Bibliography

This is the bibliography of standards, regulations, papers, software, and ontologies cited across Shepard’s documentation and design notes. It is auto-generated from a canonical BibTeX file (docs/_data/references.bib), which downstream tools (Zotero, Mendeley, LaTeX) can ingest directly.

Maintenance rule. Every aidocs/* document that cites a paper, standard, or piece of software adds an entry to docs/_data/references.bib in the same pull request. The bibliography is the single source of truth for citations. See CLAUDE.md for the durable rule.

This site also publishes a CITATION.cff at the repository root, which GitHub uses to power the “Cite this repository” button. Use that for citing Shepard itself.

Standards

• Lebo, Timothy and Sahoo, Satya and McGuinness, Deborah (2013). PROV-O: The PROV Ontology. W3C. W3C Recommendation
  Provenance ontology — used by Shepard's PROV1a capture and the f(ai)²r vocabulary.
  Citation key: w3cProvO2013
• Knublauch, Holger and Kontokostas, Dimitris (2017). Shapes Constraint Language (SHACL). W3C. W3C Recommendation
  SHACL graph is Shepard's single source of truth for domain shape constraints (see SHACL changeover findings).
  Citation key: w3cShacl2017
• Cox, Simon and Little, Chris (2017). Time Ontology in OWL. W3C. W3C Recommendation
  Citation key: w3cTime2017
• Haller, Armin and Janowicz, Krzysztof and Cox, Simon and Le Phuoc, Danh and Taylor, Kerry and Lefrançois, Maxime (2017). Semantic Sensor Network Ontology (SOSA/SSN). W3C & OGC. W3C Recommendation
  SOSA/SSN is the natural ontology shape for Shepard timeseries Observations (resolves the 5-tuple gap).
  Citation key: w3cSosaSsn2017
• OGC (2012). GeoSPARQL 1.0: A Geographic Query Language for RDF Data. Open Geospatial Consortium.
  Citation key: ogcGeoSparql2012
• ISO/IEC (2023). ISO/IEC 42001:2023 --- Information technology --- Artificial intelligence --- Management system. ISO.
  AI Management System standard; PDCA-shaped. Cited in OECD AI-in-science alignment study.
  Citation key: isoIec42001_2023
• NIST (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). National Institute of Standards and Technology.
  Citation key: nistAiRmf_2023
• DIN (2018). DIN EN 9100:2018 --- Quality management systems --- Requirements for aviation, space and defence organizations. DIN / IAQG.
  Aerospace QMS standard. Frames Shepard's audit-trail requirements for MFFD manufacturing data.
  Citation key: dinEn9100
• ISO (2022). ISO 10303-242:2022 --- Industrial automation systems and integration --- Product data representation and exchange --- Part 242: Application protocol: Managed model-based 3D engineering. ISO.
  STEP AP242 --- MBE backbone targeted by Shepard's MFFD process-chain modelling.
  Citation key: isoAp242
• ISO/IEC (2021). ISO/IEC 21838-2:2021 --- Information technology --- Top-level ontologies --- Part 2: Basic Formal Ontology (BFO). ISO.
  Citation key: isoIec21838_2_2021
• CCSDS (2024). Consultative Committee for Space Data Systems --- Blue Books (Recommended Standards). CCSDS.
  Telemetry, telecommand, and data-management standards referenced for the PLUTO mission use case.
  Citation key: ccsdsBlueBooks
• Plattform Industrie 4.0 and Industrial Digital Twin Association (IDTA) (2022). Details of the Asset Administration Shell Part 1 --- The exchange of information between partners in the value chain of Industrie 4.0 (Version 3.0RC02). Federal Ministry for Economic Affairs and Climate Action (BMWK) / Plattform Industrie 4.0. Specification document
  Version 3.0RC02 produced November 2020 to May 2022 by the sub working group ``Asset Administration Shell'' of the joint working group of the Plattform Industrie 4.0 Working Group ``Reference Architectures, Standards and Norms'' and the ``Open Technology'' Working Group of IDTA. The normative metamodel specification for the Asset Administration Shell (AAS) --- the technology positioned as the digital-twin substrate in Aerospace-X (see @weberAerospaceX2025), the ZLP data-driven-intelligence strategy (@zlpDataDrivenIntelligence2025 slides 15--17, 22), and the DIVA project AAS-Typ-3-as-active-application-protocol proposal (@krebsDIVA2026). The specification defines: the AAS metamodel (submodels, submodel elements, semantic identifiers), the AASX exchange package format, access control (ABAC), and exchange-format mappings (XML, JSON, OPC UA, AutomationML, RDF). Extends the RDF/OWL representation surveyed in @aasRdf2019. Florian Krebs (DLR e.V.) is a named contributor in the AUTHORS section (confirmed from the PDF, p.~229, ``Florian Krebs, Deutsches Zentrum für Luft- und Raumfahrt e.V.''). Author list is institutional (50+ contributors); Krebs is the only DLR contributor named. ResearchGate entry: 361547453. This is the foundational specification behind the planned shepard-plugin-aas integration (see @aasRdf2019 note) and the forward-link from Krebs's ZLP standards participation into the fork's AAS surface. Superseded as a normative release by IDTA-01001-3-0 (January 2023); V3.0RC02 remains the version published by Plattform Industrie 4.0 with Krebs in the author list.
  Citation key: aasDetailsPart1V3Rc02_2022
• Hu, Vincent C. and Ferraiolo, David and Kuhn, Rick and Schnitzer, Adam and Sandlin, Kenneth and Miller, Robert and Scarfone, Karen (2014). Guide to Attribute Based Access Control (ABAC) Definition and Considerations. National Institute of Standards and Technology. NIST Special Publication
  The reference framing for attribute-based access control. Used in aidocs/platform/110 \S 3 to anchor the ABAC column of the options matrix; complements the in-tree RBAC implementation (Roles 4-bit struct, AccessType enum) with the framing under which a future ``same-institute readers only'' rule is just an attribute predicate over subject + resource.
  Citation key: nist800162Abac2014
• OWASP Foundation (2024). Authorization Cheat Sheet.
  OWASP guidance: deny-by-default, centralize authorization logic, prefer ABAC/ReBAC over hardcoded checks, defense-in-depth at multiple layers. Cited from aidocs/platform/110 \S 1 + 24 \S 3 — the C3 fail-closed default + single PermissionsService seam already satisfy the two strongest invariants.
  Citation key: owaspAuthzCheatsheet2024
• Knublauch, Holger. Form Generation using SHACL and DASH.
  The de-facto-standard SHACL form-generation vocabulary: dash:editor / dash:viewer annotations, the dash:Editor widget catalogue, dash:singleLine, and the constraint-derived default-editor scoring algorithm. Adopted (not re-minted) by aidocs/integrations/125 \S 4 as the core widget-hint layer; Shepard extends it only via custom dash:Editor individuals (shepard-ui namespace) per the spec's sanctioned extension pattern.
  Citation key: dashFormsKnublauch

Regulations

• European Parliament and Council (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). European Union.
  Article 2(6) research exemption; Article 4 AI literacy; Article 50 transparency for synthetic content (effective 2 Aug 2026).
  Citation key: euAiAct2024_1689
• European Parliament and Council (2023). Regulation (EU) 2023/1230 of the European Parliament and of the Council of 14 June 2023 on machinery. European Union.
  Article 17 instruction-manual requirements relevant to Shepard as RDM substrate for manufacturing data.
  Citation key: euMachineryRegulation2023_1230
• EASA (2024). EASA Artificial Intelligence Concept Paper --- Issue 02. European Union Aviation Safety Agency.
  Learning Assurance framework. Levels 1/2/3 of AI autonomy. The evidence-pack target Shepard publishes against.
  Citation key: easaAiConceptPaperIssue2_2024
• EASA (2023). EASA Artificial Intelligence Roadmap 2.0. European Union Aviation Safety Agency.
  Citation key: easaAiRoadmap2
• EASA (2024). Machine Learning Application Approval (MLEAP) --- Final Report. European Union Aviation Safety Agency.
  Citation key: easaMleap2024
• EASA (2012). Commission Regulation (EU) No 748/2012 (Part 21) --- Production Organisation Approval (Subpart G). European Union Aviation Safety Agency.
  Production organisation approval; non-conformance resolution requirements that frame the Shepard NCR data-model gap.
  Citation key: easaPart21G

