Index

Catalog of every page in the SINE strategy wiki. Updated on every ingest. Read this first when answering a query — locate relevant pages here before opening them.

• Activity log — append-only record of every ingest, query and lint, with the reasoning behind each. ⟨Link added by the 2026-09-14 lint: the log was the wiki’s only orphan page, reachable from nothing.⟩

• Open tasks — positions SINE owes and questions the wiki cannot answer from its own sources. Twenty items ⟨count corrected by the 2026-11-02 lint: item 20 — what moved between CI and DSS in the FY2025 segment restatement — was added by the 2026-11-01 Inspire 2027 ingest, and the same lint extended item 11 from seven un-ingested sources to nine, with the PACT Tech Specs v2.2 and the GB/Z 185.1—2026 agent-interconnection standard, neither of which was in any register⟩. ⟨2026-11-06: item 11 is back to eight — the PACT Tech Specs v2.2 are ingested, completing Martin Pompéry’s three-document PACT bundle — and item 19’s PACT strand is closed: the protocol answers confidentiality with access control and says so, placing access management, credentials and audit logging out of scope in §7.2.1.⟩ Before that: nineteen, none closed by the 2026-10-27 OECD ingest, which narrowed one. Item 19 — three communities stating the confidentiality requirement and none reaching for cryptography — asked for a counterexample, and OECD (2026) is it: a committee-declassified report that inventories four PET categories, makes supporting them one of eight policy considerations, and bounds rather than rejects them (“Neither alone is sufficient”). That falsifies “not known” at policy level, leaves “not deployable” standing on a governance barrier (agreeing which PET), and puts a fourth reading in front: sequencing — confidentiality is the second problem, and a community that has not settled who governs and who pays has nothing to deploy a PET into. The live question is now sequencing vs. incentive, tested against iLEAP specifically. The same ingest leaves a commit overdue across two ingests and the Anthropic S-1 downloaded but un-ingested. ⟨The count bullet previously read “Seventeen as of the 2026-10-26 ACN ingest” while naming item 19 — corrected here.⟩ Before that, the ACN ingest raised item 19; it also resets the honest baseline for AI Advantage. Before that, sixteen as of the 2026-10-19 lint (item 15 closed, item 17 mostly closed), which raised no new item but corrected the count and moved four. ⟨The bullet had read “fourteen as of the 2026-10-14 ingest” and was two behind: items 16 and 18 were added by the 2026-10-16 ingests without this line being touched.⟩ The sharpest today: the GIZ tender’s contract term started on 15.10.2026 and is now four days into a twelve-month term the wiki holds no record of — eight consecutive lints on one unanswered question (item 5); CCTI Milestone 1 is nineteen days past due, the wiki’s oldest unanswered question, on the one programme in it that SINE runs itself (item 8); the un-ingested backlog is seven, not six — a source had been sitting in another item’s prose — and was passed over entirely by the heaviest ingest week in the wiki’s history, which retires capacity as its explanation (item 11); and the Hitachi engagement inverted, from “exists entirely outside the wiki” to its largest single cluster in eleven days, while three of the four problems named with it survived the inversion — including a fourth self-authored file filed into sources/ three days after the last lint proposed a header for exactly that case (item 13). Before that, fourteen as of the 2026-10-14 ingest, which raised one: how concentrated is the NatCat model layer, actually? — a new concept page’s entire spine rests on one clause in one anonymous memo, and the claim most likely to travel into SINE’s own writing is the one least able to carry a citation (item 14). Thirteen as of the 2026-10-12 lint, which raised no new item but moved four. The sharpest: the GIZ tender’s contract term starts 15.10.2026 — in three days — while its award period runs to 13.11.2026, two facts that have sat side by side on the page since August without the contradiction being named (item 5); CCTI Milestone 1 is twelve days past due and the project page was still carrying it in the future tense (item 8); the un-ingested backlog grew from four to six while one source was ingested, with the OpenMined/Prewitt essay promoted to the front of it as a third independent arrival at the Data Commons architecture that disagrees with SINE about governance sequencing (item 11); and a third self-authored file landed in sources/ — this time carrying the provenance header the previous lint asked for (item 13). Before that, thirteen as of the 2026-10-05 lint, which raised one: a live Hitachi Rail engagement that exists entirely outside the wiki — two client-facing work products filed into the immutable sources/ layer, resting on a pre-analysis that is not in the repository, carrying figures with no primary source, and stating as fact an iLEAP claim the wiki holds as an ⟨Inference⟩. No wiki page mentioned Hitachi — partly closed 2026-10-15: the Integrated Report 2026 is ingested and Hitachi, Hitachi Rail and Customer Zero now exist; the unsourced figures are registered as unsourced, not resolved. Twelve as of the 2026-10-01 iLEAP ingest, which raised: does the “Internet of Emissions Data” converge, and why is there no cryptography in it? — iLEAP has left PACT’s wire, and answers confidentiality by aggregation, in a standard SINE co-authored. Before that, eleven as of the 2026-09-28 lint — nine from the 2026-09-21 lint, one raised by the Kim ingest (is confidentiality a constraint on the corporate side of nature-related reporting? — TNFD/NGFS unread, and a load-bearing inference hangs on it), and one new at the 2026-09-28 lint: three source bundles filed 22–25 September and never ingested — the GHG Protocol’s Actions and Market Instruments RFI response summary, IPCC AR6, and CIFS Megatrends 2026 (a fourth, the Bundesbank climate roadmap 2026-27, joined them 2026-10-02). The GHGP one also closes the wiki’s largest page gap: the GHG Protocol is cited on nine pages and has no page of its own. The nine: SINE’s critical response to the ICC ledger architecture candidate (unblocked, specified in eight points), three items carried since 2026-08-17, the unrecorded outcome of GIZ Tender 10036847, the un-ingested Dafoe Cooperative AI paper, the unfiled FNG-Marktbericht, CCTI Milestone 1 (30.09.2026), on which no source newer than the 2026-06-17 proposal draft exists, and — new at the 2026-09-21 lint — five papers whose findings live only in a dossier and were never carried onto a wiki page; two ingests closed 2026-08-17. Nothing has closed since then.

Context maps (MVC)

Minimum Viable Contexts — relational maps built to give a strategy conversation a shared context before positions are taken. Tell Münzing’s artifact type: a value chain × its actors × its people × its money, with the holes marked. Working instruments, not findings.

• Coffee — Minimum Viable Context (v0.1) — the financial-markets strategy’s first test case, on the GIZ/WHH approach in coffee. Nine chain stages, each with primary data / the claim to be made believable / SINE’s five instruments / the economics / the finance touchpoint; actor register by mandate, money-and-horizon, relationships and principal-or-agent distance; people register with adjacency; register of holes. Load-bearing findings: principals sit at both ends and SINE is in the agent layer; coffee’s buy side is family- and state-controlled, not capital-market-facing; warrantage exists because land cannot be mortgaged (Ethiopia state-owned land vs. Vietnam mortgageable LURCs) and its verification technology is a padlock, which caps it at walking distance; no instrument anywhere prices a roaster’s capital against verified farm data; GCP holds a farm-level data standard with no located governance or EUDR position; the coffee PCF’s ±2.04 on a median of 2.18 is the business case for primary data; BIS/IAIS names data infrastructure as a systemic constraint on reinsurance pricing. Two contradictions to Tell’s account held open (WHH/warrantage attribution; GIZ chair succession)

Organization

• SINE Foundation — think-and-do-tank behind the Data Commons framework; originator org

• GIZ — German development-cooperation implementing agency; commissioning party for Tender 10036847; procurement mechanics (70/30 weighting, 500-point cutoff, AVB backstopping)

• ARIA — UK research agency; funder of CCTI via its Scaling Trust programme; cohort-of-creators model, Arena benchmarks/challenges; no primary ARIA source ingested yet

• Hanns R. Neumann Stiftung (HRNS) — Hamburg coffee-sector implementer in the Neumann Kaffee Gruppe orbit; proposed consortium lead for Tender 10036847; 250+ staff, 386,000 farmer families in 18 countries, endowment figure contradictory (€150M vs €30M); ICP / coffee&climate toolbox / aCLIMAtar as the two Key-Expert-2 leads; includes E.D.E. Consulting and the unverified Vietnam access

• Global Coffee Platform (GCP) — coffee-sector multi-stakeholder association, 160+ members, Geneva/Bonn, ED Annette Pensel; the invitation layer per Tell Münzing; 4C + IDH merger (2016); owns the Coffee Data Standard (15 farm-level indicators, feeding the Delta Project toward government endorsement) and ~10 country platforms, incl. Kenya funded by BMZ via GIZ; NKG and HRNS both members; no located EUDR position and no located statement on data ownership or governance — the gap that is precisely SINE’s layer

• GEO BON — Group on Earth Observations Biodiversity Observation Network; steward of the Essential Biodiversity Variables framework, convener behind Kim et al. (2026); operates the BON network model (EuropaBON, FrenchBON, MBON, Soil BON…), “BON in a Box”, and the proposed GBiOS global observing system; ⟨Inference⟩ a standard-setter that has built an intermediate data layer and arrived at federation with no governance design at all — the layer SINE claims. Stub: single-source, self-described; founding date, legal form, funding and governance all unknown

• Inside Track x (Re)Insurance — anonymous group of senior UK/European non-life insurance, reinsurance and broking professionals; publisher of Sounding the Alarm (2026). Stub by design, not by neglect: no named person, no legal form, recruitment over Signal, participants anonymous even from each other, under a disclaimer that nothing commercially sensitive is shared — read as an antitrust-shaped construction, and as a worked instance of the Kartellrecht als übersehene Bremse point from 2026-08-15. The wiki’s only primary voice from inside the insurance industry

• Welthungerhilfe (WHH) — German aid association (1962), ~37 countries, €383.5m project funds 2024; ~77% institutional/project-tied money — which is why it can invite and implement but not commit; Vorstand + Präsidium on file (incl. a Makerere engineering dean and an ex-DAX CIO on the finance committee); coffee work via Coffee Circle in Ethiopia and a DRC cooperative; no WHH-attributed warrantage work found (contradiction held open); no data platform of its own, plus published criticism of corporate power in digital agri-food

• Hitachi, Ltd. — ⟨added 2026-10-15⟩ Japanese digital-infrastructure conglomerate, FY2025 revenues ¥10,586.7bn, four sectors, management plan Inspire 2027; the operating model is Physical AI + IWIM + Lumada/HMAX + Customer Zero, with the FDE Team and Data Fabric turning internal deployment into resellable assets. EcoSpace (HMAX Energy) is the precedent: sustainability data already shipping as a product, quantified and visualised but not independently checkable. The structural finding ⟨corrected 2026-10-16⟩: Hitachi publishes Scope 3 (209.18 Mt-CO₂e, 99.8 % of the footprint — in the Sustainability Report, not the Integrated Report) but does not govern it as a quantity — Cat. 1 is spend-based with partial primary data, the targets are per unit of added value, the only compensation-linked supply-chain KPI counts partners taking action on environment and human rights (5,232 against a 6,000+ target) ⟨corrected 2026-10-29: previously, and wrongly, given as the 428-of-700 “partners with plans” KPI⟩, and the headline claim — 280 Mt/yr avoided emissions — is ~600× the assured actuals (460 kt Scope 1+2). Strong where Hitachi owns the meter, estimated where it depends on counterparties. Also holds the standardization apparatus (committee under the CTO, Rule-making Promotion Group, IEC SMB seat). ⟨updated 2026-11-01 from the Inspire 2027 deck: the four-sector dashboard on identical measures (revenue, margin, targets, Lumada ratio, backlog) — the four sectors are parallel, not nested, and Lumada is an attribute of sector revenue, not a fifth sector, which is how digital growth on physical assets avoids double counting; HMAX sized at ¥300bn / over 20% margin, ~7% of Lumada; and the two basis changes — consolidation offsets newly reflected, CI and DSS restated on a new segment structure — that sit under the 29% → 40% jump. Plus, from the Birje deck, the DSS/CI division of labour: the physical sectors own the HMAX products, the customers and the 175 Customer Zero sites; DSS owns the horizontal beneath — FDE service, AI operations, Data Fabric, ontologies — so a proposal that needs a site goes to CI, a proposal that is a mechanism goes to DSS⟩

