey, secret-types, echo-types, epistemic-types, residual-evidence-types, choreographic-types, tropical-types (README/description/state via API). Related: #115 (TYPE-CONNECTIONS receipts), echo-types#197 (clade registration).
Gaps
occupancy-types is unregistered. It is the state-resource counterpart of Tropical Types (cost): its thesis is the cost/state separation (max-plus vs the HWM monoid (p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂)) and protocol-frontier reclamation. Its ULTRAPLAN §5 vocabulary fence (cost grade ≠ occupancy grade ≠ echo index ≠ residue measure ≠ warrant) is an extension of this map's glossary that nobody has cross-checked. Its T2 ("buffer grade commutes with endpoint projection") and R5 are precisely the Tropical→Choreographic-style obligations the map tracks. It has shipped a working Session IR checker + stepper (14/14 fixtures green, 2026-09-27) and is exactly the kind of node that drifts when unlisted. (I have opened its README/glossary side of the link already: occupancy-types@1004e49.)
secret-types is un-minted (README is still {{PROJECT_NAME}} template text) yet referenced by epistemic-types#32 and by occupancy-types' ULTRAPLAN no-import list. It cannot take a map row until it has a question and a boundary; the row must wait, the tracking must not.
- Boundary notes missing for non-family type-set repos. The map's "Where things belong" covers
kategoria/typell but nothing on absolute-zero (CNO/OND) and januskey (reversible file ops). Two vocabulary collisions surfaced:
absolute-zero ships a "residue list of out-of-scope observables" — this map's glossary defines Residue as "information retained after a transformation". Three-way collision with the numerical residual already fenced here.
kategoria's "10 levels of type safety" is a strength ladder; this map states families "are complementary research questions, not successive ranks in a universal hierarchy of type-system strength". Both are legitimate; the two taxonomies need an explicit non-conflation note.
- Naming hazards (one is already half-documented):
tropical-resource-typing (used in occupancy-types ULTRAPLAN D1/R0-A) resolves to tropical-types; katagoria→ideas-to-alphas ≠ kategoria. Occupancy-types has recorded the errata in its decision log. Suggest the map's naming note grows one line covering the ULTRAPLAN spelling.
Proposed patch to TYPE-CONNECTIONS.adoc (owner wording wins)
"What each family means" table — new row:
| Occupancy Types | How do live-state footprints compose, and when are they reclaimed? Acquired-and-released resources compose by the high-water-mark monoid (p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂); closing a session endpoint names the reclamation point of its epoch. | An occupancy bound is not a cost bound: costs compose with +/max (max-plus), state composes by the HWM monoid, and neither reduces to the other. A measured live-byte count is a run statistic, not an occupancy grade. |
"Shared glossary" — additions:
| Occupancy grade | An element of a state-resource algebra (the HWM monoid) bounding live acquired-and-not-released resources. Distinct in kind from a cost grade. |
| High-water-mark monoid | (p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂), unit (0,0); sequential composition of state resources. Non-commutative, unlike max-plus. |
Optionally amend the existing Resource grade row: "…used for compositional usage or cost accounting. Resource algebras are of two kinds — cost (e.g. max-plus) and state (e.g. the HWM monoid) — and are not interchangeable."
"What each connection asks us to establish" — additions:
| Tropical → Occupancy | Costs and state are separate axes; a certificate may carry both, and one program may carry two bounds. | The separation statement with non-commutativity witnesses (occupancy-types R2 / algebra sibling tropical-types); a proof that the HWM monoid does not reduce to max-plus. |
| Occupancy → Choreographic | Compose buffer/occupancy grades when a global protocol is projected to participants. | A grade-preserving projection theorem (occupancy-types' T2; its R5 is the same question from the projection side). |
"Where things belong" — additions:
occupancy-types: the session-IR calculus, checker and stepper, the Occ occupancy spike, and stackcert. Registered here as the state-resource family node.
absolute-zero, januskey: applications/verification projects outside the five-family map. absolute-zero's "residue list" is an out-of-scope-observables list — not the glossary's Residue; reconcile or rename (its issue linked below). januskey's pending reversibility formalisation is a candidate consumer of, or sibling to, absolute-zero's CNO pillar — decided there, not here.
Naming note — one line: "occupancy-types ULTRAPLAN (D1/R0-A) uses the historical tropical-resource-typing spelling; canonical is tropical-types."
Checklist
Per-repo issue links will be edited into this body as they land.