Policy reports & position statements

• OECD (2023). Artificial Intelligence in Science: Challenges, Opportunities and the Future of Research. Organisation for Economic Co-operation and Development. DOI: 10.1787/a8d820bd-en.
  Authoritative consensus survey. Frames reproducibility as the load-bearing concern for AI in science --- exactly what Shepard's snapshot-bounded forging cycle answers.
  Citation key: oecdAiInScience2023
• OECD (2024). OECD AI Principles (2019, updated 2024). OECD.
  Five values-based principles: inclusive growth, human rights, transparency/explainability, robustness, accountability.
  Citation key: oecdAiPrinciples2024
• UNESCO (2021). Recommendation on the Ethics of Artificial Intelligence. UNESCO.
  Ten ethical principles; auditability & traceability principle directly underpins Shepard's PROV-O capture.
  Citation key: unescoEthicsAi2021
• COPE (2023). Authorship and AI tools (COPE position statement).
  De-facto consensus across Elsevier / Springer Nature / Wiley / T&F / SAGE: AI cannot be listed as author; AI use must be disclosed in Materials & Methods.
  Citation key: copePositionAuthorshipAi
• The Royal Society (2024). Science in the Age of AI.
  Royal Society 2024 project on AI in research integrity, skills, and ethics.
  Citation key: royalSocietyAiInScience2024
• Helmholtz Metadata Collaboration (2023). HMC FAIR Data Requirements. Helmholtz Association.
  Helmholtz's operational expression of FAIR; Shepard's Unhide publish path targets HKG / Unhide harvester compliance.
  Citation key: hmcFairRequirements
• European Commission (2024). Horizon Europe Programme Guide. European Commission.
  Citation key: horizonEuropeProgrammeGuide
• Deutsche Forschungsgemeinschaft (2022). Guidelines on the Handling of Research Data.
  Citation key: dfgResearchDataGuidelines
• Clean Aviation Joint Undertaking (2024). Clean Aviation Joint Undertaking.
  EU public-private partnership funding the next generation of green-aviation research. MFFD's institutional sponsor.
  Citation key: cleanAviationJu
• European Commission (2026). Code of Practice on AI-generated content. European Commission --- DG CONNECT.
  Draft Code of Practice operationalising EU AI Act Article 50.
  Citation key: euCodeOfPracticeAiContent
• Amazon Web Services (2024). Customer Carbon Footprint Tool: methodology and regional grid intensities. Amazon Web Services --- Sustainability.
  AWS-published grid carbon intensities + PUE used as the AWS-US proxy in aidocs/sustainability/00. us-east-1 (~370 gCO2eq/kWh) adopted as the conservative end of the band; us-west-2 (~80 gCO2eq/kWh) as the optimistic end.
  Citation key: awsCustomerCarbonFootprint2024
• Anthropic (2026). Prompt caching --- developer documentation.
  Source for the ~10:1 fresh-input-to-cache-read energy ratio used in aidocs/sustainability/00. Anthropic does not publish a precise per-token Wh figure for cache reads; the 10:1 ratio is derived from the documented KV-cache reuse path.
  Citation key: anthropicPromptCachingDocs2026
• Zachgo, Jochen (2025). Übergreifende forschungspolitische Ziele für die Programmorientierte Förderung Periode V (PoF V, 2028--2034). Bundesministerium für Forschung, Technologie und Raumfahrt (BMFTR). Ministerial letter to the President of the Helmholtz Association
  Two-page letter from Dr.~Jochen Zachgo (Leiter Abteilung~4 ``Hochschul- und Wissenschaftssystem; Bildungsfinanzierung'', BMFTR) to Prof.~Dr.~Otmar D. Wiestler (President, Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V., Berlin), dated 24~September~2025 (Bonn), with cc to Dr.~Martin Keller. Transmits the overarching research-policy goals for the fifth period of Programmorientierte Förderung (PoF~V, 2028--2034) to be agreed with the Helmholtz presidium. The three named overarching goals are: (a) a purely declaratory reference to the validity of the PFI (Pakt für Forschung und Innovation) research-policy goals, (b) the development of a joint sustainability strategy, and (c) the development of a joint and overarching architecture concept for the data-management infrastructure (``ein gemeinsames und übergreifendes Architekturkonzept für die Datenmanagement-Infrastruktur''). Goal (c) is the load-bearing strategic context for Shepard's HMC-aligned federation work; see aidocs/strategy/90-hmc-phase-2-positioning.md \S 1.1. PDF metadata names ``Wolf, Frank /423'' as document author (the BMFTR drafting staffer); signed by Zachgo on the printed letter. Uploaded to AI working memory 2026-05-23.
  Citation key: bmftrPofV2025
• Weber, Holger (2025). Aerospace-X: Charting a course through regulatory storms towards sustainable aerospace solutions.
  Consortium positioning deck for the BDLI-initiated, Airbus-Operations-led Aerospace-X project (the aerospace branch of the German BMWK-funded Manufacturing-X programme). Records: the regulatory-tsunami timeline (REACH~2007 through DPP~2030 and minimum recycling quotas 2036); the four pain-points motivating a sovereign aerospace dataspace (no seamless data transfer, paper-based documentation, multi-scan degradation, manual processes); the export-control and counterfeit-part traceability use cases; the project shape (consortium of 30+ companies/associations/research institutions, kickoff April~2024, 27-month duration, scope ``Supply Chain for Aircraft Manufacturing in Germany focusing fuselage, Cabin and Engine''); the four named use cases (Demand-and-capacity, CO\textsubscript2/PCF, Circularity/R-Strategies, End-to-End Quality); the technical vision (Catena-X+AAS+Eclipse Tractus-X as substrate, EDC connectors federating); and the worked E2E-quality use case for safety-critical aero-engine components citing Bergs & Ganser IMEC~2024. Primary source for aidocs/strategy/92-aerospace-x-regulatory-context.md. Uploaded to AI working memory 2026-05-23.
  Citation key: weberAerospaceX2025