• Mobility sector — ⟨added 2026-10-15⟩ the engagement target; CEO Giuseppe Marino; transforming “from a rail-focused OEM into a multimodal digital mobility company”, margin path 8.2% (FY25) → ca. 11% (FY27) through digital/service expansion. HMAX Mobility: 2,500 trainsets, over 35% built by someone other than Hitachi, pipeline ¥200bn → ¥600bn; Omnicom acquired “which brings sensor technologies”; RAILA/Project Jackie+ runs 90 AI agents in engineering (bidding, diagrams, code — not procurement). Three absences that reshape the pitch: Customer Zero is never claimed for Mobility, circular economy belongs to Connective Industries, and there is no supplier-data or Scope 3 content at all. Holds the register of unsourced load-bearing claims carried by the engagement — the 14.93% / 76-supplier figure, the competitor facts (the “Enhanced by HMAX” entry was closed 2026-10-15, see below; the HTR 412 SPD entry is partly sourced the same day from trade press — EPD International, suppliers from 23 countries, 1–4 risk scale, valid to 2028, platform ambition for 20,000+ suppliers — while the 38/8/11-of-26 breakdown stays open). Also carries the one piece of demand-side evidence the repo holds: Hamburger Hochbahn’s 94% recycling-rate requirement in its 2022 metro tender. ⟨updated 2026-11-01: the CEO deck’s Mobility line — revenue ¥1.3tn, margin 9.2% excluding integration cost (the Integrated Report’s 8.2% includes it; both bases recorded), Lumada ratio 30% → 33%, backlog ¥7.1tn; the smallest and lowest-margin of the four sectors, carrying by far the longest order book. HMAX Mobility is pitched on a signalling installed base of over 30,000 km — not trainsets — into a ¥3tn railway-AI market by 2030. ⟨2026-11-01⟩ Rail also appears as a client of Hitachi Digital — “the FDE model for Smart Operations, including Factory IQ and vehicle manufacturing optimization”: the live digital engagement with Rail is in the factory, not on the fleet⟩

• Hitachi Digital (Digital Engineering & AI Solutions BU) — ⟨added 2026-11-01⟩ the merged Hitachi Digital Services + GlobalLogic unit inside DSS, called Hitachi Digital ⟨the renaming is undated — no source here gives a date; corrected by the 2026-11-02 lint, which removed “in H2 FY2026” as unsourced⟩; CEO Anand Birje (Santa Clara; ex-Encora, which he sold to Coforge for $2.35bn in April 2026; ex-HCLTech $4.5bn digital services; Emerson/Capgemini SCADA-MES background), in post since Aug/Sept 2026 and reporting into sector CEO Jun Abe. What it owns is the horizontal under HMAX, not an HMAX product: the Physical AI FDE service, HMAX AI Operations, and the HMAX Data Fabric (data platform + IWIM/AI, “ontologies and physical models”), plus HARC as the industry-ontology/context layer. The growth split is the finding: the internal channel — “sell-to / sell-with Hitachi Front Business Units” — ran at +46% YoY in FY2026/H1 while the external top-100 accounts, 65% of the business, compound in single digits; three of its four named reference cases are Hitachi companies (Rail: the FDE model for Smart Operations, Factory IQ, vehicle manufacturing optimisation). ⟨Inference: the likeliest home for a horizontal trust mechanism — the layer, the IP-led economics and the internal-first channel all match — with a clear falsifier: the deck contains no confidentiality, provenance, verification or standards vocabulary at all, and the HMAX-focused CTO office that would own such a component is only being stood up in FY26⟩ ⟨updated 2026-11-01 with Abe’s sector deck: the Data Fabric is explicitly DSS’s and funded — “enhance data fabric capabilities” is technology bullet one of the ¥30bn FY2026 organic investment — and the sector stack has a Security pillar the BU stack lacks (OT availability, not provenance between parties; adjacent, not identical). Now also carries “Where SINE attaches, and to whom”: the Data Fabric-vs-AI-Operations layer answer, the six-audience internal map, the narrative for approaching a counterpart function that does not exist yet, six translations for making SINE legible to the DSS CEO/CTO, and the circular-economy nuance for Hitachi Rail’s CSO — every Data Fabric arrow runs Hitachi-to-Hitachi, so the CSO’s supplier-data problem sits exactly where the arrows stop⟩

• Hitachi-AIST Circular Economy Cooperative Research Laboratory (CECRL) — ⟨added 2026-10-16⟩ joint lab of Hitachi and AIST (Japan’s largest public research institute, under METI), established to “create internationally accepted rules and standards originating from Japan”. Its named output is CVP. Read as the research end of a standards pipeline whose other end is already built: a Group Standardization Committee under the CTO, a Rule-making Promotion Group, Priority Theme Working Groups, a seat on the IEC Standardization Management Board (15 members worldwide), and two METI Industrial Standardization Awards in FY2025. ⟨Inference: a distribution channel SINE cannot replicate — which is the argument for entering as a contributor to Hitachi’s metric rather than as the author of a rival one.⟩ ⟨updated 2026-10-16: the CVP route is now known — ISO/TC 323 WG3, as an annex to ISO/AWI 59011, plus the WBCSD Global Circularity Protocol as a second, de facto route; the named people on both sides; and a second, entirely undescribed half of its proposal, “data model standards for information on goods including quality information”, which is the layer closest to SINE. The April 2026 research report is now in sources/, not yet ingested.⟩

• Circular Economy Association (Japan) — ⟨added 2026-10-16⟩ Japanese industry body founded 26 Feb 2021 around ISO/TC 323, the venue where Hitachi and AIST are socialising CVP. Its Study Group on Common Infrastructure Standards for Circular Business (7 Feb 2025) was constituted as a “(confidentiality and consultation) committee” — Panasonic, Mitsubishi Electric, Mitsubishi Materials, INFRONEER, Kajima, JEMAI, ABeam, academics from Hosei/Toyo/U. Tokyo, with METI and the Ministry of the Environment as “audience members”. ⟨Inference: the confidentiality test fires before the content does — the clearest single piece of evidence in the wiki that this is a bargaining failure, not a coordination failure.⟩

• Röchling Stiftung — ⟨added 2026-11-06⟩ German foundation behind the POLYPROBLEM report series and SINE’s counterpart in the engagement Tell Münzing opened around its Foundation Camp (2026-11-05). Founded 1990 by the Röchling entrepreneurial family, which also owns the Röchling group — “international führende[r] Anbieter von Hochleistungs-Kunststoffen” — so the foundation’s subject matter and its funder’s business are the same material. Theme is narrow and explicit (Kunststoff und Umwelt); the operationally relevant sentence is its self-description: “sieht sich dabei nicht als reiner Förderpartner, sondern ist vor allem interessiert am Aufbau neuer sektorenübergreifender Kooperationen und Netzwerke” — i.e. a convenor, which is why the reports read as convening instruments rather than research. POLYPROBLEM is a joint initiative with Wider Sense GmbH (B Corp since 2015), two reports a year since a 2019 study that found “eine globale Agenda weiterhin nicht in Sicht”. Carries the report table (06 ingested, 07 pending, 08 ingested), the named Stiftung staff, and the caveat that the six ChatGPT preparation documents in sources/roechling/ are second-hand and must not be cited as evidence. ⟨Inference: read in sequence the three reports are one argument the foundation has never assembled itself — each report’s unanswered question is the next one’s subject⟩

Dossiers

Analysis products written for the group, versioned in dossiers/ in this repository. Self-authored — cite the underlying source, not the dossier.

• dossiers/2026-11-04-polyproblem-06-07-08-drei-reports.md — ⟨added 2026-11-05 (written 2026-11-04), for Tell Münzing ahead of the Röchling Foundation Camp⟩ “Drei Reports, ein Henne-Ei-Problem” (German, PDF alongside, 9 pp.). Reads POLYPROBLEM 06 Circularity Code (2023), 07 Die Kluft im Kopf (2025) and 08 Wechsel-Geld (2026) as one argument the Röchling Stiftung has never assembled itself: 06 is the supply side (the data exist and do not flow, blocked by Betriebsgeheimnis), 07 the demand side (the offer exists and is not used), 08 the capital side (the value exists and is not bookable). All three call it a Henne-Ei-Problem; none cross-references the others. Carries every quotation with page reference, plus the finding that the foundation reversed its own explanation between 2023 (“Es ist eine Frage der Haltung”) and 2025 (“keine individuelle Entscheidung … kollektive Verantwortung”). ⚠ Report 06 has since been ingested properly ⟨2026-11-06⟩ — cite polyproblem-circularity-code-2023.md, not this dossier. Report 07 is still chat-only and has no wiki page, so its findings here are not yet citable — open task 21
• dossiers/2026-11-01-hitachi-where-sine-attaches.md — ⟨added 2026-11-01, for Tell Münzing to share with the team⟩ “Where SINE attaches inside Hitachi’s digital business” (English, PDF alongside, rendered --external). The briefing form of the wiki section on Hitachi Digital: the four facts (the Data Fabric is named, owned by DSS, funded through “enhance data fabric capabilities” in the ¥30bn FY2026 line, and has no architect yet); which layer — written in the Data Fabric, argued from AI Operations, claiming neither partner-held box; who to talk to — the three-beat narrative for a counterpart function that does not exist, the six-audience map, and the CE / value-chain nuance for Hitachi Rail’s CSO (every Data Fabric arrow runs Hitachi-to-Hitachi, so the supplier-data constraint sits where the arrows stop); how to be legible to DSS — five translations into the sector’s own scoreboard. Cites Abe and Birje
• dossiers/2026-10-15-hitachi-customer-zero-signierte-sensordaten.md — ⟨added 2026-10-15, revised 2026-10-16⟩ “Signierte Sensordaten als Customer-Zero-Fall”: the SINE × Hitachi fit, re-cut around Hitachi’s own Customer Zero doctrine as the entry point (German). Seven parts — the shared logic of circular economy and sensor data (both are claims about an object that cannot be checked on the object; all three of Hitachi’s CE KPIs measure its own waste handling); what HMAX and Lumada 3.0 are built for commercially (Lumada 80-20, growth via “services to third-party assets” — so confidentiality sits on the critical path of the central financial target); Customer Zero as entry; the case in Hitachi’s own operations (the STI-linked site-decarbonisation KPI as a Heiss-shaped meter problem, run at a Connective Industries plant); what crosses the company boundary (product passports and declarations: Hitachi Rail’s own HTR 412 SPD already records the supplier data gap as a risk note; CBAM covers the materials a train is made of, not the train, and prices missing primary data through punitive defaults at reasonable-assurance level — so SINE proposes an attachment to existing documents, not a new one); a five-stage learning path to Rail with a named artefact per stage; and the Physical AI business case on the now-sourced 2027 sensing commitment. Revision 2026-10-16 (now eight parts, 19 pp.): a new part 5 on H-AIST CE Lab. — the lab measured the same information failure in plastics recycling, prices it as “uncertainty-related costs”, and its own revenue side (Shinkai) and METI (Mimaki) state the gap while prescribing only trust; the CVP/residual-value grading split, with the grading half identified as the layer SINE touches and CVP as the accounting layer to leave alone; the association→ISO personnel union as the access route; part 6 re-cut around the stated DPP division of labour (Europe upstream, Japan downstream + the IEC Common Data Dictionary) and around the attestation layer that is unoccupied in both architectures; a parallel stage-3 track on secondary materials; three added falsifiers. Feeds Hitachi, Ltd., Hitachi Rail, Customer Zero, Residual-value grading, Hitachi-AIST CECRL
• dossiers/2026-08-22-ki-marktstruktur-im-vergleich-zu-adtech.md — “Zwei Inversionen und eine Leerstelle”: market structure across the whole AI value chain (compute → foundation models → inference API/MaaS → orchestration/routing/agents → distribution) compared with adtech property-by-property rather than layer-by-layer — eleven structural properties, each with one of three verdicts (German, quotations in the original language; 12 pp. PDF alongside). The stickiness inversion, the opacity inversion, the missing auction; four agencies locating the chokepoint on four different layers and none of them naming the routing layer; three explicit corrections to Dossier 1. Feeds AI stack market structure and binary-verifiable remedies
• dossiers/2026-08-21-adtech-als-regulatorischer-praezedenzfall.md — 18 research papers on market structure, market failure and remedies in programmatic advertising, read as a precedent for AI-stack regulation (German, quotations in the original language; 17 pp. PDF alongside). Six mechanism families, a provenance audit, six transfers and three breaks. Feeds adtech as a regulatory precedent and binary-verifiable remedies
• dossiers/2026-08-14-giz-tender-10036847-konsortialszenarien.md — GIZ tender, second dossier: consortium scenarios A′/A/B/C, person assessment against the ToR profile, five formalia, conflict of interest. Integrated into consortium constellation
• dossiers/2026-08-13-giz-tender-10036847-anforderungen-und-sine-fit.md — GIZ tender, first dossier: hard vs. soft requirements, two ToR contradictions, three Bruchstellen, options A–D
• dossiers/2026-08-15-cooperation-failure-taxonomies.md — cooperation-failure taxonomies (basis of coordination vs. bargaining failure)
• dossiers/2026-08-06-sine-foundation-executive-summary.md — SINE Foundation executive summary

Projects

• CCTI — Centre for Cryptographic Trust Infrastructure — ARIA-funded Cambridge/SINE project, Jul 2026–Dec 2027; cryptographic trust infrastructure for AI agents; six work packages, three Go/No-Go gates, SINE leads WP-C (EIC) and WP-E (institutional lens); SINE’s second centre of gravity next to the sustainability-data line; published prior art by its own people ⟨2026-08-17⟩ — Heiss (2023) and Man & Kleppmann (2026), which give WP-A/M5 a concrete baseline and the supply-chain pathway a precedent

Strategy & positioning

• Hypothesenpapier: Strategische Standortbestimmung (2026-08-05) — internal discussion paper for SINE leadership; what SINE does/doesn’t do, notable gaps, competitor landscape (data-exchange standards + Tech-for-Good NGOs), ten hypotheses, open questions

Discussions & transcripts

Verbatim records of strategy conversations. Raw sources, not wiki pages — cite them, don’t edit them.