Context
docs/TYPE-CONNECTIONS.adocis the estate's coherence mechanism for the type-research families: shared glossary, boundary statements, connection obligations, ownership routing. A 2026-09-27 sweep of the type-set shows three drift risks the map does not yet cover, plus two naming hazards. This issue tracks the sweep; per-repo issues link back here.Checked this date:
nextgen-typing,occupancy-types,kategoria,typell,absolute-zero,januskey,secret-types,echo-types,epistemic-types,residual-evidence-types,choreographic-types,tropical-types(README/description/state via API). Related: #115 (TYPE-CONNECTIONS receipts), echo-types#197 (clade registration).Gaps
occupancy-typesis unregistered. It is the state-resource counterpart of Tropical Types (cost): its thesis is the cost/state separation (max-plus vs the HWM monoid(p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂)) and protocol-frontier reclamation. Its ULTRAPLAN §5 vocabulary fence (cost grade ≠ occupancy grade ≠ echo index ≠ residue measure ≠ warrant) is an extension of this map's glossary that nobody has cross-checked. Its T2 ("buffer grade commutes with endpoint projection") and R5 are precisely the Tropical→Choreographic-style obligations the map tracks. It has shipped a working Session IR checker + stepper (14/14 fixtures green, 2026-09-27) and is exactly the kind of node that drifts when unlisted. (I have opened its README/glossary side of the link already: occupancy-types@1004e49.)secret-typesis un-minted (README is still{{PROJECT_NAME}}template text) yet referenced by epistemic-types#32 and by occupancy-types' ULTRAPLAN no-import list. It cannot take a map row until it has a question and a boundary; the row must wait, the tracking must not.kategoria/typellbut nothing onabsolute-zero(CNO/OND) andjanuskey(reversible file ops). Two vocabulary collisions surfaced:absolute-zeroships a "residue list of out-of-scope observables" — this map's glossary defines Residue as "information retained after a transformation". Three-way collision with the numerical residual already fenced here.kategoria's "10 levels of type safety" is a strength ladder; this map states families "are complementary research questions, not successive ranks in a universal hierarchy of type-system strength". Both are legitimate; the two taxonomies need an explicit non-conflation note.tropical-resource-typing(used in occupancy-types ULTRAPLAN D1/R0-A) resolves totropical-types;katagoria→ideas-to-alphas≠kategoria. Occupancy-types has recorded the errata in its decision log. Suggest the map's naming note grows one line covering the ULTRAPLAN spelling.Proposed patch to TYPE-CONNECTIONS.adoc (owner wording wins)
"What each family means" table — new row:
| Occupancy Types | How do live-state footprints compose, and when are they reclaimed? Acquired-and-released resources compose by the high-water-mark monoid
(p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂); closing a session endpoint names the reclamation point of its epoch. | An occupancy bound is not a cost bound: costs compose with+/max(max-plus), state composes by the HWM monoid, and neither reduces to the other. A measured live-byte count is a run statistic, not an occupancy grade. |"Shared glossary" — additions:
| Occupancy grade | An element of a state-resource algebra (the HWM monoid) bounding live acquired-and-not-released resources. Distinct in kind from a cost grade. |
| High-water-mark monoid |
(p₁,n₁)·(p₂,n₂) = (max(p₁, n₁+p₂), n₁+n₂), unit(0,0); sequential composition of state resources. Non-commutative, unlike max-plus. |Optionally amend the existing Resource grade row: "…used for compositional usage or cost accounting. Resource algebras are of two kinds — cost (e.g. max-plus) and state (e.g. the HWM monoid) — and are not interchangeable."
"What each connection asks us to establish" — additions:
| Tropical → Occupancy | Costs and state are separate axes; a certificate may carry both, and one program may carry two bounds. | The separation statement with non-commutativity witnesses (occupancy-types R2 / algebra sibling
tropical-types); a proof that the HWM monoid does not reduce to max-plus. || Occupancy → Choreographic | Compose buffer/occupancy grades when a global protocol is projected to participants. | A grade-preserving projection theorem (occupancy-types' T2; its R5 is the same question from the projection side). |
"Where things belong" — additions:
occupancy-types: the session-IR calculus, checker and stepper, the Occ occupancy spike, and stackcert. Registered here as the state-resource family node.absolute-zero,januskey: applications/verification projects outside the five-family map.absolute-zero's "residue list" is an out-of-scope-observables list — not the glossary's Residue; reconcile or rename (its issue linked below).januskey's pending reversibility formalisation is a candidate consumer of, or sibling to,absolute-zero's CNO pillar — decided there, not here.Naming note — one line: "occupancy-types ULTRAPLAN (D1/R0-A) uses the historical
tropical-resource-typingspelling; canonical istropical-types."Checklist
type-connections.dot+ SVG rebuilt (just type-map)1004e49)Per-repo issue links will be edited into this body as they land.