Papers & pre-prints

• Wilkinson, Mark D. and Dumontier, Michel and Aalbersberg, IJsbrand Jan and Appleton, Gabrielle and Axton, Myles and Baak, Arie and Blomberg, Niklas and Boiten, Jan-Willem and da Silva Santos, Luiz Bonino and Bourne, Philip E. and others (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3(1), 160018. DOI: 10.1038/sdata.2016.18.
  Foundational FAIR-principles paper. Shepard's vision (aidocs/42) and Helmholtz HMC alignment derive from it.
  Citation key: wilkinsonFair2016
• Huerta, E. A. and Blaiszik, Ben and Brinson, L. Catherine and Bouchard, Kristofer E. and Diaz, Daniel and Doglioni, Caterina and Duarte, Javier M. and Emani, Murali and Foster, Ian and Fox, Geoffrey and others (2023). FAIR for AI: An interdisciplinary and international community building perspective. Scientific Data, 10, 487. DOI: 10.1038/s41597-023-02298-6.
  Extension of FAIR to AI artefacts (datasets, models, workflows). Cited in the CODATA / RDA position on AI in science.
  Citation key: huertaFairForAi2023
• Alder, Marko and Moerland, Erwin and Häßy, Jasper and Nagel, Björn (2020). Recent Advances in Establishing a Common Language for Aircraft Design with CPACS.
  DLR's flagship aircraft-configuration schema. Cited in the ontology catalog (aidocs/agent-findings/dlr-ontology-catalog.md).
  Citation key: cpacsCommonLanguage2020
• Samuel, Sheeba and König-Ries, Birgitta (2020). Combining P-Plan and the REPRODUCE-ME Ontology to Achieve Semantic Enrichment of Scientific Experiments. arXiv: 2006.12117.
  Citation key: mlPipelinesFairProvenance2020
• Bader, Sebastian R. and Maleshkova, Maria (2019). The Semantic Asset Administration Shell. arXiv: 1909.00690.
  RDF/OWL representation of AAS submodels --- referenced by the planned shepard-plugin-aas design.
  Citation key: aasRdf2019
• Luccioni, Sasha and Jernite, Yacine and Strubell, Emma (2024). Power Hungry Processing: Watts Driving the Cost of AI Deployment?.
  Per-token / per-inference energy measurements across decoder-only frontier LLMs. Comparable-class proxy used in aidocs/sustainability/00 for Claude Opus 4.7 inference accounting (Anthropic does not publish per-token Wh).
  Citation key: luccioniPowerHungry2024
• Salvadores, Manuel and Alexander, Paul R. and Musen, Mark A. and Noy, Natalya F. (2013). BioPortal: Enhanced Functionality via New Web Services from the National Center for Biomedical Ontology to Access and Use Ontologies in Software Applications. Semantic Web, 4(3), 277--284.
  Ontology-browser usage study. Cited in docs/reference/ontology-term-search and frontend/utils/ontologyTermTree as evidence for the namespace-derived two-level hierarchy fallback used by the L4 term-search surface when no explicit broader/narrower edges are available.
  Citation key: salvadoresBioportalUsage2013
• Poveda-Villalón, María and Espinoza-Arias, Paola and Garijo, Daniel and Corcho, Oscar (2020). Coming to Terms with FAIR Ontologies.
  FAIR vocabulary / ontology URI-design guidance (hash vs. slash namespaces). Cited in docs/reference/ontology-term-search and frontend/utils/ontologyTermTree for the IRI-split convention behind the L4 namespace tree.
  Citation key: poveda2020fairUriDesign
• Solanki, Dhwani and Ciuciu-Kiss, Jenifer Tabita and Qui\~nones, Nelson and Ravinder, Rohitha and Venkatesh, Suhasini and Rebholz-Schuhmann, Dietrich and Garijo, Daniel and Castro, Leyla-Jael (2025). FAIR4ML, a vocabulary to describe Machine/Deep Learning models. DOI: 10.5281/zenodo.16735334.
  CoRDI 2025 Aachen; documents v0.1.0 design rationale, adopter list (InesData, MLentory), and next-release roadmap (provenance, metrics, hyperparameters — gap that f(ai)²r could fill).
  Citation key: solankiFair4mlCordi2025
• Akhtar, Mubashara and Benjelloun, Omar and Conforti, Costanza and Gijsbers, Pieter and Giner-Miguelez, Joan and Jain, Nitisha and Kuchnik, Michael and Lhoest, Quentin and Marcenac, Pierre and Maskey, Manil and Mattson, Peter and Oala, Luis and Ruyssen, Pierre and Shinde, Rajat and Simperl, Elena and Thomas, Goeffry and Tykhonov, Slava and Vanschoren, Joaquin and Vogler, Steffen and Wu, Carole-Jean (2024). Croissant: A Metadata Format for ML-Ready Datasets. DOI: 10.1145/3650203.3663326.
  MLCommons-led ML-dataset metadata format; schema.org extension. Adopted by Hugging Face, Kaggle, OpenML, Google Dataset Search, TFDS. FAIR4ML uses cr:Dataset for trainedOn/validatedOn/testedOn/evaluatedOn references.
  Citation key: akhtarCroissant2024
• Mukherjee, Pritam and Ghosh, Rabindra K. and Mukherjee, Sourav and Lock, Mary and Bashir, Mustafa R. (2025). A practical framework for appropriate implementation and review of artificial intelligence (FAIR-AI) in healthcare. npj Digital Medicine. DOI: 10.1038/s41746-025-01900-y.
  Operationalises FAIR-for-AI as a clinical-AI evaluation checklist (pre- and post-implementation). Distinct from the RDA FAIR4ML vocabulary track --- a deployment-framework derivative of Huerta et al. 2023.
  Citation key: mukherjeeFairAi2025
• Krebs, Florian and Willmeroth, Mark and Haase, Tobias and Kaufmann, Patrick and Glück, Roland and Deden, Dennis and Brandt, Lars and Mayer, Manuel (2021). Systematische Erfassung, Verwaltung und Nutzung von Daten aus Experimenten. DOI: 10.25967/550315.
  DOI 10.25967/550315 confirmed via ORCID 0000-0001-6033-801X record (retrieved 2026-05-23 from \urlhttps://pub.orcid.org/v3.0/0000-0001-6033-801X/record). Load-bearing peer-reviewed publication of record for Shepard's architecture. Krebs is lead author; the four Zenodo software-citation authors (Haase, Glück, Kaufmann, Willmeroth on @haaseShepard2021) are co-authors here. The paper describes the polyglot-persistence architecture (Neo4j + InfluxDB + MongoDB), the Collection/DataObject/Container/Reference data model citing the RDA Research Data Collections WG recommendation, the JWT/OAuth2/API-key authentication model, the OPC-Router + Data-Reference-Generator (DRG) shop-floor ingest stack, and the thermoplastic AFP use case at the ZLP Fiber Placement Cell (FPZ). The §5 outlook anticipates git integration, CAD/PDM integration, search/filter expansion, and access-rights management --- all of which the fork has since shipped or designed. Contact address florian.krebs@dlr.de. See aidocs/strategy/86 §5 and aidocs/98 §1.5.
  Citation key: krebsDlrk2021
• Audet, Julien-Mathieu and Fortin, Yves and Côté, Gabriel and Vistein, Michael and Brandt, Lars and Krebs, Florian and Monsarrat, Bruno (2024). Iterative Robot Calibration with Accuracy Saturation Termination and Application for AFP Processing of a Thermoplastic Aerostructure with Real-time Elastic Path Compensation.
  Manuscript 2640. Krebs co-authored (DLR side: Vistein, Brandt, Krebs; NRC Canada lead: Audet, Fortin, Côté, Monsarrat). The paper introduces the ICAST (Iterative robot Calibration with Accuracy Saturation Termination) algorithm and applies it to in-situ elastic calibration of the DLR ZLP T-AFP robotic workcell --- a KUKA KR 120 HA on a rail system with an AFPT MTLH tape-laying machine and 6 kW Laserline LDM laser. Demonstrator: an 8-section thermoplastic rear pressure dome for next-generation aircraft. Calibration reduced from 31 to 17 poses with unaffected accuracy (0.057--0.064~mm side deviation). The cell is the same FPC described in @krebsDlrk2021 §3. Peer-reviewed publication evidence for the FPC research output. See aidocs/strategy/87 §3.x.
  Citation key: audetIcra2024
• Petsch, Michael and Kohlgrüber, Dieter (2018). PANDORA --- A python based framework for modelling and structural sizing of transport aircraft.
  PANDORA aircraft-sizing framework at DLR --- referenced as a sidebar in aidocs/strategy/86 §4.4 (iDMS bridged to PANDORA via the idms_pandora_converter, but PANDORA is NOT in the Shepard lineage).
  Citation key: petschPandora2018
• Rowley, Jennifer (2007). The Wisdom Hierarchy: Representations of the DIKW Hierarchy. Journal of Information and Communication Science, 33(2), 163--180. DOI: 10.1177/0165551506070706.
  Canonical DIKW (Data $\to$ Information $\to$ Knowledge $\to$ Wisdom) hierarchy paper. Cited on slide 3 of krebsForInfPro2026 as the theoretical scaffold for the ``data-driven'' framing that motivates Shepard's contextualisation layer.
  Citation key: rowleyDikw2007
• Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N. and Kaiser, Lukasz and Polosukhin, Illia (2017). Attention Is All You Need.
  Original Transformer paper. Cited in krebsGenKi2026 (ref [1]) as the architectural foundation of contemporary LLMs.
  Citation key: vaswani2017attention
• Hoffmann, Jordan and Borgeaud, Sebastian and Mensch, Arthur and Buchatskaya, Elena and Cai, Trevor and Rutherford, Eliza and others (2022). Training Compute-Optimal Large Language Models (Chinchilla).
  Compute-optimal scaling laws (the ``20 tokens per parameter'' Chinchilla regime). Cited in krebsGenKi2026 (ref [12]) as the constraint that frames the 2026--2032 data-bottleneck argument.
  Citation key: hoffmann2022chinchilla
• Shumailov, Ilia and Shumaylov, Zakhar and Zhao, Yiren and Gal, Yarin and Papernot, Nicolas and Anderson, Ross (2024). AI models collapse when trained on recursively generated data (model collapse). Nature, 631, 755--759. DOI: 10.1038/s41586-024-07566-y.
  Empirical demonstration of degenerative learning when LLMs train on synthetic data. Cited in krebsGenKi2026 (ref [13]) as the failure mode that constrains naive ``train on the web's AI output'' strategies; relevant to Shepard's f(ai)$^2$r AI-provenance vocabulary, which makes the human/AI/synthetic split visible in the data substrate.
  Citation key: shumailov2024modelCollapse
• Schlenz, Hartmut and Bronger, Torsten and Selzer, Michael and Nestler, Britta and Riem, Leo and Enahoro, Salome (2026). Research Data Management --- A Practical Introduction. DOI: 10.5281/zenodo.18468308.