• Strategiegespräch Themenfelder (2026-08-15) — sources/discussions/2026-08-15-strategiegespraech-themenfelder.md. Martin Pompéry × Elinor, Slack. Full transcript. Relevanzfilter korrigiert auf Transformationsbeschleunigung als Nordstern; Meta-Hebel nach Funktion sortiert (Preise/Nachfrage/Regeln/Durchsetzung/Ermöglichung) inkl. Versicherung, Kohlenstoffmärkte, Haftung, Prüfindustrie; Externalitäten-Frame in fünf Punkten bestritten; Koordinations- vs. Verhandlungsversagen (→ Konzeptseite); Ratenzerlegung Fenster × Trefferquote × Durchlaufzeit; ökonomische Traditionen als rivalisierende Hypothesen (Dixit/Pindyck Realoptionen, Akerlof, Heller Anticommons, DiMaggio/Powell Isomorphie); Kartellrecht als übersehene Bremse; Hypothese „Vertraulichkeit löst das Eintrittsproblem und beschädigt den Verbreitungsmechanismus”. Themenfeldrecherche bewusst noch nicht gestartet.

Market & competitors

Data-exchange standards

• Gaia-X — EU trust-framework/certification layer; contested membership figures; Scaleway exit; Bundeskartellamt-confirmed Catena-X link
• IDSA — connector-based data-sovereignty standard (IDS-RAM); DIN SPEC 27070; personnel overlap with Catena-X
• Catena-X — automotive data ecosystem; €245M budget/€100-108M BMWK subsidy; Cofinity-X commercial spin-off
• CIRPASS-2 — WEF-incubated battery passport standard ahead of the EU Battery Regulation; Systemiq overlap
• GS1 — barcode/GTIN standards body; GS1 US $188.8M FY2024 revenue (IRS Form 990) — the mature, fee-self-funded reference point
• Ocean Protocol — tokenized Compute-to-Data; 2025 ASI governance collapse and SDNY federal lawsuit — cautionary case
• Commercial clean-room vendors (Decentriq, Duality, Habu) — for-profit MPC/HE/confidential-computing vendors; Habu acquired by LiveRamp 2024

Tech-for-Good / governance-tech NGOs

• Mozilla Foundation — non-profit/for-profit sibling structure benchmark; revenue-concentration-risk case study; MozFest, Privacy Not Included, small-grants tactics
• OpenSSF — tiered membership funding with a Governing Board/Technical Advisory Council firewall
• MyData Global — federated local-hub model for personal-data rights
• Creative Commons — mass small-donor funding vs. single-funder concentration risk
• Stiftung Neue Verantwortung — German tech-policy think tank; concentrated foundation/government funding
• AlgorithmWatch — ADM-systems advocacy; dedicated Head of Fundraising since 2024; “Automating Society” flagship report
• Bertelsmann Stiftung (digital ethics) — “algo.rules” framework cited by the EU Commission; independence criticism despite no-voting-rights structure
• Ada Lovelace Institute — embedded inside the Nuffield Foundation; NHS Algorithmic Impact Assessment
• Open Data Institute (ODI) — transparent membership pricing; 20+-country Node franchise network

German sustainability research institutes (Ecornet peer group)

• Ecologic Institut — Berlin gGmbH, 1995, project-funded with no government base funding; the only institute in the German peer group with data governance as a named research strand (circular-economy data space, DPP); Ecornet member — cooperation candidate, not competitor
• IÖW — Institut für ökologische Wirtschaftsforschung — Berlin, founded 1985, 69 staff (end 2024), explicitly no growth strategy; researcher-owned — ~50 individual Gesellschafter*innen plus VÖW e.V., incl. Ecologic’s founder Andreas Kraemer; Ecornet member. Profiled by a full crawl of all 110 person pages at the level of individuals: the overlap with SINE is thematically broad and technically null — Datenraum, Gaia-X, Produktpass, CSRD, ESRS, Scope 3, Blockchain, Interoperab*, Datensouveränität and vertraulich* all score zero profiles. The asymmetry: IÖW asks whether and why firms report, SINE asks how a figure becomes believable without disclosure. Three people one step away (Schöpflin, Lautermann, Pentzien); assets SINE lacks: standing inside the ESRS text, and the Ranking der Nachhaltigkeitsberichte (× future e.V., since 1994). Cooperation candidate, not competitor — with the predictable objection recorded: an institute whose norm is disclosure may read an instrument that makes disclosure unnecessary as helping firms avoid transparency
• ⟨Not yet written, from the same 2026-10-29 survey: Ecornet itself (8 institutes, ~800 staff, founded 1977–1995 “outside the established science system” — the peer group IÖW belongs to), Wuppertal Institut (closest on DPP artefacts; GS1 Exchange 2026), Öko-Institut, ifeu (BISKO municipal CO2 accounting standard), ISOE, IZT, UfU, plus the non-Ecornet adelphi, CSCP, Borderstep. Source note: sources/research/2026-10-29-german-sustainability-research-institutes.md⟩

Impact & ESG data vendors

• The Upright Project (Upright Oy) — Helsinki net-impact data company; ~800-topic proprietary Universal Impact Structure, vertically-trained Leena-2 model; not a competitor but the control group — same coordination failure as Data Commons/PACT, opposite institutional answer (open inputs, enclosed outputs, patent pending); zero commons/Ostrom/cybernetics vocabulary across 41 pages; independently arrived at the EIC’s problem shape; all method claims vendor-self-published and unreviewed

People & partners

People

• Tell Münzing — SINE co-founder (10/2020, Bolzano); parallel Director Transformation & Sustainability at Performance Consultants International (Whitmore’s GROW firm, since 2009) and Founder/CEO 1Transformation; ex-SustainAbility (Triple Bottom Line house, 1998–2007); source of the first-hand SINE chronology (held on the SINE Foundation page); principal behind the coaching stack, incl. the GROW-provenance note; Key Expert 1 (Data Governance) for Tender 10036847 — assessment on the consortium page; mandatory CV fields (languages, nationality, availability, references) missing from every source. Reference implementation of the person-page schema (Status block / critical judgements kept / tender assessment linked out)
• Michael Opitz — Managing Director Hanns R. Neumann Stiftung 2005–2026 (coffee-sector implementer, farmer-family livelihoods), 21 years there; 13 years MD E.D.E. Consulting; 4 years GIZ extension expert (1988–1992); MSc Agricultural Economics; proposed Team Leader for Tender 10036847 — the only candidate who documents rather than argues the ToR profile; availability unasked
• Stefan Dierks — Director Sustainability Melitta Group (since 2019, reporting to group management); 17 years Tchibo incl. Head of Corporate Responsibility; supply-chain KPI/compliance systems, biodiversity & natural capital, Omnibus/CSRD debate; BAUM e.V. board, International Bankers Forum chapter “Mittelstand & Transformation”; freelance from 09/2026, originated the Tender 10036847 constellation, fallback Team Leader — regional and DC experience absent from the profile
• Mathias Mogge — Secretary General & Vorstandsvorsitzender, Welthungerhilfe, since 2018; agricultural engineer, at WHH since 1998; live relationship via Tell Münzing — wants to learn how SINE’s approach can support WHH’s work (warrantage, climate-adaptation finance, food-security standard, transition finance); described an approach “brought from Niger to Burkina” that no source corroborates as WHH’s; adjacency = Sahel rural finance + WHH’s country network + a serious governance sceptic; his conversation notes are the outstanding item
• Ingrid-Gabriela Hoven — GIZ Managing Director since 10/2020; portfolio incl. Finance, Sector & Global Programmes, Evaluation, Europe/Med/Central Asia; ex BMZ Director-General Global Issues and Commissioner for Climate Policy and Finance (2014–2020), ex World Bank Executive Director for Germany (2010–2014); named by Tell as one of two board members driving GIZ’s finance link; the “next Vorstandsvorsitzender” claim is not publicly reported (Schäfer-Gümbel chairs since 11/2022) — held as Tell’s non-public information; no relationship path recorded
• Anna Sophie Herken — GIZ Managing Director since 08/2023; portfolio Digital Transformation & IT Solutions, Insurance, Africa, Auditing, Legal — read twice: three of SINE’s concerns in one board mandate; ex Allianz Asset Management division head and Allianz Life (USA) board, CFO HPC Germany, MD Hertie School, earlier World Bank + EBRD; Yale sustainable-finance exec ed (2022); structurally the best-matched individual on the coffee map if the insurance-first thesis holds; no relationship path recorded
• Elinor Ostrom — academic source for Data Commons’ three values; author of Governing the Commons (1990)
• Martin Kleppmann — Co-PI of CCTI at the University of Cambridge (which leads WP-A/WP-B; the proposal assigns the work packages to the institution, not to him personally); the UK/Cambridge anchor of the project; research output now on file — Man & Kleppmann (2026) on verifiable cloud emissions, incl. the transparency-log-instead-of-blockchain position
• Jonathan Heiss — TU Berlin ISE group (Stefan Tai), ZKP/blockchain researcher; first author of the 2023 verifiable-carbon-accounting paper; part of SINE inside CCTI per Martin Pompéry, but not named in the CCTI proposal draft — his work package and role are unrecorded
• Patrick Schöpflin — wiss. Mitarbeiter IÖW since 2018, Gesellschafter*in; first author of the 2026 ESRS report and sole author of the 2024 ESRS × due-diligence-law compatibility study — the one person at IÖW inside the standard text; second bridge Nature Credits (tradable “quantitatively measurable and verifiable” biodiversity units); circular-economy projects ÜBER-AUS (textile streams, Tracking) and REPAID. Closest operational contact of the three; no relationship path recorded
• Dr. Christian Lautermann — Forschungsfeldleiter Unternehmen, Wertschöpfungsketten und Konsum at IÖW, Gesellschafter; co-leads the Ranking der Nachhaltigkeitsberichte (since 1994); We Care steering committee; nearer bridge is his Corporate Digital Responsibility strand (2024 journal article with Dörr) — the one place where data itself is the object of responsibility. Half his output is alternative-economy research unrelated to SINE; no relationship path recorded
• Jonas Pentzien — co-head of the same IÖW field since 01/2024, Gesellschafter; 2019/20 Fellow at Trebor Scholz’s Institute for the Cooperative Digital Economy (The New School); AG seat in Plattform Industrie 4.0 since 2022 — the closest institutional contact point to SINE’s world in all 110 profiles; platform-cooperativism publication line 2019–2023 (Building Platforms Differently, 2023). Caveat: his 2025 output is deforestation/biodiversity policy, not platforms; and his normative answer is cooperative ownership where SINE’s is cryptographic

Partner organizations

• Einstein Center Digital Future (ECDF) — Berlin digitization research center; co-authored the 2023 DPP whitepaper
• Systemiq — “system-change” advisory/B Corp; co-author of the 2022 Data Commons whitepaper and the 2023 DPP follow-up
• Deutsche Bank — co-author of the 2022 Data Commons whitepaper; ESG/banking use case
• TLGG Consulting — co-author of the 2022 Data Commons whitepaper
• Ink & Switch UK Ltd. — CCTI sub-contractor; zk-SNARKs, zkVMs, human-agent interaction (stub)
• Light Squares Ltd. — CCTI sub-contractor; TEEs and cryptographic protocol engineering (stub)
• WBCSD — hosts/publishes PACT; profiled via its own PACT Methodology v3.0 (2025) Foreword, Scope 3/CBAM framing, 2,500+ companies engaged
• Carbon Measures — International Chamber of Commerce (45M+ companies, 170+ countries), its Carbon Measures coalition and Technical Expert Panel on Carbon Accounting; a standard-setter with a pre-committed answer (the ledger), positioned one layer above WBCSD/PACT; E-Ledgers Institute / EFI Foundation / CASI as the affiliated approaches. Known gap: the Panel’s Mission, Opportunity, and Principles statement is not in sources/
• OECD — ⟨added 2026-10-27⟩ Digital Policy Committee and the Data Governance and Privacy Unit (Clarisse Girot); publisher of the 2026 data spaces report, host of the DFFT Institutional Arrangement for Partnership and of the Global Repository on PETs named in the G7 Dec 2025 declaration. Relevant as a neutral citable authority rather than a funder or client — its reports are declassified by member governments, which is what distinguishes it from IDSA, Gaia-X or Fraunhofer publications on the same subject. Known gap: the PETs repository is not in sources/