  Version 1.1 (Feb 2026); first published July 2025 (J\"ulich and Karlsruhe). CC BY-SA 4.0. Funded by DFG via NFDI4ING CADEN archetype (codes D.B.B01953, D.B.C02632). Initially presented at the international CoRDI 2025 conference at RWTH Aachen. Authors are the maintainers of JuliaBase (Bronger, FZJ), Kadi4Mat (Selzer/Nestler/Riem, KIT) and SciMesh (Bronger). German companion edition: \hrefhttps://doi.org/10.5281/zenodo.1846823910.5281/zenodo.18468239. The community-standard reference for engineering-research RDM in Germany; covers FAIR life-cycle, DFG-aligned DMP template (\S 4.1--4.6), the three free ELNs (JuliaBase / eLabFTW / Kadi4Mat), data-quality fundamentals, ELN-to-ELN data exchange via SciMesh (\S 7.1), MetaData4Ing process model (\S 7.5), Zenodo publication walk-through (\S 8.2), and ML training-data preparation (\S 8.4). Shepard positioning vs. this handbook is captured in aidocs/strategy/88-nfdi4ing-alignment.md \S 5 and the thesis-positioning \S 4.1 of aidocs/98-thesis-perspective.md.
  Citation key: schlenzRdmHandbook2026
• Bergs, Berend and Ganser, Patrick (2024). Aerospace-X --- Towards multi-tier traceability in the manufacturing of safety critical aerospace components.
  The cited IMEC 2024 conference paper from which the E2E-quality worked example on slide 13 of weberAerospaceX2025 is sourced. Establishes the three-step traceability use case (production monitoring~$\to$ root-cause analysis at MRO~$\to$ containment) as the canonical Aerospace-X demonstrator. Not directly retrieved; cited as the deck cites it.
  Citation key: bergsAerospaceXImec2024
• Haase, Tobias and Glück, Roland and Görick, Dominik and Kaufmann, Patrick and Krebs, Florian and Mayer, Monika (2023). Towards a Closed-Loop Data Collection and Processing Ecosystem. DOI: 10.25967/610241.
  Cited in aidocs/strategy/103 §4 (Cluster B'). The ecosystem-around-shepard paper -- where the ``agentic data management'' framing begins. Six-author cohort confirms the 2021 core team carried the work forward two years on. PDF available via eLib.
  Citation key: haaseClosedLoop2023
• Glück, Roland and Krebs, Florian (2023). Toward Ontology-based Production.
  Cited in aidocs/strategy/103 §4 (Cluster B'). Glück--Krebs joint pivot toward ontology-driven design --- direct intellectual predecessor of Shepard's SHACL-single-source-of-truth direction (per `feedback_shacl_single_source_of_truth.md`).
  Citation key: gluckOntologyBasedProduction2023
• Weber, Holger and Glück, Roland and Krebs, Florian (2025). Leveraging semantic interoperability in Industry 4.0. DOI: 10.25967/650173.
  Cited in aidocs/strategy/103 §4 (Cluster B', G). The Industry 4.0 framing of Shepard's ontology direction. Confirms Krebs ↔ Weber edge that links Cluster B' (Shepard arc) and Cluster G (Aerospace-X external industrial peer).
  Citation key: weberGluckSemanticInteroperability2025
• Kaufmann, Patrick and Vistein, Michael and Nieberl, Dorothea and Buchheim, Andreas (2025). Data Management For Quality Assurance in Semi-Automated Processes.
  Cited in aidocs/strategy/103 §4 (Cluster B'). First recent Shepard-team paper without Krebs as author --- proves the cohort has its own publishing voice. Cross-links to Cluster A via Nieberl (AZIMUT 2013--14 with Nuschele).
  Citation key: kaufmannQaDataManagement2025
• Glück, Roland and Vinot, Mathieu and Unger, Nicolas and Kamble, P. and Toso, Nathalie (2023). Project MEMAS: ontology-based storage system for manufacturing and simulation data.
  Cited in aidocs/strategy/103 §4 (Cluster B'). Sister project at ZLP on ontology-based storage. Useful contrast point: MEMAS chose a different substrate split; cross-links Glück to the Toso/Vinot composite-simulation group around Voggenreiter.
  Citation key: gluckMemas2023
• Weber, Holger and Kaufmann, Patrick and Vistein, Michael (2023). Low-Cost Synchronization Techniques for KUKA Robots and External Axes. DOI: 10.5220/0012207900003543.
  Cited in aidocs/strategy/103 §4 (Cluster G). Weber + Kaufmann + Vistein triangle predating the 2025 semantic-interoperability paper.
  Citation key: weberKaufmannKukaSync2023
• Babaei Giglou, Hamed and Auer, Sören (2025). OntoAligner: A Comprehensive Modular and Robust Python Toolkit for Ontology Alignment. DOI: 10.1007/978-3-031-94578-6_10.
  Cited in aidocs/strategy/103 §4 (Cluster I). Closest external work to Shepard's ontology layer --- LLM-driven ontology alignment, open-source Python toolkit, from TIB Hannover (Auer's group). Worth direct outreach for thesis defence.
  Citation key: babaeiGiglouOntoAligner2025
• Bronger, Torsten (2024). SciMesh: RDF graphs for scientific knowledge.
  Cited in aidocs/strategy/103 §4 (Cluster C). The closest external precedent for Shepard's RDF backbone. JuliaBase + SciMesh combined ambition = data-store + RDF backbone = exact peer of Shepard's design. Bronger ORCID 0000-0002-5174-6684.
  Citation key: brongerSciMesh2024
• Brandt, Nico and Griem, Leon and Herrmann, Christoph and Schoof, Ephraim and Tosato, Giovanna and Zhao, Yinghan and Zschumme, Philipp and Selzer, Michael (2021). Kadi4Mat: A Research Data Infrastructure for Materials Science. Data Science Journal. DOI: 10.5334/dsj-2021-008.
  Cited in aidocs/strategy/103 §4 (Cluster C). The Kadi4Mat foundational paper. Sibling project to Shepard at KIT. Selzer is lead developer; Nestler is chair-holder.
  Citation key: brandtKadi4Mat2021
• Selzer, Michael and Nestler, Britta and others (2025). ML-extendable framework for multiphysics-multiscale simulation workflow and data management using Kadi4Mat. Scientific Data. DOI: 10.1038/s41597-025-05027-3.
  Cited in aidocs/strategy/103 §4 (Cluster C). Kadi4Mat 2025 ML-extension paper. Shepard's ML-extensibility framing should cite or contrast with this.
  Citation key: selzerKadi4MatMlExtendable2025
• Mössinger, Ines and Raps, Lukas and Fricke, Daniel and Freund, Jonathan and Löbbecke, Miriam and Chadwick, Ashley R. (2024). Characteristics of in-situ automated fiber placement carbon-fiber-reinforced low-melt polyaryl ether ketone laminates part 1: Manufacturing influences. Journal of Composite Materials. DOI: 10.1177/00219983241244882.
  Cited in aidocs/strategy/103 §4 (Cluster F). The Raps + Chadwick AFP / LM-PAEK paper that is data-adjacent to the MFFD use case Shepard demos. Krebs has no published co-authorship with this group --- defence-fragile claim documented in aidocs/strategy/103 §7.4 and §8.
  Citation key: mossingerRapsAfpLmpaek2024
• Audet, Julien-Mathieu and Fortin, Yves and Côté, Gabriel and Vistein, Michael and Brandt, Lars and Krebs, Florian and Monsarrat, Bruno (2024). In-situ Elastic Calibration of Robots: Minimally-Invasive Technology, Cover-Based Pose Search and Aerospace Case Studies (vol 89, 102743, 2024) --- corrigendum. Robotics and Computer-Integrated Manufacturing. DOI: 10.1016/j.rcim.2024.102773.
  Corrigendum / erratum to @audetIrecRcim2024 (10.1016/j.rcim.2024.102743). Surfaced via ORCID 0000-0001-6033-801X record retrieved 2026-05-23. Confirms the WOS-indexed canonical form of the Audet/Krebs in-situ elastic calibration paper. Same authorship and ZLP FPC cell as the original.
  Citation key: audetIrecRcim2024Erratum
• Jejkal, Thomas and Pfeil, Andreas and Schweikert, Jan and Pirogov, Anton and Barranco, Pedro Videgain and Krebs, Florian and Koch, Christian and Günther, Gerrit and Curdt, Constanze and Weinelt, Martin (2022). Realizing FAIR Digital Objects for the German Helmholtz Association of Research Centres. Research Ideas and Outcomes, 8, e94758. DOI: 10.3897/rio.8.e94758.
  Author list verified 2026-05-23 against \urlhttps://riojournal.com/article/94758/list/authors/ (canonical: Jejkal T, Pfeil A, Schweikert J, Pirogov A, Barranco PV, Krebs F, Koch C, Guenther G, Curdt C, Weinelt M). An earlier ORCID-anchored draft of this entry carried an over-broad author list (HMC working-group contributors); corrected to the canonical paper authors. Lead author Thomas Jejkal is at KIT-SCC and is the HMC PID-services lead. Krebs is 6th-named on a 10-author paper. This is the defining anchor of \emphCluster J (HMC technical cohort): the strongest single piece of peer-reviewed evidence that Krebs's HMC Phase 2 PID work-package commitments (@hmcPhase2WpKrebs2025 WP-1) sit on pre-existing peer-reviewed working relationships --- not on a green-field promise. Cited in aidocs/strategy/103 §3 Cluster J, §4 narrative, §6 publications table; aidocs/strategy/104 §3 paragraph and §5.1 peer-reviewed list. The Helmholtz Kernel Information Profile (KIP) for FAIR Digital Objects is articulated here, backed by Handle-system PIDs and machine-actionable PID Records.
  Citation key: krebsFairDigitalObjectsHelmholtz2022
• Brandt, Lars and Deden, Dennis and Bremen, Christian and Hassan, A. and Krebs, Florian and Stefani, Tobias and Glück, Roland and Kupke, Michael (2020). Automated manufacturing of large composites utilizing a process orchestration system. Procedia Manufacturing. DOI: 10.1016/j.promfg.2020.10.066.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Procedia Manufacturing paper on the ZLP process-orchestration system --- direct ancestor lineage to Shepard's automation surface. Confirms Brandt, Deden, Glück, Kupke as recurring co-authors with Krebs across the 2020--2024 cell-engineering arc. Author list synthesised from publisher record; ORCID lists only the title.
  Citation key: krebsProcessOrchestrationPromfg2020
• Krebs, Florian and others (2020). Development of a Lean Production Process for a Thermoplastic Composite Upper Stage Propellant Tank.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Crosses Krebs's MFFD/CFRP cell expertise with upper-stage propellant tank application --- a precursor to the LUMEN engine demo's composite-pressure-vessel ergonomics. No DOI in the ORCID record; primary access via DLR eLib 140218. Conference: SAMPE Europe 2020.
  Citation key: krebsSampeTpsTank2020
• Krebs, Florian and Larsen, Lars and Braun, G. and Dudenhausen, Wolfgang (2019). Wrinkle and boundary detection of fiber products in robotic composites manufacturing. Assembly Automation. DOI: 10.1108/aa-05-2019-0094.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Vision-based defect detection at the ZLP FPC --- direct predecessor for the ADR (Automated Defect Recognition) capability listed in @krebsInventur2025 slide-22 technology inventory. Establishes Krebs's vision-detection lineage that the planned image-region-annotation feature continues.
  Citation key: krebsAssemblyAutomationWrinkle2019
• Krebs, Florian and others (2019). Conversion of a PLC-Based Gripper System into a Modular, Cross-Linked Production Unit with Innovative Control System.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). CANCOM 2019 paper on PLC modernisation --- the predecessor-systems arc captured in aidocs/strategy/86 (PLC-step-sequence into software-orchestrator lineage). No DOI; eLib 131179.
  Citation key: krebsCancomGripper2019
• Krebs, Florian (2019). Introduction to Smart Manufacturing.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Krebs-led introductory talk at CANCOM 2019 on smart-manufacturing framing. Companion to @krebsCancomGripper2019. No DOI; eLib 131178.
  Citation key: krebsCancomSmartMfg2019
• Krebs, Florian and Larsen, Lars and Wilhelm, F. and Braun, G. and others (2019). Model-Based Elastic Deformation Compensation for a Multi-Robot Work Cell.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). ICCA 2019 paper on elastic-deformation compensation for the multi-robot work-cell --- foundational work for the IREC (in-situ robot elastic calibration) line later published in @audetIrecRcim2024. WOS-indexed; ORCID does not list a DOI.
  Citation key: krebsIccaMultirobotDeformation2019
• Krebs, Florian and Larsen, Lars and Braun, G. and others (2018). Towards plug and work OPC UA as middleware of modern automation systems.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Early articulation of OPC UA as the integration substrate that later anchors Shepard's shopfloor-ingest stack (OPC-Router + DRG, per @krebsDlrk2021). ORCID provides no external identifier; venue-only citation.
  Citation key: krebsIsrOpcUa2018
• Krebs, Florian and Larsen, Lars and Braun, G. and Dudenhausen, Wolfgang (2018). Vision-based deformation and wrinkle detection for semi-finished fiber products on curved surfaces. 2018 IEEE 14th International Conference on Automation Science and Engineering (CASE).
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). CASE 2018 precursor to @krebsAssemblyAutomationWrinkle2019 (the journal-form follow-up). WOS-indexed; no DOI in ORCID record.
  Citation key: krebsCaseVisionWrinkle2018
• Tu\vrek, Vit and Bernhard, F. and Krebs, Florian and others (2017). Inverse Model Command Shaper for a Flexible Gantry Robot. DOI: 10.1109/IROS.2017.8205973.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). IROS 2017 paper on flexible-gantry control. Same paper appears twice in ORCID (journal-article + conference-paper variants); both DOIs map here. Establishes the control-systems lineage that feeds the IREC line.
  Citation key: krebsIrosGantry2017
• Krebs, Florian and Larsen, Lars and Braun, G. and Dudenhausen, Wolfgang (2016). Design of a multifunctional cell for aerospace CFRP production. International Journal of Advanced Manufacturing Technology. DOI: 10.1007/s00170-014-6022-1.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). IJAMT paper documenting the MFZ (Multifunktionale Roboterzelle) at ZLP Augsburg --- the cell @krebsJungDlrMagazin140 covers in popular-science form. Krebs-led core paper. Direct architectural ancestor of the cell whose data Shepard later substrates.
  Citation key: krebsIjamtMultifunctionalCell2016
• Krebs, Florian and Larsen, Lars and Braun, G. and Dudenhausen, Wolfgang (2015). Development of a Vision-Based 6D Pose Estimation End Effector for Industrial Manipulators in Lightweight Production Environments. DOI: 10.1109/ETFA.2015.7301469.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). ETFA 2015 paper on the 6D pose-estimation end-effector. Krebs-led; vision-detection lineage anchor (the metrology stack later covered in @krebsMetrologyI42025 builds on this).
  Citation key: krebsEtfaPoseEstimation2015
• Krebs, Florian and others (2015). Kamerabasierte Bestimmung der Position und Orientierung eines Endeffektors.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). German-language companion to @krebsEtfaPoseEstimation2015 at DLRK 2015. No DOI; the DLRK Krebs-publication line eventually leads to @krebsDlrk2021 (the Shepard architecture paper of record).
  Citation key: krebsDlrkPoseEstimation2015
• Krebs, Florian and others (2015). Towards Interactive Verification of Programmable Logic Controllers using Modal Kleene Algebra and KIV. DOI: 10.1007/978-3-319-24704-5_15.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). LNCS chapter on interactive verification of PLC programs using Modal Kleene Algebra + KIV --- a less-known facet of Krebs's research arc: formal methods for shopfloor PLC verification. Companion: @krebsRamicsPlcKiv2015 (the conference-form sibling).
  Citation key: krebsLncsPlcKiv2015
• Krebs, Florian and others (2015). Towards Interactive Verification of PLC Programs using MKA and KIV.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Conference companion to @krebsLncsPlcKiv2015. RAMiCS 2015 is the venue for the LNCS post-proceedings.
  Citation key: krebsRamicsPlcKiv2015
• Krebs, Florian and others (2015). Towards Multi-functional Robot-based Automation Systems.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). ICINCO 2015 paper articulating the multifunctional-robot-system vision that the MFZ (cf. @krebsIjamtMultifunctionalCell2016) realises in hardware. No DOI in the ORCID record.
  Citation key: krebsIcimcoMultifunctional2015
• Krebs, Florian and others (2012). Automated Production of Large Preforms Based on Robot-Robot Cooperation.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Earliest robot-robot-cooperation paper for CFRP preform production. Conference: European Conference on Materials and Structures in Aerospace.
  Citation key: krebsEucomasPreforms2012
• Krebs, Florian (2012). Kostenreduzierung in der Qualitätsprüfung durch robotisch basierte zerstörungsfreie Prüfung.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Report on cost reduction in NDT via robot-based non-destructive inspection. Establishes Krebs's NDT lineage --- thread later picked up in @krebsMfzSteuerung2010's commitment to ``Integration des NDT Frameworks ($\to$ PI-NDT)''.
  Citation key: krebsRobotNdt2012
• Krebs, Florian (2011). Möglichkeiten und Herausforderungen der robotisch basierten zerstörungsfreien Prüfung.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Krebs's 2011 NDT keynote / report at the Fachtagung Carbon Composites. Earliest dated ORCID-listed work, predating Krebs's 2010 MFZ-Steuerungskonzept (which is uploaded separately as @krebsMfzSteuerung2010).
  Citation key: krebsCarbonNdt2011
• Krebs, Florian and others (2008). Relational Cognitive Structures for Intelligent Agent and Robot Control. DOI: 10.1109/ICSMC.2008.4811781.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). Earliest Krebs publication on the ORCID record (IEEE SMC 2008) --- predates his DLR employment (2009-10) per ORCID employment timeline; presumably arose from his Diplom work at Universität Augsburg (Informatik). The ``relational cognitive structures'' framing is a deep-prior signal for the ontology-first design philosophy later articulated in the Shepard fork.
  Citation key: krebsSmcRelationalCognitive2008
• Krebs, Florian and others (0000). Learning in Layered Module Architectures.
  ORCID-listed contribution (0000-0001-6033-801X, retrieved 2026-05-23). No year, no journal, no DOI on the ORCID record --- preserved as a stub for the record's completeness, not yet usable for citation. May be a draft Krebs uploaded to ORCID then never finalised. Honest-reconstruction: marked unverified.
  Citation key: krebsLearningLayeredModules
• Pang, Ruoming and Caceres, Ramon and Burrows, Mike and Chen, Zhifeng and Dave, Pratik and Germer, Nathan and Golynski, Alexander and Graney, Kevin and Kang, Nina and Kissner, Lea and Korn, Jeffrey L. and Parmar, Abhishek and Richards, Christopher D. and Wang, Mengzhi (2019). Zanzibar: Google's Consistent, Global Authorization System.
  Canonical relationship-based-access-control (ReBAC) paper; introduces the zedtoken consistency model that downstream OSS implementations (SpiceDB, OpenFGA, Permify) inherit. Cited from aidocs/platform/110 \S 3 (Option E framing) and \S 4 (migration-cost reasoning for Option C).
  Citation key: pang2019zanzibar
• Luo, Yang and Shen, Qingni and Wu, Zhonghai (2019). An Authorization Model for Cross-Enterprise Collaborations. Proceedings of ACM SIGSAC.
  Casbin authors' framing of the RBAC/ABAC/ReBAC unification model that the Casbin policy engine implements. Cited from aidocs/platform/110 \S 3 as background for the Option C row's flexibility column; Casbin is the lighter-weight alternative to a Zanzibar-style engine when org-scale consistency is not a requirement.
  Citation key: casbinLuo2019
• O'Connor, Martin J. and Gonçalves, Rafael S. and Willrett, Debra and Graybeal, John and Musen, Mark A. and others (2024). Ensuring Adherence to Standards in Experiment-Related Metadata Entered Via Spreadsheets. Scientific Data.
  CEDAR template-driven spreadsheet metadata entry with controlled-terminology cell validation and interactive error-fixing (HuBMAP deployment). Peer-reviewed precedent for the shape-spine Excel round-trip in aidocs/integrations/125 \S 6 (BTKVS-C2).
  Citation key: oconnorSpreadsheetMetadata2024
• Sundaram, Sowmya S. and others (2025). Toward total recall: Enhancing data FAIRness through AI-driven metadata standardization. GigaScience.
  CEDAR-template-guided metadata standardisation lifted dataset retrieval recall from 17.65% to 62.87% — the quantified payoff of structured, template-enforced capture cited in the Role 5 (RDM) verdict, aidocs/integrations/125 \S 7.
  Citation key: sundaramTotalRecall2025
• Karimova, Yulia (2019). Flexible metadata models and controlled vocabularies for research data description in multiple domains. Bulletin of the IEEE Technical Committee on Digital Libraries.
  Controlled-vocabulary-backed description measurably improves metadata quality and ease of value entry in multi-domain research data repositories. Cited in the Role 5 verdict, aidocs/integrations/125 \S 7.
  Citation key: karimovaFlexibleMetadata2019