Concepts

• Data Commons — SINE’s Ostrom-inspired governance framework for shared data ecosystems; three core values (Reciprocity, Security, Openness), architecture choice, key-questions checklist; two independent arrivals at the same architecture — the ICC Review from regulation, GEO BON from biodiversity — the second of which builds the federation and skips the governance layer entirely ⟨2026-09-24⟩
• Ostrom’s Commons Design Principles & SINE’s governance approach — the theoretical basis, and per Martin Pompéry, the operating logic SINE governs its own outputs by
• Coordination failure vs. bargaining failure — synthesis: the two failure shapes inside commons problems, the one-question diagnostic, PACT/iLEAP as coordination vs. MPC-line/CCTI as bargaining, and the proposed core description (“makes claims believable without disclosure”); objections on record; two worked cases from outside — the ICC landscape as one clean instance of each shape, plus external support for the distributive-fight objection; and ⟨2026-09-24⟩ biodiversity monitoring as the near-pure coordination control case, where the cryptographic instrument is never reached for at all; a third worked case ⟨2026-10-14⟩ — the reinsurance model layer, and the first run forwards: the page’s own rule (confidentiality machinery in use ⇒ a bargaining component exists) predicted the shape from the memo’s anonymity before its content was weighed, and holds. New there: the instrument was reached for and is a bad one — anonymity buys confidentiality by destroying attribution, so the document cannot be checked. Demand demonstrated by substitution. Plus a limit on the distinction itself: the three-step coverage withdrawal fits neither column — it is an ordinary externality, which neither a focal point nor a credible commitment touches, and the honest version of the diagnostic therefore has a third outcome (“not SINE’s problem”)
• Epistemic Integrity Certificate (EIC) — SINE’s novel CCTI primitive: agents privately commit beliefs + cooperation conditions, match securely over them, then negotiate the further evidence needed; over conjugate-family beliefs
• Cryptographic trust infrastructure (CCTI stack) — four layers: TEE/zk-SNARK/MPC/zkVM primitives → attested evidence types → zkDB query/rule layer (under an explicit kill switch) → matching, reputation, agents
• Agent identity and delegated authority — ⟨added 2026-10-26⟩ the four questions that get collapsed into one (who acted / was it allowed / who consumed what / who is answerable), and the finding that the ACN white paper files accountability under the entity level and answers it with identity plus tamper-evident logs; the one sentence in §7.4 where it escapes that frame (“accountability cannot be established solely from packet logs”) and the next paragraph where it returns to it. Carries the correction that the EIC is not an accountability primitive — it is ex-ante commitment for bargaining; the relevant CCTI objects are WP-A’s business-data attestation and recursive proofs ⟨per Martin Pompéry, 2026-10-26⟩
• Data space boundary conditions — ⟨added 2026-10-27⟩ the OECD’s four-part qualification test for when a federated data space beats bilateral contracts, centralised platforms, open APIs or PETs: three or more parties; competitive or jurisdictional relationships precluding a trusted intermediary; non-excludable collective value; ongoing flows. Plus the second-stage sector preconditions and the five domains with structural advantage. Includes a reconstruction of the mechanism comparison (Table 3.1 did not survive PDF conversion) and a first pass at scoring SINE’s own cases — PACT scores on all four, CCTI’s reinsurance case turns on condition 3
• The bootstrapping problem — ⟨added 2026-10-27⟩ why qualified data spaces still fail to start: two levels (collective infrastructure before network effects; individual onboarding before scale), the club-good structure, and the DGA neutrality rule disabling the cross-subsidisation platforms normally bootstrap with. The operative test is Frischmann’s — binding only when no subset of participants can justify the investment from their own returns — scored against five cases from Fintraffic (state anchor, 24 % measured CO₂ reduction) and ABtC (regulation raises every anchor’s return at once) to Agdatahub (bankrupt Nov 2024) and Covisint
• Data space business models and financial viability — ⟨added 2026-10-27⟩ what the field charges for and what it has failed to charge for: the marketplace model dying on three independent signals (18 %→8 % adoption, 2 of 96 charging a fee, orchestrators disowning the framing), the infrastructural-value pattern that survives, the OECD’s funding critique (instruments built for pre-competitive R&D, used for ecosystem creation; compliance measurable, viability not), the demand-side instruments almost nobody runs, and AI demand as the viability pathway — against Snowflake/Databricks at >USD 10 bn combined. Also carries the unanswered IP question: usage control never says who owns the value derived from pooled data
• Agent protocol interoperability — ⟨added 2026-10-26⟩ the O(n²)→O(n) claim made for an MCP/A2A/ACP translation gateway, read against SINE’s own five years of the same argument: O(n) holds only while the intermediate representation is semantically complete; the saving is bought with institutional centralisation; and the proposed translation path is probabilistic across the authorisation boundary. The PACT/iLEAP separation as the lived case — what survived was a mapping table, what was never claimed was semantic interoperability. Three transferable claims SINE could put into a conversation where it currently has no voice
• AI Advantage — CCTI’s measurable claim: AI-with-CCTI vs. AI-without vs. human negotiation, effect sizes, publishable either way; “cooperation-unlocking decision” as the qualitative counterpart; CoCoCo as a methodological precedent for the M4 gate ⟨2026-09-10⟩; externally prescribed ⟨2026-11-02⟩ — Capraro et al. reach the assisted-vs-unaided design independently and supply three secondary measures (calibration, abstention, ability to reject wrong advice)
• The output–process gap — ⟨added 2026-11-02⟩ similar outputs do not imply similar processes, so an evaluation that scores what a system produces leaves open how it got there and whether the competence extends. The general form of the wiki’s coherence critique, stated from outside by Capraro et al. (2026), with Loru et al.’s human/LLM credibility judgements that agree while the criteria differ as the worked case. Carries the three replacement capacities (generalization under rephrasing, appropriate suspension of judgement, adaptive reframing), the wiki’s three instances of the same error — Turing-test claims, CoCoCo, and prompt-injection hardening that left the model vulnerable to being told rather than asked — and the honest limit: processes are not observable either, the gap is narrowed by varying the input, not by opening the box. ⟨Inference⟩ which is where the EIC restates itself as a procurement problem
• Coherence as a truth substitute — ⟨added 2026-09-10⟩ what you optimise for when the claim has no ground truth: internal consistency across a large set of judgements, solved and trained on; checkable, transferable, falsifiable in one direction only; the failure mode is a loop with no exogenous anchor (coherent bias); the EIC’s problem shape reached independently by Upright, and a commitment device with no monitor
• Digital Product Passport (DPP) — EU product sustainability data standard; fragmentation problem, 4 target use cases, EU Battery Passport
• Residual-value grading — ⟨added 2026-10-16⟩ the second half of the H-AIST proposal, named in the white paper as “data model standards for information on goods, including quality-related information” and described nowhere until the Third Open Forum report called it “grading for the visualization of residual value”. Built on a plastics-recycling study of what data actually supports value transfer; two failures (end-use requirements not reaching sorters; “product quality and provenance fails to reach stakeholders on the downstream sector”), three requirements (basic transactions, upstream-downstream partnership and traceability, interoperability by reference to existing standards), target architecture downstream integration + the IEC Common Data Dictionary. This, not CVP, is the layer SINE’s work touches — claims about objects travelling to parties who cannot inspect them — and it names the willingness-to-pay directly: substantiated quality “reduc[es] uncertainty-related costs”, with the benefit landing on the party who verifies. One caution carried on the page: only the second of the two failures is confidentiality-shaped; the first is specification matching, and conflating them would be wrong about half the problem
• Circular Value-added Productivity (CVP) — ⟨added 2026-10-16; rewritten the same day when the defining equations arrived⟩ Hitachi’s proposed CE metric, being standardized via the Hitachi-AIST CECRL towards an ISO/AWI 59011 annex. It is a financial ratio: circular value added over circular expenditure, CVP = (S_CE − I_CE) / (C_CE + G_CE) — sales, intermediate-goods costs, COGS and SG&A, all standard financial-statement items, with a company-chosen “CE ratio” deciding which share of each is circular. Its diagnosis is correct and evidenced: of ~1,400 CE indicators screened down to 45, none has a cost-and-value-added focus. Eq. (3) extends it across a value network — so the metric’s boundary is not the company. The computable and believable boundary still is: the study group records that VN-wide disclosure of the numerator and denominator “is difficult” and offers only trust-based remedies. SINE’s slot is therefore sharper than first stated: not supplying the data CVP lacks, but making an additively decomposable sum checkable without disclosure — see the confidentiality gap. ⚠ Two of this page’s original claims were wrong (a DPP-style material-data numerator; a definitional company boundary) and are retained as documented corrections. ⟨extended 2026-10-16 from the primary source: worked values, Methods 1–3 and the consolidation analogy, the §3.3 third-party collector, and the authors’ own four limitations; then from the June 2025 overview: the formula’s first published form, and the finding that no generation of it contains an investment term — capital enters only as depreciation, so the “investment efficiency” language overstates what the ratio measures.⟩
• Value networks and the ISO 59000 series — ⟨added 2026-10-16, answering Tell Münzing’s question⟩ What ISO means by a value network (59004), how 59010/59020/59040 and the under-development ISO/AWI 59011 divide the work, and what the VN perspective does to SINE’s company-boundary thesis. Short version: it corrects the definitional half — CVP is defined over multiple organizations, so “the metric stops at the firm” is false — and sharpens the operative half: the standard crosses the boundary on paper and has no instrument for crossing it in practice. Four remedies are on the table (internal-only use, certification of the final value, year-on-year coarsening, a financial-institution data trustee); none makes a figure checkable without showing it. ⟨Inference: Eq. (3) is an unusually well-shaped MPC problem — additively decomposable, with its one cross-party term an amount both sides of each transaction already know.⟩
• IWIM — Integrated World Infrastructure Model — ⟨added by the 2026-11-02 lint⟩ Hitachi’s “social infrastructure intelligence foundation model designed for AI-enabled operations” (Integrated Report 2026, pp. 21–22), one of the four concepts carrying the Inspire 2027 operating model. IWIM is a safety argument before it is a product: its stated reason to exist is preventing “large-scale power outages and railway accidents resulting from incorrect AI decisions”, and its claimed asset is multilayered OT knowledge and physical causality rather than data volume. Commercially the cross-domain layer for Lumada synergies (Watt-Bit integration, wide-area mobility, physical internet; SPP/NVIDIA transmission planning at an anticipated −80% study time; Naivy inside HMAX Industry). Organizationally it is the “IWIM / AI” half of the HMAX Data Fabric on Birje’s stack slide — inside the horizontal DSS owns, not inside the HMAX products the sectors sell. ⟨Inference: the strongest provenance argument Hitachi has already made for itself — “what if the AI decides wrongly” implies “what did it know, and was that true”, which cannot be answered from the model side; same falsifier as the Hitachi Digital page, no provenance or verification vocabulary anywhere in the source⟩. Created by the lint because four pages discussed IWIM and one linked it as a concept before it had a page
• Pathfinder Network — Scope 3 carbon-data exchange; organizational/network layer (2,500+ companies, 40+ conformant solution providers); also the Data Commons whitepaper’s flagship validating use case; the external assessment (ICC 2026) — promoted as the leading exchange framework, demoted below the accounting layer; contradiction with the 2022 reading held open; ⟨2026-10-29⟩ promised-vs-built against PACT’s own 2022 figure, the architectural exclusion (no storage, no calculation, no central intermediary), and confidentiality stated without cryptography
• PACT PCF Calculation Methodology — cradle-to-gate scope, dual biogenic-CO2 PCF values, allocation/PCR hierarchies, PCF validity and exemption rules
• PACT Data Quality & Verification — Primary Data Share (PDS, mandatory now) and Data Quality Ratings (DQR, mandatory from 2027) formulas, two-horizon verification roadmap, SME grace period
• iLEAP — ⟨added 2026-10-01⟩ the logistics sibling of PACT, SINE + Smart Freight Centre (2025); ISO 14083/GLEC as borrowed methodology, own data model (TCE/TOC/HOC/ShipmentFootprint/AggregatedReport/TAD) and wire protocol; v1.1.0 decouples from PACT — standalone API mandatory, PACT integration deprecated and scheduled for removal in v2.0.0 after the PACT v2 sunset of 01.04.2026; Data Transaction 4 answers confidentiality with aggregation, not cryptography, and says so; verification and semantic correctness explicitly out of scope; conformance gate runs on act.sine.dev. Open: no Smart Freight Centre source, no adoption evidence, everything new is still draft
• Interoperability profile — ⟨added 2026-11-06, term and frame supplied by Martin Pompéry⟩ a specification whose job is to select from and constrain the standards that already exist, not to invent one or crown one as canonical. The page exists to kill a recurring wrong finding — standard X does not define Y, therefore gap — which is invalid for a profile. Four moves: selection, constraint, obligation levels, and tolerance (the one usually left out, and the one that decides whether a heterogeneous network actually interoperates — in PACT it is Postel’s law). Worked example: the Tech Specs for PCF Data Exchange v2.2, which invent almost nothing. The transferable finding is the boundary: a profile fixes representation (units, vocabularies, identifier schemes, envelopes, lineage) and cannot reach accountability (verification, conservation of quantities, audit trail) — because accountability is a question about state, and a document-moving profile has none. Also: profiles stack (Catena-X CX-0136 on PACT, iLEAP inside PACT), and “both conform to X” ≠ “their data is exchangeable” — which is where the agent-protocol page’s an architecture’s real n² count is the count of its profiles lands; a profile is a frozen cut of moving standards (the 01.04.2026 PACT v2 sunset, iLEAP’s decoupling), so its governance outweighs its content; and conformance to a profile certifies the envelope, not the method. Bears directly on the DPP harmonisation debate: identifiers stay plural, interoperability happens anyway
• Ledger-based carbon accounting — the architecture candidate: counting/inventory vs. accounting/ledger (Table 1b in full), why a PCF is not a measurement, the E-Ledgers/EFI/CASI family and the E-liability rules, chain-of-custody models left unresolved, what §6.1–6.3 call settled vs. missing; what the ledger is actually for ⟨2026-08-17⟩ — the closing constraint, and why it is the complement to cryptographic proof rather than its rival
• Verifiable Carbon Accounting — the joint reading of Heiss et al. (2023) and Man & Kleppmann (2026): the shared trusted-edge/ZK-circuit/public-commitment architecture, a seven-row comparison of the two designs, both assessed against the eight ICC criteria, the closing-constraint thesis, where they collide with SINE’s own commitments (trusted hardware vs. M5; blockchain vs. transparency log), and what is still missing
• The eight decision-grade criteria for product-level carbon data — ICC’s A–H with verbatim definitions and provenance; ⟨Inference⟩ that E/F/H do the real work and the set is engineered for a ledger; Table 3 criteria→regulation mapping; objections on record (Martin Pompéry: A/C/D/G are near-synonyms — assessed, partly upheld; and H contradicts C — checked against the text and upheld)
• Temporal resolution and the service case — ⟨Inference⟩ thesis page on two construction faults in the ICC Review, both raised by Martin Pompéry and verified against the source: Criterion H’s demand for transaction-level, near-real-time records runs against Criterion C, because period-averaging is how LCA removes temporary variance and makes figures comparable — and “any reporting window” relocates the averaging rather than avoiding it; and the word “service” appears exactly once in the document, in the ICC boilerplate, in a framework addressed to what is mostly Scope 3. Includes the constructive reformulation (decouple recording frequency from reporting frequency)
• The confidentiality gap in ledger-based carbon accounting — ⟨Inference⟩ thesis page: Criterion B states confidentiality, supplies no mechanism, and it appears in no later instrument — while transaction-level ledgers sharpen exactly that exposure; the largest unclaimed space in the ICC document, with caveats against over-reading it. Checked against SINE’s own research 2026-08-17 and upheld — B answered, C unexpectedly with it, F and G only inside a closing constraint, E untouched
• Product-level carbon regulation (CBAM, LCFS, CFMS) — the demand side: eight-dimension comparison of the three regimes (Table 2), legal bases, Australia’s facility-level Safeguard Mechanism as contrast case, and “harmonise the data layer rather than the regulatory designs that sit on it”
• Nature-related reporting frameworks (KM-GBF, SEEA EA, TNFD, NPI, NGFS) — the demand side for biodiversity data, and the deliberate counterpart to the carbon-regulation page: KM-GBF prescribes indicators (42 headline + 177 others, reported via NBSAPs/NRs), SEEA EA prescribes structure (UN global statistical standard since 2024, 41 states, five accounts, exchange values only) — two opposite theories of how comparability is achieved, drawing on the same observations; the Aichi Targets’ failure attributed in part to “the dearth of reporting measures”; the corporate/financial side (TNFD, Nature Positive Initiative, NGFS) thin and second-hand, no primary source — with the ⟨Inference⟩ that central banks on nature-related risk is the BIS/IAIS coffee finding generalised, and that the confidentiality constraint enters here and only here
• EOV) — the proposed supply side: the intermediate layer between raw observations and policy indicators, four framework families, the EBV class→subclass→specific-variable hierarchy, “space-time-biology” data cubes and the three operations (interpolation, modelling, aggregation); Table 1 transcribed in full (data dimensions per EBV subclass and EESV class) and Table 2, the crosswalk to SEEA EA — which is where the interesting failures are: abiotic ECT condition classes uncovered, a head-on semantic collision on “supply”/“use”, and exchange-values-only excluding relational value
• The NatCat model layer — ⟨added 2026-10-14⟩ the layer on which the insurance industry’s view of physical climate risk is produced, treated as a market-structure object: the industry “collectively rel[ies] on a small number of base case models” because internal models are costly and “robust external data comes at a steep price”. Three consequences — errors are correlated, not averaged (a supervisor reading convergent exposure estimates is reading one model counted many times); the layer that defines what counts as risk sits outside the regulated perimeter, which is the adtech rule-setting-layer-is-not-the-deepest-layer finding transferred; and the concentration is an affordability artifact, i.e. a function of the price of data. Plus the second mechanism that rides on it, projections given a commercial “haircut”, and the tension left unresolved: a shared base case is also how comparability is produced — the wiki’s third instance of that shape and the first where the comparability-producing artifact is a priced commercial product. Single-source; its spine is open task 14
• The insurance protection gap — ⟨added 2026-10-14⟩ reprice → restrict → withdraw, the ordered strategy by which an industry exits a risk it can no longer price, stated by its own practitioners as commercially correct rather than malicious; the 12-month renewal cycle as the mechanism that lets step 2 complete unnoticed; the gap as a deliberate negotiating position (“impending damage that insurers cannot and should not be expected to cover”) rather than a finding; the mortgage/GFC transmission claim and why the analogy is the memo’s weakest analytical move; and the narrowing of the insurance-first thesis — better information does not make an uninsurable risk insurable, it moves the line and re-prices individual assets across it
• ESG reporting in the financial sector — EBA disclosure rules and the SME primary-data gap
• Adtech as a regulatory precedent for the AI stack — the actor taxonomy as agencies converge on it (CMA/ACCC/Bundeskartellamt), the five-plus-one transferable findings (rule-setting layer ≠ deepest layer; opacity’s yield is a learning-lag rent, so the regulable quantity is change frequency; cryptographic auditability stops at completeness; measure singlehoming not share; demand concentration as countervailing power), three places the analogy breaks — including the unresolved Acemoglu et al. vs. Farboodi & Veldkamp question about whether data returns are diminishing, on which two of the transfers hang — and why the field has almost no disinterested empirics
• AI stack market structure — the AI value chain assessed against adtech property-by-property: the stickiness inversion (adtech’s rule-setting layer was also its stickiest, which is what let it hold a rent; in AI those are different layers, so the pathology has no single home), the opacity inversion (know what you pay, not what you got), the missing auction (and therefore half the transferable adtech findings have no target); the router taxonomy (only the integrated first-party router has self-preferencing structurally built in); four agencies locating the chokepoint on four layers; rent at 0–5.5 % and payment-fee-shaped, not spread-shaped; network effects denied outright by Korinek & Vipra; a capital-and-geography entry barrier that is, unlike adtech’s, measurable; and the load-bearing non-finding — no accusation of router self-preferencing exists, and no instrument to raise one either
• Binary-verifiable remedies — Witte & Krämer’s construction rule (prefer obligations checkable as a yes/no observation over ones needing an ongoing judgement) and what it rules out: non-discrimination duties, functional separation (“wooden dollars”), data siloing, log mandates; Hovenkamp’s inverted-U counterweight read as a sequencing rule; and the objection that the AI stack may have no binary remedy available where the affiliated relationship never crosses a visible API boundary — corrected 2026-08-22: one already exists in law, and at the compute layer (France’s SREN Law Art. 26–28, including a statutory reporting duty on self-preferencing), so the measurement gap is localisable rather than general
• SAFE project (GIZ) — Sustainable Agriculture for Forest Ecosystems, Team Europe Initiative on Deforestation Free Value Chains; DPI ecosystem for deforestation-free ag value chains; context for Tender 10036847
• Ecosystem financing — ⟨added to the index 2026-11-06; the page itself predates it⟩ financing the system rather than the actors inside it (PREVENT investor, via Wechsel-Geld p. 6). Why a circular business case does not close inside one balance sheet — Schwalbe as the positive instance, PreZero’s −€386m as the negative one — plus the report’s two unanswered questions (who finances the shared infrastructure, how the resulting value is distributed), the full instrument inventory from both Wechsel-Geld and the EEA, and the two instruments only POLYPROBLEM 06 names (industry offtake guarantees, an earmarked single-use levy) because they construct revenue rather than cheapen capital
• Circular ecosystem maturity stages — ⟨added 2026-11-06, on Tell Münzing’s framing⟩ SINE-authored model, built from POLYPROBLEM 06, 07 and 08 read as one set for the transition. Five stages (S0 unvalued → S1 single-actor isolated → S2 bilaterally underwritten → S3 multilaterally synchronised → S4 bookable), each anchored in quoted voices. Two claims carry the page: the stages interlock rather than sequence — in one chain actors sit on different rungs at once and the chain performs at its least mature link, which is the mechanism behind Wechsel-Geld’s “synchronisiert” correction and behind Schuck’s 0–80 % variation across stores of the same chain — and the financing gap is a mismatch between financier valuation patterns and chain maturity, not a capital shortage (VC reads S1, impact funds S1–S2, guarantors S2, MDBs S2–S3, commercial banks S4 only; nothing public covers S2→S3, which is where the valley of death sits, and Rabobank/Intesa cover it privately). Reconciles Hafenmayer and Betancourt Munoz (“capital is not scarce”) with Nusseck (“nothing flows”). Carries the temporal to-do set T1–T7 with SINE coverage marked honestly — T2–T5 ours, T1 pricing, T6 regulation and T7 behaviour execution not — and the capability-window argument: the offer is worth most across S2→S3 and nowhere else. Stress-tested against biodiversity ⟨2026-11-06⟩: the ladder transfers, and the test forces two amendments — S4 splits into demand for disclosure (TNFD/NPI/NGFS exist; PPWR and recyclate quotas are the circular equivalent) and bookability (neither domain has it), which reframes Pöppelmann’s early-mover penalty; and the S2 rung exists only where there is a bargaining component, so near-pure coordination domains run S1→S3 directly, making the presence of an S2 rung diagnostic rather than a measure of progress. Net effect on the window: it is a stage plus a bargaining component, which in biodiversity is met only at the corporate boundary. Carries the counter-finding that biodiversity’s shared layer is financed because it was claimed as a public statistical function, a move the circular layer never made
• Warrantage — ⟨added by the 2026-09-07 lint⟩ the canonical page for the instrument the coffee map calls its hinge: dual-locked warehouse credit against stored produce, Niger 1999/FAO origin, three preconditions, the Ethiopia-vs-Vietnam land-tenure argument for why the instrument exists at all, and the ⟨Inference⟩ that a verified data stream is a claim-verification technology without the padlock’s radius — with the limit stated against it (custody→verification moves risk from theft to fraud) and the WHH attribution contradiction held open