Ontologies & vocabularies

• NFDI4Ing (2025). metadata4ing --- An ontology for describing the generation of research data within a scientific activity. Version 1.4.0.
  NFDI4Ing engineering-process vocabulary. Shipped in Shepard's internal semantic repository.
  Citation key: nfdi4ingMetadata4ing
• Krebs, Florian (2026). f(ai)²r --- FAIR-for-AI Research vocabulary.
  Vendor ontology extending PROV-O for AI provenance. Every AI interaction becomes a typed Activity. Vocab namespace: https://noheton.org/f-ai-r/ns#.
  Citation key: fair2rVocab
• Castro, Leyla-Jael and Garijo, Daniel and Rebholz-Schuhmann, Dietrich and Solanki, Dhwani and Ciuciu-Kiss, Jenifer Tabita and Katz, Daniel S. and Eklund, Lars and Bharathy, Gnana and Research Data Alliance FAIR4ML Task Force (2024). FAIR4ML Metadata Schema v0.1.0.
  RDA FAIR4ML Interest Group; schema.org extension for ML model metadata. Adopted by InesData and MLentory (NFDI4DataScience). Companion to Croissant (datasets), CodeMeta (code), m4i (process).
  Citation key: castroFair4ml2024
• IFOMIS / BFO Consortium (2020). Basic Formal Ontology (BFO) 2.0.
  Upper-level ontology. ISO/IEC 21838-2:2021. Underpins IAO, EMMO, NFDIcore, PMD-Co, MWO.
  Citation key: bfoOntology
• EMMC consortium (2025). EMMO --- Elementary Multiperspective Material Ontology. Version 1.0.0.
  Citation key: emmoOntology
• EMMC / NanoMECommons (2024). CHAMEO --- CHAracterisation MEthodology Ontology. Version 1.0.0.
  EMMO-aligned methodology for materials characterisation; CEN CWA 17815.
  Citation key: chameoOntology
• FIZ Karlsruhe (ISE) / NFDI (2024). NFDIcore 2.0.
  BFO-compliant mid-level ontology for NFDI consortia; mapped to schema.org.
  Citation key: nfdiCoreOntology
• Platform MaterialDigital (2025). PMD Core Ontology (PMDco) 3.0.0.
  Citation key: pmdCoreOntology
• Helmholtz Metadata Collaboration (2024). Helmholtz Digitisation Ontology (HDO).
  Mid-level ontology for digital assets/processes across Helmholtz. CC0. Direct DLR-institutional alignment.
  Citation key: helmholtzDigitisationOntology
• QUDT consortium (2024). QUDT --- Quantities, Units, Dimensions and Types. Version 2.1.
  Citation key: qudtOntology
• DCMI (2020). DCMI Metadata Terms.
  Citation key: dublinCoreTerms
• schema.org community (2024). schema.org --- structured-data vocabulary.
  Citation key: schemaOrg
• Research Object Alliance (2023). RO-Crate 1.1 Specification.
  JSON-LD-on-schema.org packaging for research outputs. Shepard's canonical Collection-export format.
  Citation key: roCrate11
• DataCite (2024). DataCite Metadata Schema 4.5.
  Citation key: dataCiteMetadataKernel45
• SEMIC / EU JRC (2023). DCAT-AP 3.0 --- Data Catalogue Vocabulary Application Profile for European data portals.
  Citation key: dcatAp3