Tenders

• GIZ Tender 10036847 — consortium constellation and staffing — the bid-side companion page: who bids in which role; HRNS as consortium lead (under examination), Michael Opitz as proposed Team Leader, Tell Münzing as Key Expert 1, Dierks as fallback, KE2 open; Team-Leader profile comparison, scenarios A′/A/B/C, five formalia (incl. the mandatory GIZ prior-employment disclosure), pilot countries Uganda + possibly Vietnam, conflict of interest, blocking items as of 2026-08-17; Abgabe-Checkliste (Gap-Analyse 2026-08-17) — five missing tender annexes, eligibility gates, five CVs with per-role gaps, the 10-page concept, financial bid
• GIZ Tender 10036847 — Data Governance in Agricultural Value Chains — SAFE-project tender, farmer-centric participatory data governance + open-source data wallet prototype, 2 pilot countries, 12-month term (15.10.2026–15.10.2027); strong thematic fit with Data Commons, key staffing gap on data-wallet software engineering (WP2); binding Q&A confirms tandem Team Leader OK; official Bekanntmachung confirms extended submission deadline 21.08.2026, 12:00; all 6 tender documents + Bekanntmachung ingested; includes eligibility thresholds (turnover, references) and contract mechanics (advance payment, invoicing, VAT)

Coaching stack

Wissensgrundlage für einen später zu entwerfenden Coaching-Agenten (“Karl”-Stack): Coaching-Schulen, Fragelogiken, Vergleiche und Design-Notizen.

• Cynefin (Dave Snowden) — Fünf-Domänen-Entscheidungsframework nach Constraint-Typ (fix/regulierend/ermöglichend/abwesend); je Domäne eine andere Fragereihenfolge (sense-categorise/analyse/probe/act-respond), jetzt gegen Kurtz & Snowden (2003) und Snowden & Boone (2007) primärquellenabgeglichen — Terminologie-Drift known/knowable→simple/complicated→clear/complicated und Wertungswechsel Unordnung→aporetisches Zentrum als Hauptbefunde; Signature Moves (Ritual Dissent, SenseMaker, safe-to-fail probes, Liminalität); Kritik von Michael C. Jackson/Critical Systems Thinking
• GROW (Whitmore) und Inner Game (Gallwey) — Gallweys Self1/Self2-Dualismus und die Formel P = p − i als Haltungsfundament; Whitmores vierphasige GROW-Session-Choreographie (Goal/Reality/Options/Will) plus GROWS-Erweiterung (HAPA-basierte Selbstwirksamkeitsfragen); dünne Evidenzbasis (Müller & Kotte 2020), Scoular & Linleys „sacred cow”-Befund, ungeklärte est/Graham-Alexander-Genealogie
• CTI) — vier Cornerstones (u. a. NCRW-Postulat „people are naturally creative, resourceful and whole”), fünf gleichzeitig aktive Kontexte statt Phasenfolge, Designed Alliance als Meta-Verhandlung der Beziehung; Machtasymmetrie-Kritik am dyadischen Gleichheitsanspruch (dreiseitiger Organisationskontrakt), Individualisierungs-/Neoliberalismuskritik mit Co-Active als explizitem Beispielfall
• Heidelberger Schule und zirkuläres Fragen (Mailänder Tradition, Karl Tomm) — konstruktivistisches Fundament (von Foerster, zweiter-Ordnung-Kybernetik) und Tomms vierteilige Haltungs-/Fragetypologie (lineal/zirkulär/strategisch/reflexiv) primärquellenbasiert (Tomm 1987 Teil I+II, von Foerster 1991 vollständig gelesen); reflexive Bedeutungsverschiebung über CMM-Hierarchie-Reversal (strange/charmed loop) als Wirkmechanismus; Vierfeldertafel nur sekundär rekonstruiert (Teil III unbeschaffbar); de Shazers Wunderfrage trotz Zusatzrecherche nur schwach sekundär belegt; feministische Gewaltschutzkritik bewusst nicht recherchiert
• Angrenzende Frageschulen (Übersicht) — breite, flache Übersicht zu zehn Kontrastschulen (SFBT, Clean Language, Immunity to Change, Ontologisches Coaching, Motivational Interviewing, Appreciative Inquiry, Theory U, plus Kurzprofile Focusing/Sokratisches Gespräch/Adaptive Leadership/Action Learning/NVC); primärquellenbasiert für Ontologisches Coaching (Shabi 2015) und MI (Miller/Rollnick-Handout 2010), sonst Sekundärliteratur; markiert Grundkonflikt Zukunfts- vs. Gegenwartsorientierung gegen Cynefin
• Bateson (Nora Bateson, Gregory Bateson) — Warm-Data-Lab-Fragebewegung durch Kontexte („about” vs. „within”); Gregory Batesons Lernebenen 0–III als Theorie der Rahmen-Frage (Lernen II) mit vier Therapeutenmanövern; Symmathesy (Ablehnung von „System”/Teile-Ganzes-Denken, kein Selbst auf Lernen-III-Ebene) und Aphanipoiesis (unsichtbare Vorgeschichte jeder Hypothese); radikale Ablehnung von Messbarkeit/Objektivität als Wertsetzung; Korrektur: „pattern which connects” nicht in Steps to an Ecology of Mind verifizierbar
• Bioregioning (Indy Johar) — institutionelle Entsprechung der Lernen-II-Frage auf Zivilisationsebene: Meta-Target Validity Matrix (zehn Fragen) prüft, ob operative Ziele „civilizational optionality” erweitern statt nur Buchhaltung liefern; „viability hourglass”, Bioregioning als nicht-lineare, multigenerationale Praxis; Wertsetzung („durable, shared, viable”) als scheinbar analytisches Kriterium; keine unabhängige externe Kritik auffindbar
• Fragelogiken im Vergleich — Epistemologie, Ontologie, Axiologie — Syntheseseite über alle 18 ingestierten Schulen: Vergleichstabelle entlang der drei Annahmenschichten plus Belegstärke-Tabelle (Primärtext vs. Sekundärliteratur — die Tabelle suggeriert sonst Gleichrangigkeit, die die Quellenlage nicht deckt); acht echte Divergenzen (Ganzheit vs. Beobachterkonstrukt, Zielklarheit als Voraussetzung vs. Kategorienfehler, Absichtslosigkeit vs. Frage-als-Intervention, Konfrontationsverbot vs. Konfrontationspflicht, Problembenennung, Sprachstatus, Analyseeinheit, Wertfreiheitsanspruch); Schicht-0-Reflexivitätsfalle mit Ostroms Framework-vs-Modell-Unterscheidung und Symmetrieannahme als Argumentationskern
• Coaching-Agent — Entwurf — ENTWURF (nicht implementiert, Stand 2026-08-08) nach Tell Münzings Beschluss A+C+D: eigener fragender Agent, der seine Voreingenommenheit explizit macht; Pflicht-Repertoire (GROW/Inner Game, Co-Active, Heidelberger Schule, Cynefin) plus Adjacency-Feld; Adjacency-Prinzip (Linse ansagen, Wechsel markieren, Widerspruch stehen lassen); erweiterte Fragen-Trias in fünf Fragen inkl. ausgeschlossener Zukünfte und der inneren Verfasstheit des Fragenden; Reasoning-Stil Pfadabhängigkeit/Systeme zweiter und dritter Ordnung, ausdrücklich nicht ergodisch; acht offene Entwurfsfragen
• Annahmenregister — Vorlage — ausfüllbare Sitzungsvorlage (Datum, Thema, verwendete Schulen mit Fragen im Wortlaut, Freigelegtes, Bias des Frageformats selbst, Ungeprüftes) plus ein vollständig ausgefülltes fiktives Beispiel; der Mechanismus, der die nicht auflösbare Reflexivität sichtbar hält, statt sie in Techniken zu begraben
• Customer Zero — ⟨added 2026-10-15⟩ Hitachi’s own named doctrine: “leveraging Hitachi’s own diverse business sites as a testbed to deliver value to customers faster” (CEO message). Stated at group level and in Connective Industries and DSS; absent from Mobility. The mechanism is internal deployment → FDE Team → Data Fabric → “reusable assets” → Lumada/HMAX offering. Why it matters: it dissolves SINE’s missing-reference-customer problem (under Customer Zero, having no external reference is the expected starting condition), it fixes the shape of a first engagement (internal validation run, not a procurement pilot), and it sets the pass criterion — the output must be resellable. The caution: Customer Zero framing commits the proposal to becoming a product, so the SINE-supplies-mechanism / Hitachi-owns-asset split has to be stated early. ⟨updated 2026-11-01 from the Inspire 2027 deck: the doctrine now has a size — “Customer zero — 175 in-house manufacturing sites”, attached to HMAX Industry, confirming Connective Industries as its operational home and making it a programme with a site population rather than a concept. ⟨2026-11-01, Birje deck⟩ And it is practised without the name by the digital business: the term appears nowhere in the DSS BU’s own strategy deck, yet three of its four reference cases are Hitachi companies and the internal channel grows at +46% YoY against single digits externally — so use the shape, not the label⟩
• Offene Fragen an Tell Münzing und Martin Pompéry — laufend wachsende Liste offener Entscheidungen zum Coaching-Stack