Software

• KIT Data Manager Group (2026). NovaCrate: web-based interactive editor for RO-Crates. Version 1.11.1.
  Apache-2.0. Standalone Next.js 16 + React 19 web application; not published as an npm package (private:true). Supports RO-Crate 1.1 and 1.2. Evaluated in R4 (aidocs/data/r4-novacrate-evaluation.md) — WATCH verdict; no Vue 3 embed path available as of 2026-05-26.
  Citation key: novacrate2026
• Describo (2025). Describo Crate Builder Component (archived).
  MIT. Vue 3 component for RO-Crate metadata editing. Archived 2025-08-29; no further maintenance. Evaluated in R4 as the Vue-native alternative to NovaCrate; rejected due to archived status.
  Citation key: describo_crate_builder_archived
• DLR-shepard contributors (2025). shepard --- Storage for HEterogeneous Product And Research Data (upstream). Version 5.4.0.
  Upstream codebase that this fork tracks. Frozen API surface preserved for byte-compatibility (see CLAUDE.md API-version policy).
  Citation key: upstreamShepard
• Krebs, Florian (2026). shepard (noheton fork).
  Fork with additive /v2 surface, plugin SPI, MCP server, AI plugin, Unhide publish, S3 file storage.
  Citation key: shepardFork
• Helmholtz Metadata Collaboration (2024). Helmholtz Unhide --- Research Data Harvester.
  Helmholtz-wide harvester. shepard-plugin-unhide publishes /v2/unhide/feed.jsonld for it.
  Citation key: helmholtzUnhide
• DLR-SL (2025). CPACS --- Common Parametric Aircraft Configuration Schema.
  Citation key: cpacsSoftware
• Barrasa, Jesús (2024). neosemantics (n10s) --- RDF & Semantics toolkit for Neo4j.
  Neo4j plugin that backs Shepard's internal semantic repository.
  Citation key: n10sPlugin
• Red Hat (2025). Quarkus --- Supersonic Subatomic Java. Version 3.27.x.
  Shepard backend runtime (Java 21).
  Citation key: quarkusFramework
• Haase, Tobias and Glück, Roland and Kaufmann, Patrick and Willmeroth, Mark (2021). shepard --- storage for heterogeneous product and research data. Version 1.0.0. DOI: 10.5281/zenodo.5091604.
  Original Zenodo software citation for upstream shepard. Apache-2.0. All four named authors are co-authors with F. Krebs on the DLRK 2021 architecture paper [@krebsDlrk2021] --- the personal continuity from CUBE iDMS (Krebs-architected, 2017--2020) into the 2021 Shepard team is therefore directly attested, not just institutional. See aidocs/strategy/86 §5 for the revised continuity reading.
  Citation key: haaseShepard2021
• Haase, Tobias and Glück, Roland and Kaufmann, Patrick and Willmeroth, Mark (2021). shepard --- storage for heterogeneous product and research data (concept DOI). DOI: 10.5281/zenodo.5091603.
  Concept DOI for the Zenodo shepard record --- always resolves to the latest version (currently 5.1.2 per Zenodo API retrieved 2026-05-23). Distinct from @haaseShepard2021 (10.5281/zenodo.5091604), which is the 1.0.0 version DOI. Surfaced via ORCID 0000-0001-6033-801X record (retrieved 2026-05-23 from \urlhttps://pub.orcid.org/v3.0/0000-0001-6033-801X/record). Krebs is listed via ORCID self-attribution.
  Citation key: haaseShepardConceptDoi
• Authzed, Inc. (2024). SpiceDB: Open Source, Google Zanzibar-inspired permissions database.
  Canonical OSS Zanzibar implementation. Maintains its own consistency model (zedtokens, schema, dual-write windows). Cited from aidocs/platform/110 \S 3 (Option C) and \S 4 (migration-cost reasoning: external engine introduces a second consistency boundary Shepard does not have today).
  Citation key: authzedSpicedb2024
• Cloud Native Computing Foundation and Okta, Inc. (2024). OpenFGA: Fine-grained authorization at scale, inspired by Google Zanzibar.
  Alternative OSS Zanzibar implementation; CNCF Sandbox. DSL is more approachable than SpiceDB's schema language. Cited from aidocs/platform/110 \S 3 (Option C) as the plugin-impl target if D's AuthzProvider SPI evolves toward C.
  Citation key: openfgaZanzibar2024
• Pluskiewicz, Tomasz. shaperone: Form UI elements driven by SHACL Shapes.
  Most active OSS SHACL form renderer (web components, RDF/JS); consumes DASH editor annotations and demonstrates SPARQL-backed dash:AutoCompleteEditor. Cited from aidocs/integrations/125 \S 1.3 as the reference implementation FORM-DESCRIPTOR-1's DASH scoring should track.
  Citation key: shaperoneHypermedia