Papers

• Data Commons whitepaper (2022) — SINE + Deutsche Bank + Systemiq + TLGG, the original framework and its three use cases
• DPP & Data Commons whitepaper (2023) — SINE + ECDF + Systemiq, applies Data Commons to DPPs
• Governing the Commons (Ostrom, 1990) — primary source behind the Data Commons values
• PACT, June 2022) — ⟨added 2026-10-29⟩ PACT’s founding document, written before the Network existed; the earliest primary PACT source here and usable as a baseline. Names SINE as PACT’s “technology partner” (McKinsey Sustainability: knowledge partner) and lists 38 core collaborators incl. Catena-X and Japan’s Green x Digital Consortium. Three load-bearing findings: Figure 4 marks only the data exchange mechanism as already specified in 2022 — identity, authentication, plausibility-checking and fraud prevention were intent, auditing was an “add-on feature” — so the ICC’s 2026 charge sheet lands on boxes PACT itself left white; the Network “cannot be used to calculate a carbon footprint or to store any individual footprint data”, which puts the ledger layer out of scope by design rather than by omission; and confidentiality is stated at full strength and answered with governance plus data minimisation, cryptography nowhere in the document — a fourth instance of open task 19, on a document SINE was technology partner for. Also the first attestation of “internet for emissions data”, and of a third pillar, Pathfinder Ecosystem, the wiki does not track
• PACT, January 2023) — ⟨added 2026-11-01⟩ the version SINE co-worked on (Martin Pompéry), 78 pages, direct predecessor of v3 and renamed “PACT Methodology” by it. The wiki’s hardest promised-vs-built evidence: v2’s Figure 17 committed to reasonable assurance from 2030 across three horizons, and v3 (2025) has two horizons with the phrase “reasonable assurance” appearing nowhere — a dated commitment measured against its own successor. Two deadlines slipped (DQR 2025 → 31.12.2027, self-attributed by v3 to its rebuild of the matrix; DQR indicators 5 → 3, scale 1–3 → 1–5) while one rule tightened (exemptions <1%/process AND <5% aggregate → <3% aggregate): a mixed direction, not a one-way weakening. Figure 19’s six “key values” replace the Vision Paper’s five design principles — Data quality and integrity drops out, Inclusivity and Reach are new. Fifth and most precise instance of open task 19: zero cryptographic vocabulary in 78 pages, confidentiality defined as adjustable access rights, one optional non-mandatory digital signature that v3 drops. SINE’s role is stable (Foreword: “our tech collaborator”; page 2 and back matter: “Technology partner”), the Disclaimer credits only McKinsey, and SINE’s own self-description here is “governance tools and open-source software” — not PETs. Collaborator circle 38 → 50. Appendix B gives the citable lineage to the Tech Specs
• PACT, April 2024) — ⟨added 2026-11-06⟩ source 3 of 3 of the PACT bundle, and the first time the wire format itself is in the wiki. 2,421 lines, native Markdown. Read, on Martin Pompéry’s framing, as an interoperability profile: it invents almost nothing and instead selects and constrains OAuth2/OIDC, a named subset of OData v4 $filter, CloudEvents in Structured Content Mode, RFC 8288 pagination, URN namespaces over GTIN/CAS/InChI, UN CPC, ISO 3166, SI units — with Postel’s law as the governing principle (a mandatory floor of HTTPS and four actions, liberal acceptance above it: “should strive to accept an event”, best-effort filtering, HTTP 202 partial results, mandatory retry with “should not rely on this behavior”). Both terms are Martin’s; neither occurs in the document. Five load-bearing findings: §7.2.1 answers the Vision Paper’s five unbuilt trust pillars by scoping them out — authentication-as-access-decision, credentials management and audit logging belong to “solutions”, not to the network; the complete security inventory is TLS + OAuth2 + Bearer + optional OIDC, which makes this the sharpest instance of open task 19 precisely because a profile could have picked up JWS as cheaply as it picked up OData; SINE is inside the specification (public-dev@pathfinder.sine.dev, every normative example against api.pathfinder.sine.dev, server: Pathfinder) though authorship is nowhere stated in it; §6’s lifecycle is a full provenance graph made of self-assertions — the slots a proof would occupy already exist as claims; and four live divergences with the rest of the stack (DQI five indicators on 1–3 vs v3’s three on 1–5; dqi mandatory for reporting periods including 2025 while v3 defers DQR to 31.12.2027; exemptedEmissionsPercent cap to be removed in v3 while the methodology tightened to <3 %; assurance.level: "reasonable" alive in the schema and absent from v3). Licence is WBCSD’s own, New York law: implementations commercial, the document non-commercially shareable only — against the Vision Paper’s “all knowledge and technical specifications will be publicly available”
• PACT Methodology v3.0 (WBCSD, 2025) — primary WBCSD source (Foreword by Dominic Waughray); cradle-to-gate PCF calculation, PDS/DQR, verification roadmap, PACT Network
• The Carbon Accounting Landscape Review (ICC, 2026) — landscape/gap analysis by the ICC Technical Expert Panel Secretariat, written on the openly stated premise that a ledger-based architecture is the answer; the first external assessment of PACT in this wiki (“emerging as the leading interoperability and data-exchange framework” but “exchanges data rather than accountable quantities”); three findings that bear on SINE, incl. the confidentiality gap and allocation-as-market-design
• iLEAP Technical Specifications v1.1.0-draft.2 (2026) — the first primary iLEAP source in the wiki, retrieved 2026-10-01 from specs.ileap.global; a wire protocol and data model, not a methodology; five load-bearing findings (decoupling from PACT, Data Transaction 4 / AggregatedReport, verification and semantic correctness out of scope, TAD as SME on-ramp, SINE at the conformance chokepoint) plus the draft’s own in-line list of unresolved issues
• Agentic Communication Networks (Li et al., 2026) — ⟨added 2026-10-26⟩ arXiv white paper, eight academics across CN/FR/DE/US/IT, staking a term on a future standardisation agenda; agents as first-order network entities across three planes (entity / service / execution) and six network-visible expressions. Same premise as CCTI, different answer — identity, policy and tamper-evident logs, with ZK, MPC and reasoning-attestation absent entirely. Carries the wiki’s third independent instance of a community stating the confidentiality requirement and not reaching for cryptography, and an unnamed naming game in its capability-vocabulary assumption, relevant to the CCTI foundational-games question
• Data Spaces between Co-ordination Challenges, Financial Viability and Public Policy (OECD, 2026) — ⟨added 2026-10-27, handed over by Martin Pompéry⟩ OECD Digital Economy report, declassified by the Digital Policy Committee 10 June 2026, drafted by Christian Reimsbach-Kounatze under Clarisse Girot, funded by Japan. The first committee-level external assessment of the data-space field in this wiki, and it is brutal: 235 registered initiatives, 25 operational, none scaling, against EUR 1.7 bn of EU public money — a figure the report traces to a Gaia-X AISBL portfolio written by its own board members. Four things bear on SINE. (1) Four boundary conditions that say when federated infrastructure beats bilateral contracts, platforms and PETs — the first disciplined statement in the wiki of when the Data Commons thesis does not apply. (2) It reaches the Data Commons thesis independently — “governance structure, not technology, is the binding constraint” — derived from Arrow, Williamson and four failed technology waves (EDI, SOA, APIs, control towers) rather than asserted. (3) It is the counterexample open task 19 asked for: a body that does inventory PETs (Box 3.2; §5.5; the G7 declaration of Dec 2025) and bounds rather than rejects them — “Neither alone is sufficient” — which replaces the three candidate readings with a fourth, sequencing. (4) The funding critique: the EU instruments “were simply not designed for the problem they are now expected… to solve”, and data monetisation “has not proven a viable funding model” — 2 of 96 initiatives charge a fee. Method note: Marker emptied Tables 3.1 and 3.2; both reconstructed from the prose and marked as such
• Verifiable Carbon Accounting in Supply Chains (Heiss, Oegel, Shakeri, Tai, 2023) — IEEE TSC 17(4); zk-SNARK proofs (ZoKrates) that a supplier’s reported PCF was computed correctly from signed sensor data under an audited MRV schema, without disclosing the data; four requirements, four-step architecture, measured cost (214 s setup, ~40 s operational, 602,461 Gas on-chain); the authors’ own flagged limits — recurring audit cost, per-instance non-scaling, the auditor as a double trust concentration, sensors assumed trustworthy
• Emission Impossible (Man & Kleppmann, 2026) — E-Energy ‘26; verifiable cloud-provider emissions reporting, by CCTI’s Cambridge Co-PI. Couples carbon to financial accounting so that “in order to lie about its emissions, the company would then also have to lie about its finances”; recursive proofs in o1js with no trusted setup; needs only one public hash commitment per period, “a task that can also be fulfilled by a simpler system such as a transparency log”; the single-electricity-meter assumption that “allowed us to side-step” completeness
• From Data to Decision (Kim et al., 2026) — Conservation Letters; GEO BON’s own community writing its interface to the UN reporting regimes — read as a position paper with a review’s apparatus. Proposes essential variables as the intermediate layer between observations and the KM-GBF’s 219 indicators and SEEA EA’s five accounts, and publishes the crosswalk. The first source in this wiki on biodiversity and natural capital, and the reason three things now stand: it is the near-pure coordination control case — 23 pages, no confidentiality, no cryptography, because the reporting entities are states with no rent to divide; it arrives at the Data Commons federation architecture independently and with no governance layer; and its corporate/financial section (TNFD, NPI, NGFS) is the same disclosure-ahead-of-the-data-layer shape the wiki knows from carbon. Method note: Marker destroyed the two landscape crosswalk tables; both were re-extracted from the PDF directly
• Evaluating the ability of LLMs to quantify consequences (Ojala, 2025) — Upright self-published report UPTR-202501; defines CoCoCo (consistency, explicitly not correctness), 3,200 pairwise comparisons × 11 models (Claude 3.7 Sonnet thinking 73.25 % → GPT-3.5 Turbo 30.47 %); the greenwashing-in-the-training-corpus finding and ⟨Inference⟩ the closed reporting→corpus→assessment loop; not peer-reviewed, benchmark authored by the interested party
• When prompt engineering just isn’t enough — Leena-2 (Ojala, 2026) — Upright UPTR-202601, patent pending; 14B Mistral base post-trained on the coherent solution to its own pairwise judgements, plus inference-time solving; unplanned generalisation to the AI2 Fermi benchmark; “steering models, and steering money”; ⟨Inference⟩ the loop has no exogenous error signal — where WP-A’s attested data would be the missing half. Do not cite the Meaning & Joy figures
• Sounding the Alarm (Inside Track x (Re)Insurance, 2026) — ⟨added 2026-10-14⟩ six-page anonymous memo by senior UK/European (re)insurance professionals, addressed to government, customers and investors; the first primary source in this wiki on the insurance meta-lever, named since 2026-08-15 and until now carried on a single commissioned research note. A position paper by an interested party — no data, no models, no citations, no named author — so its evidentiary weight sits in the claims that run against its authors’ interest: the three-step withdrawal, the “haircut” on NatCat projections, and the base-case model concentration. Three findings bear on SINE: the model-layer concentration is a data-market finding; it inverts the BIS/IAIS line the wiki treats as its strongest external corroboration (there the data is missing; here it exists and is shaded); and the agricultural supply chain is named, unprompted, among those the industry expects to drop. Its anonymity is itself the artifact — an industry solving SINE’s problem with the cheapest instrument available, and paying the full price in believability
• AI’s shortcomings in understanding consequences are a practical alignment problem (Ojala, 2026) — 45 prompt-injection breaks against 25 LLMs across 12 lab-sponsored scenarios (30-day embargo on technique); five-level alignment ladder; the load-bearing finding — models are hardened against being asked and vulnerable to being told, so the attack surface is facts, not instructions, which makes provenance the only defence and is the best outside argument for WP-A; the EIC commitment rhyme ⟨per Martin Pompéry⟩
• Human likeness is a misplaced benchmark for artificial intelligence (Capraro, Marcus & Quattrociocchi, 2026) — ⟨added 2026-11-02, handed over by Martin Pompéry⟩ PsyArXiv preprint of 30 Oct 2026, 10.31234/osf.io/zn5u9_v1; Milano-Bicocca / NYU / Sapienza. The wiki’s first disinterested source on AI evaluation — everything else it holds on the subject is Upright measuring its own product. Two separable moves. Evidential: of 35 papers applying the Turing-test label to LLMs (16 positive under their own criteria), only three run the standard three-party conversation and only one reports a pass — Jones & Bergen’s GPT-4.5 at 73 % of five-minute trials, conditional on a humanlike-persona prompt; of the other two, one is negative and one passes the simple test and fails the enhanced one. The largest category never holds a conversation at all. Normative: human likeness is the wrong goal even if the evidence held — it reproduces human bias, builds a manipulation surface, and runs into Brynjolfsson’s Turing trap (replacement over augmentation). Three findings bear on SINE. (1) The general statement of the output–process gap, which the wiki’s coherence critique was a special case of. (2) The institutional recommendation — “compare assisted with unaided performance” on accuracy, confidence calibration and users’ ability to detect and reject incorrect advice — is WP-D’s three-arm design prescribed independently, and hands it three secondary measures. (3) “Recognizes the limits of its competence and withholds judgment” is the EIC stated as a capacity, with the demand-side argument the wiki lacked: their own experiment found AI advice suppressed willingness to say “I don’t know”, lowered accuracy and raised confidence even when the advice was wrong and accuracy was incentivized. ⚠ Marker scrambled Table 1; figures were read against the PDF
• 24 Sep 2026) — ⟨added 2026-10-15⟩ two-page release, URL supplied by Tell Münzing the same day to close the wiki’s most load-bearing unsourced claim. The commitment is real and quotable: “all new rolling stock being equipped with proprietary HMAX sensor technology from 2027”, ETR1000 included; ≈2,500 HMAX trains today → “almost 3,500” through the order pipeline; edge compute “from partners such as NVIDIA” (IGX Thor); ≈50% of railway costs are maintenance and energy, with 15%/15% reductions claimed. Two things it does not contain: any material, origin, carbon or circularity vocabulary, and any retrofit commitment — the 2027 pledge covers new rolling stock only. ⟨Inference: proprietary means signing at the sensor is an internal engineering decision; and a 2027 fleet-wide commitment implies a specification freeze under way now⟩
• Hitachi Integrated Report 2026 — ⟨added 2026-10-15⟩ supplied by Tell Münzing as the one document joining Hitachi’s sustainability, digitalisation and organisational-transformation strategies; the first primary source in the wiki on the live Hitachi engagement, which until now ran entirely on SINE’s own work products filed into sources/ (see open task 13). Carries the Customer Zero doctrine, IWIM/Physical AI, the PLEDGES KPI set and the five-year ESG data. Read as much for its absences as its content: no Scope 3 line, no PACT/PCF/DPP vocabulary anywhere, no “Enhanced by HMAX”, and Customer Zero never claimed for Mobility — the Scope 3 and standardization absences are document-specific, both appear in the sibling Sustainability Report ⟨2026-10-16⟩
• Hitachi Sustainability Report 2026 — ⟨added 2026-10-16⟩ 163 pages, published the same day as the Integrated Report and meant to be read with it; fetched on Tell Münzing’s pointer to p. 36. Ingested so far: the standardization chapter (the rule-making apparatus and CVP), the Scope 3 inventory the Integrated Report omits (209.18 Mt-CO₂e vs 0.46 Mt Scope 1+2; Cat. 11 191.46 Mt, Cat. 1 15.09 Mt, spend-based with partial primary data), and the circular-economy chapter. The spine: Hitachi converts environmental quantities into productivity ratios wherever it governs them, each conversion individually defensible and jointly leaving no absolute quantity governed. ⚠ Conversion defect — Marker’s Markdown twin has reversed word order per line for this PDF (a defect of the PDF’s own text stream, not Marker); citations are anchored to printed page numbers and quotes come from a de-reversed reading copy. The remaining chapters (Planet in full, ESG data, human capital, governance) are an open task
• Progress of “Inspire 2027” (CEO deck to financial analysts, 27 Apr 2026) — ⟨added 2026-11-01, supplied by Tell Münzing⟩ 24 slides, Tokunaga reporting FY2025 — the first year of Inspire 2027 — and the only source here that puts all four sectors on identical measures and sizes HMAX as a business. The dashboard: Lumada revenue ratio Energy 25% / Mobility 33% / CI 43% / DSS 62%, group 40% (from 29%) against a 50% target. HMAX is ¥300bn at over 20% margin — ~7% of Lumada’s ¥4.1tn, under 3% of group revenue: the most profitable sliver, not the engine, which constrains any argument routed through HMAX rather than through Lumada. Two footnote findings do the real work: the group Lumada figures newly reflect “revenue consolidation offsets … started from this period”, and CI and DSS alone are restated “based on the new reporting segment structure” with structural-reform operations (home appliances, ATM) out of the denominator — so the two headline digital jumps are partly a redrawn boundary, and the deck never says what moved between CI and DSS. Also: “Customer zero — 175 in-house manufacturing sites”, attached to HMAX Industry; HMAX Mobility resting on over 30,000 km of signalling, railway-AI market ¥3tn (2030); EMEA the largest region at ¥2.8tn under “autonomous decentralized global management”; the growth story told entirely as Physical AI + Agentic AI, over ¥100tn by 2030, with no circular-economy, carbon, PCF or DPP vocabulary anywhere in 24 slides
• Hitachi Digital — Anand Birje (28 Oct 2026) — ⟨added 2026-11-01, supplied by Tell Münzing⟩ 12 slides from the CEO of the Digital Engineering & AI Solutions BU inside DSS, and the only source here that shows what DSS itself sells. Answers the question it was supplied to answer: the word “Lumada” never appears, nor Inspire 2027, nor Customer Zero — because Lumada is a revenue tag for investors with no product surface, because this is a BU one level below the P&L that reports a Lumada ratio, because the deck’s own benchmark is “the gap to best-in-class multiples” against IT-services peers, and because what he owns is the machinery (HMAX Data Fabric, FDE, IWIM, HARC) rather than the story. Carries the stack slide that settles the CI division of labour — physical sectors own the HMAX products and the sites, DSS owns the horizontal beneath — and the growth split (+46% YoY internal vs single-digit external on 65% of the business). Fifteen ecosystem partners incl. Claude and OpenAI, with “APIs / MCP / Orchestration” as a named stack row. ⚠ Marker’s twin inverts the two headline numbers on p. 5; quote from the rendered slide
• Hitachi DSS Sector Business Strategy — Jun Abe (10 Jun 2026) — ⟨added 2026-11-01, supplied by Tell Münzing⟩ 25 slides from the DSS sector CEO, Investor Day; the level between the group CEO’s deck and Birje’s BU, and the deck that states outright what the wiki had been inferring. p. 13 tags the technology DSS contributes into each HMAX market — digital twin (DSS) → HMAX Energy, Edge AI (DSS) → HMAX Mobility, autonomous evolution AI (DSS) → HMAX Industry — so the physical sector owns customer, asset and product while DSS supplies the mechanism. p. 16 labels the Data Fabric band “DSS Sector” and gives its job as “enable cross-sector and cross-industry data utilization”, with three value examples that are all Hitachi-to-Hitachi. p. 14 adds a Security pillar that Birje’s stack has no counterpart for. p. 18 carries the money: ¥30bn organic FY2026 whose first technology bullet is “enhance data fabric capabilities”, plus capacity for up to ¥1tn inorganic FY26–27. Numbers: revenue ¥3,068.3bn → ¥3,190.0bn plan, adj. EBITA 15.5% → 16–18% by FY27, Lumada ratio 62% → 70% plan, FY27 target raised to 75–80% from ca. 65%; GlobalLogic stand-alone $1,950m → $2,000m (+2.6%) against synergies $1,450m → $2,500m (+73%) — the external book is flat and all growth is internal. ATM/OKI integration starts Oct 2026. ⚠ Marker flattened this deck throughout; nine slides were read as rendered images — do not cite the twin without checking the slide
• Proposal for Social Implementation of the CVP (CE Association, 2026) — ⟨added 2026-10-16, on Tell Münzing’s pointer⟩ 19-page white paper by Hitachi, AIST and AIST Solutions for the Circular Economy Association’s confidentiality study group: the implementation document behind CVP, and the only source in the repo that gives its equations. Carries the indicator survey (~1,400 → ~770 → 45), the obstacle table naming VN data disclosure as the blocker, the two-track standards strategy (ISO for legal force under the WTO/TBT Agreement, WBCSD for de facto adoption), and an unresolved sequencing tension between the two tracks that this wiki represents rather than resolves. Forced two corrections to pages written earlier the same day. ⚠ Marker destroyed its Table 1; hand-restored in tables-recovered-from-pdf.md next to the twin
• CVP research report (H-AIST CE Lab, April 2026) — ⟨added 2026-10-16, supplied by Tell Münzing⟩ 12 pages, the primary definition of CVP — everything the wiki said about it before came from secondary statements. Forced no corrections: Eqs. (1)–(3) match. What it adds: the worked calculations (a single company scores 0.64; a three-company network 1.27 with intra-network purchase costs eliminated, 0.49 without, against standalone values of 0.47, 0.50 and 0.49 — so the whole value-network claim is a factor of 2.6 that exists only if the parties reconcile each other’s figures); the netting’s stated model, “the same concept as eliminating intra-group transactions when preparing consolidated financial statements”, i.e. a consolidation without a group; a five-type VN topology including the platform case; and §3.3, which names “a trusted, impartial third-party organization” that collects each participant’s figures and computes the ratio “without public disclosure to the entire VN” — the role specified, the instrument left open. Plus four limitations the authors state about their own indicator, one of which (arbitrary CE segmentation) SINE’s verifiability argument does not cover
• Driving Circularity Transitions (H-AIST CE Lab., June 2025) — ⟨added 2026-10-16, fetched on Tell Münzing’s instruction⟩ 4 pages, the earliest public statement of CVP and the document nobody could find (unlinked, not search-indexed, reachable only via the white paper’s reference list). Settles that CTI = WBCSD Circular Transition Indicators and dates the WBCSD route — “integrated into the ‘Value the Loop’ module of CTI” — to a year before the CE Association study group existed. Names the programme’s evidence base: an unpublished 2023 Yordas Group commissioned assessment, whose three findings the metric answers. Two things it shows by omission: the formula never had an investment term despite the comparative argument against CTI claiming one, and cross-sectoral benchmarking is asserted here but retracted into a precondition by 2026. Contains nothing on quality information, data models, passports, declarations or value networks — closing candidate 1 of open task 17
• Third Open Forum report (H-AIST CE Lab., July 2026) — ⟨added 2026-10-16, on Tell Münzing’s pointer⟩ ~8,100-word event report, and the most informative single source on this track — because it is where the lab describes everything that is not CVP. Three findings. (1) The missing half of the proposal, found: residual-value grading, built on a plastics-recycling study of where quality and provenance information fails to travel, aimed at the IEC Common Data Dictionary — which closes open task 17 except for the fact that no document exists. (2) Japan’s DPP position is a division of labour, not a rivalry: Europe holds upstream disclosure, Japan proposes downstream integration plus the property dictionary. (3) The route to ISO is personnel, not process — the association’s chair heads Japan’s TC 323 mirror committee, its co-founder works on ISO 59011, and the WG2 chair also coordinates standardization for METI’s Circular Partners. Plus the confidentiality gap stated by METI (“Data is a strategic asset … will require deeper trust”) and by Hitachi’s own Green DX division (“not the sheer number of data fields, but whether the information required by the next player … can be transmitted seamlessly”)
• Wider Sense, Aug 2026) — ⟨added 2026-11-04, handed over by Tell Münzing; index entry written 2026-11-05, see log⟩ 44 pages, journalistic and interview-driven, the document the Röchling Foundation Camp agenda is built on. Its framing question is why capital does not flow and its answer across all four actor chapters is not scarcity but steering — Nusseck (Circle Economy): “Es mangelt nicht an Kapital, sondern an dessen Lenkung”; Hafenmayer: “Es gibt genug Geld”. The report then corrects its own opening thesis in the sentence that is the most useful line in it for SINE (p. 23): the challenge is not only to finance the transition “sondern synchronisiert zu tun – unternehmensübergreifend und entlang der Wertschöpfungskette” — a coordination problem stated by a financing report in financing language. Carries the wiki’s two circular-finance concept pages (definition problem, valuation gap) and the single most SINE-relevant interview in any source here: Andreas Gruber (DKB) names the unbookable residual value (“In einem zirkulären Bau haben Bauteile einen langfristigen Wert, den wir als Bank anders bilanzieren müssten, aber noch nicht können”), the ten-year/front-loaded-risk mismatch, his own bank’s gap (“Wir haben keine Circularity-Strategie und kein genuines Circularity-Produkt”), and the remedy — “bessere Datengrundlagen und Standards”, “digitale Gebäuderessourcenpässe”, “Erst dann könnten diese Werte auch in Finanzierungsentscheidungen berücksichtigt werden”. Held-open contradiction: Hafenmayer rejects the whole framing (“Der Glaube an den grünen Bonus ist naiv”). Three absences, all in the camp’s own agenda document: nothing on Kartellrecht/IP/governance of joint learning, nothing on rebound effects or additionality, nothing on how cross-boundary data would be made trustworthy. ⚠ Of the three entity pages it links to, Röchling Stiftung now exists ⟨2026-11-06⟩; Circle Economy and Silke Stremlau still do not — see open tasks Nachtrag ⟨2026-11-06⟩ on Tell Münzing’s note that accounting/financial innovation and value-chain ecosystem development were under-covered: two new sections on the page. (a) Accounting and financial innovation — the report’s “Der Wert, den niemand bilanziert” (p. 32) split into an accounting problem (no ledger line, so no risk model) and an instrument problem (wrong duration, wrong collateral), with the instrument table the report never assembles: Intesa Sanpaolo €11.7bn (2023) assessed on its own EMF-built CE methodology and Rabobank’s Circular Entrepreneurship Desk, €270m — two banks each building a private valuation methodology because no public one exists; Stremlau’s blended finance/first-loss tranches and the 90-10 rule; Gruber’s Einspeisevergütung analogue. The bank-channel gap in one figure: 13% of circular firms use bank credit vs 46% of European SMEs (PREVENT 2025, p. 12). Stremlau’s taxonomy critique as three failure modes — coverage, usage, polarity (“Ich hätte eher definiert, was nicht nachhaltig ist”), “keine echte Steuerungsfunktion”. Held-open contradiction: Stremlau’s misestimated risk vs Betancourt’s “andere Art von Risiko”. (b) The value chain as the unit of ecosystem development — Schwalbe (value “entsteht in der Verbindung verschiedener Akteure”, p. 22), Stammer/TecPart (p. 24: evolution not greenfield, two time horizons, Mittelstand finances from cashflow, and “Klarheit” wanted over capital) and Betancourt’s isolated loops, read as one unit-of-account mismatch
• Wider Sense, Dez 2023) — ⟨added 2026-11-06 on Tell Münzing’s instruction “06 07 ingest”, closing half of open task 21⟩ 60 pages, the data-and-digitalisation report of the trio (06 data, 07 behaviour, 08 finance) and the operational pre-history of Wechsel-Geld. Its thesis is SINE’s own, three years early and from the plastics shop floor: “Nicht die Technologie stellt die große Hürde dar. Es ist eine Frage der Haltung”, and “eine gemeinsame Architektur anstelle von Insellösungen”. Six findings carry: the confidentiality gap attested in plastics — Claussen-Markefka (R-Cycle): firms hesitate to send production data “schon allein” because of custody, permitted use and Betriebsgeheimnisse; the first named mechanism in production anywhere in this wiki — Kugler (SKZ) on föderiertes Lernen with the Hamburg supplier Katulu, which answers the model problem and not the claim problem; the bankability line (headline p. 12, “Die Technik ist vorhanden, das Geschäftsmodell nicht”) with two instruments no finance report here lists — Helftewes (Der Grüne Punkt) asking for langfristige Abnahmegarantien der Industrie and an earmarked single-use levy for reuse logistics; a documented commons-vs-interfaces contradiction (Urff/ReFrastructure’s “gesellschaftliches Gemeingut” vs. Barthel/Vytal’s “keine gemeinsame Daten-Plattform”) whose shared defect is that neither has a business model — “Niemand verdient mit der Rücknahme”; source backing for the ABtC inference on bootstrapping — the 2023 Mehrweg-Angebotspflicht “hat … nichts geändert”; and the sharpest in-source objection to passport-first, Moesslein (Polysecure): over half the mass is shredded fragments, “Was nützt eine Datenbank, wenn ich einen Eintrag nicht sicher mit der dazugehörenden Datenbank verknüpfen kann?” Also: cirplus initiated and financed DIN SPEC 91446; KOLEKT and CleanHub monetising attestation; GIZ counting 20+ collection apps in Indonesia alone; consumer apps with “Wirkung unbekannt”. ⚠ Two caveats on the page — the source file is an unreviewed machine extraction of a two-column layout (only running prose quoted), and the report names the standard DIN SPEC 91446 on p. 13 but DIN Spec 19446 on p. 38, recorded not corrected
• 2026) — ⟨added 2026-11-05, on Tell Münzing’s instruction after the EMF/finance search⟩ 32 pages, doi:10.2800/5422329, CC BY 4.0; a synthesis briefing resting on ETC ST/ETC CE (2026) and EIB-EC (2026). Ingested for one reason: it states the wiki’s own circular-finance diagnoses from a source with no commercial interest in them — until now the definition problem rested on Circle Economy, which sells circularity assessment, and the valuation gap on a bank describing its own product gap. §3.1.1 convicts four instruments by name: the EU Taxonomy criteria “still only partially cover circular activities and make little reference to performance aspects”; financial ecolabels “are not currently designed in a way that allows investments linked to circular economy strategies to be mapped”; SFDR standards “do not fully allow the benefits of circular businesses to be quantified. Consequently, the benefits cannot be factored into financing decisions”; CSRD circular reporting is “still immature”. §3.1.2 gives the mechanism — “traditional credit assessments remain anchored in solvency metrics and physical collateral”, and finance teams “rely on tools unsuited to integrating circular business models into credit assessment”. §3.2.2 restates the bootstrapping problem for physical infrastructure: circular infrastructure “displays characteristics of public goods … require collective investment but generate dispersed benefits”, and pioneer knowledge “spills over to competitors, reducing first-mover incentives”. The numbers for an argument: +68 % = +EUR 82 bn/year, EUR 1,229 bn gap 2025–2040 (the same EIB-EC figure the Wechselgeld report carries second-hand), private circular investment flat at ≈0.8 % of GDP since 2019, circular material use rate +1.5 pp 2010–2024 against a Clean-Industrial-Deal target of doubling to 24 % by 2030, cost of inaction EUR 84.5 bn/year, EIB EUR 5.1 bn over 153 projects 2020–2024. And the finding that matters most: in 32 pages on why circular value cannot enter financing decisions there is no digital product passport, no materials passport, no ISO 59000 and no data space — the enablers chapter answers an information problem with taxes, EPR modulation and blended finance. The one line where SINE’s instrument is implied and not named: “results-based loans tied to verified waste reduction”. ⚠ Tables 2.1, 3.1 and 3.2 came out of conversion with collapsed columns; everything cited is from running prose