DLR-internal sources

• Welzmüller, Florian and Dannemann, Frank (2024). Research Data Management for Space Missions: Practical Experiences and Lessons Learned.
  DLR eLib 215120. Motivates the PLUTO use case and describes FAIR requirements for satellite-mission data that Shepard must meet.
  Citation key: welzmueller2024Pluto
• DLR Augsburg / ZLP (2025). Multifunctional Fuselage Demonstrator (MFFD) --- JEC World Innovation Award 2025, Aerospace --- Parts.
  Award-winning thermoplastic CFRP upper fuselage demonstrator. The reference real-world dataset informing Shepard's manufacturing-data design.
  Citation key: mffdJecAward2025
• DLR Lampoldshausen (2025). LUMEN --- Liquid Upper Stage Demonstrator Engine.
  DLR's methane / LOX research engine. Shape the source of the synthetic LUMEN showcase dataset.
  Citation key: dlrLumenProgram
• Nuschele, Stefan and Schmidt, Thomas and Kießig, Mildred and Mühlhausen, Thomas and Wagenfeld, Bastian and Voss, Hajo (2014). Datenbank PRAESTO: Speicherung von CFK-Forschungsdaten auf Fertigungsniveau.
  DLR eLib 94077. Conference paper. Predecessor system to Shepard at DLR ZLP Augsburg --- see aidocs/strategy/86. Full text not available in eLib.
  Citation key: nuschelePraesto2014paper
• Nuschele, Stefan (2014). Datenbank PRAESTO: Speicherung von CFK-Forschungsdaten auf Fertigungungsniveau.
  DLR eLib 94078. Conference talk (sister record to eLib 94077). Predecessor system to Shepard --- see aidocs/strategy/86.
  Citation key: nuschelePraesto2014talk
• European Commission CORDIS (2011). XPRESS --- Flexible Production experts for reconfigurable assembly technology. European Commission. FP6 Integrated Project factsheet
  EU 6th Framework Programme Integrated Project, call FP6-2004-NMP-NI-4. Duration 2007-01-01 to 2011-06-30. Total cost EUR 12,477,505; EU contribution EUR 7,999,324. Coordinator Harms & Wende GmbH & Co KG (Hamburg). Seventeen participants including KUKA Schweissanlagen GmbH, Airbus Deutschland GmbH, Premium Aerotec GmbH, Centro Ricerche Fiat, Fraunhofer-Gesellschaft. Load-bearing for the consortium-shape continuity claim in aidocs/strategy/86 \S 1.7 (the same KUKA + Airbus + Premium-Aerotec partner set re-appears in AZIMUT and MFFD).
  Citation key: xpressFp6Cordis
• Hochschule Karlsruhe --- Institute for Sensor and Information Systems (2011). XPRESS project page.
  Secondary source for FP6-26674 XPRESS. Project lead at HKA-ISRG: Prof. Dr. Norbert Link. Describes the ``Expertons'' approach (intelligent networked production units; 50\,% ramp-up-time-reduction target). Cited from aidocs/strategy/86 \S 1.7 as the partner-list context corroborating cordis.europa.eu/project/id/26674.
  Citation key: xpressFp6Hka
• DLR Zentrum für Leichtbauproduktionstechnologie (2014). AZIMUT --- Automatisierung zukunftsweisender industrieller Methoden und Technologien für CFK-Rümpfe.
  DLR ZLP Augsburg project page for AZIMUT. LuFo IV-3. Industrial lead Premium Aerotec; research partners DLR ZLP Augsburg + Stade with KUKA Systems as cell-integration partner. Scope: economic production of CFRP fuselage structures along the entire process chain. Cited from aidocs/strategy/86 \S 1.7 as the AZIMUT partner-shape primary source.
  Citation key: azimutDlrZlp
• Gerngroß, Tobias (2015). Projekt AZIMUT --- Automatisierungslösungen für große Faserverbundbauteile in der Luftfahrt. Ingenieurspiegel, 72--74.
  DLR eLib 110578. Public-facing summary of the AZIMUT project (DLR ZLP Augsburg, c.~2010-2013) covering automation solutions for large doubly-curved CFRP structural components in aerospace. Companion to the internal final report at eLib 94104 [@koerberAzimut2014]. Cited from aidocs/strategy/86 \S 1.7 as the AZIMUT public summary + as the analog reference for the operator-testimony AZUR sibling project.
  Citation key: gerngrossAzimut2015
• DLR Institut für Bauweisen und Strukturtechnologie (2013). AZIMUT --- Automated Manufacturing of a CFRP Rear Pressure Bulkhead.
  DLR-BT institute video documenting the AZIMUT project headline demonstrator --- automated manufacture of a CFRP rear pressure bulkhead (Druckkalotte). Funding attribution on the video page: Bundesministerium für Forschung, Technologie und Raumfahrt. Cited from aidocs/strategy/86 \S 1.7 as the visual primary source for the project scope.
  Citation key: dlrAzimutVideo
• Körber, Marian and Nieberl, Dorothea and Schmidt-Eisenlohr, Clemens and Nuschele, Stefan and Krebs, Florian and Kaps, Robert (2014). AZIMUT Abschlussbericht. Deutsches Zentrum für Luft- und Raumfahrt e.V.. DLR-Interner Bericht
  DLR eLib 94104. Internal final report for the LuFo IV-3 AZIMUT project. Full-text not publicly available (nicht ver\"offentlicht). Affiliations: Institut für Bauweisen und Strukturtechnologie + Institut für Faserverbundleichtbau und Adaptronik (Augsburg + Braunschweig). Keywords: Automatisierung, CFK, Effects-of-Defects, Door-Surround, NDT, Vakuumaufbau, flächiges Greifen. First peer-archived multi-author project-report record naming Krebs --- load-bearing for aidocs/strategy/86 \S 1.7 and aidocs/strategy/104 \S 2.
  Citation key: koerberAzimut2014
• KUKA Systems GmbH (2010). Automation for CFRP processing.
  KUKA Systems press release announcing the strategic partnership with DLR ZLP Augsburg for the Multifunktionale Fertigungszelle (MFZ). At 2026-05-23 retrieval the canonical KUKA-news URL returned an empty filter page; the press-release content is preserved in search-result summaries and corroborated by the academic record (Springer 10.1007/978-3-319-00557-7_42, ``Design of a Multifunctional Cell for Aerospace CFRP Production''). Cited from aidocs/strategy/86 \S 1.7 as the MFZ-partnership announcement timing-anchor (one month before Krebs's MFZ control-concept deck).
  Citation key: kukaDlrMfz2010
• DLR Zentrum für Leichtbauproduktionstechnologie (2009). Zweiter Spatenstich: Neues DLR-Zentrum für Leichtbauproduktionstechnologie in Augsburg.
  DLR ZLP Augsburg symbolic groundbreaking ceremony, 2009-09-07. Institutional start-of-site date. Cited from aidocs/strategy/86 \S 1.7 as the timing-anchor for Krebs's 2009-10 KUKA-Roboter \(\to\) DLR transition (one month after groundbreaking).
  Citation key: dlrZlpGroundbreaking2009
• Weyrauch, Florian (2023). ResearchGate profile --- Florian Weyrauch.
  Master of Engineering; 25 years of fibre-reinforced-plastic experience in automotive and aerospace; held the role of Project Manager at DLR Stuttgart for federally-funded aerospace projects in cooperation with the German aircraft industry, with named responsibility for the automation of CFRP-structure manufacture and the set-up of the ZLP Augsburg site; 20 publications listed; cited by 119 at retrieval 2026-05-23. Load-bearing for the Weyrauch-was-DLR-side reading in aidocs/strategy/86 \S 1.7 (corrects a small misattribution in Krebs's 2026-05-23 recall placing Weyrauch on the Airbus/KUKA side).
  Citation key: weyrauchResearchGate
• Cevotec GmbH (2023). Welcome, Florian Weyrauch!.
  Cevotec onboarding announcement for Florian Weyrauch as Team Lead Application Development. Verifies the DLR Stuttgart \(\to\) Cevotec transition; provides the biographical anchor for the DLR-era affiliation claim in aidocs/strategy/86 \S 1.7. Calendar year of move is the Cevotec announcement page's stated month (August); exact year inferred from peer evidence; treat as approximate.
  Citation key: weyrauchCevotec2023
• Krebs, Florian Benedikt (2026). ORCID public record --- 0000-0001-6033-801X.
  Public ORCID record of the author. Anchors the KUKA Roboter GmbH (Augsburg, 2008-04 to 2009-09) employment and the DLR Institute of Structures and Design (2009-10 onward) transition that are load-bearing for aidocs/strategy/86 \S 1.7 and aidocs/strategy/104 \S 1--2. ORCID iD volunteered by F. Krebs 2026-05-23 and reconciled against pub.orcid.org/v3.0/0000-0001-6033-801X/record.
  Citation key: orcidKrebs
• Krebs, Florian and Jung, Bernadette (2013). Hightech auf vier Säulen --- Glanzstück im DLR Augsburg: multifunktionale Roboteranlage für die Fertigung von Flugzeugbauteilen. DLR magazin(140).
  Public-facing institutional exposition of the MFZ (Multifunktionale Roboterzelle) at ZLP Augsburg, authored by Krebs + Jung. Describes the cell's commissioning in September 2013, its five hang-down KUKA robots on a 30~m linear axis, two portal robots (240~kg payload each), 450~m² cell floor area, and the convex-tooling capability that the bodenfreie Hang-Down-Konfiguration enables. Names Voggenreiter and Kupke as contact professors. Issue #140 is the "Roboter" themed issue; the local filename "06" is a sequence number, NOT the issue number. Useful as evidence of how DLR positioned the ZLP/MFZ externally in 2013 --- the same time-window in which Krebs's MFZ-Steuerungskonzept work (@krebsMfzSteuerung2010) had matured into operational hardware. See aidocs/strategy/87 §3.1 and aidocs/98 §1.5.
  Citation key: krebsJungDlrMagazin140
• DLR Center for Lightweight-Production-Technology (ZLP) Augsburg (2024). shepard --- storage for heterogeneous product and research data (project page).
  Official DLR ZLP project page. Notably contains no mention of predecessor systems (CUBE / iDMS / KIBID / PRAESTO) --- the absence is itself recorded in aidocs/strategy/86 §1.
  Citation key: dlrZlpShepard
• DLR Augsburg / ZLP (artefact collection) (2026). UNAS ZLP code-drop --- 102 source-code zips representing the Augsburg software ecosystem (PRAESTO/KIBID/iDMS/Shepard era).
  Primary historical source of unusual provenance: a researcher's NAS snapshot retrieved 2026-05-23 containing PRAESTO-era client code (kibid_exporter, opcua_kibid_adapter, c.~2017), the full seven-component iDMS project (Java/Python/Jupyter clients, Flask frontend, Kafka bridge, RCE integration, importers; c.~2020), and adjacent ZLP tooling. Establishes the existence and shape of internal systems not represented in eLib. See methodological discussion in aidocs/strategy/86-shepard-predecessor-systems.md.
  Citation key: unasZlpCodeDrop2026
• Krebs, Florian (n.d.). Input Florian --- PRAESTO.
  Load-bearing primary source for aidocs/strategy/86 \S2.1. Names the PRAESTO supplier as Kisters AG (first primary-source identification --- prior to this artefact, \S2 carried the testimony-only claim ``supplier not named here''). Records the engagement as ``DLR seit 2010'' --- four years before the DLRK 2014 paper, moving PRAESTO's DLR-presence lower bound from 2014 to 2010. Architecture diagram on both slides scopes PRAESTO to \emphinspection-only workflows (\emphInspektionsprozess, \emphPr\"ufung Trockenfaserablage, \emphPr\"ufung Endbauteil) --- spatial-data-fusion analyser for inspection, not full process-data acquisition platform. The deck does not state a grievance or proposed adjustment; the misfit thesis remains a testimony claim, with this deck supplying the scope-evidence that makes the testimony legible. Filename 8c1fe026-Input_Florian_PRAESTO.pptx. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: krebsInputPraesto
• DLR Zentrum für Leichtbauproduktionstechnologie (ZLP) Augsburg (2026). Wissenschaftliches Kompetenzportfolio: DLR Zentrum für Leichtbauproduktionstechnologie (ZLP) Augsburg.
  Institutional competence portfolio for ZLP Augsburg. Author not identified on the document; PDF metadata shows ``Skia/PDF m147 Google Docs Renderer''. Synthesises the five research groups (Flexible Automation; Assembly & Joining; Production-Integrated Quality Assurance; Processes & Automation; Technical Centre for Production Processes), the four flagship cells (MFZ, FPC, TPC, TEC), and names Shepard as part of the centre's digital backbone. Used as primary source for aidocs/strategy/87-dlr-zlp-positioning.md. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: zlpKompetenzportfolio2026
• Krebs, Florian (2026). Projektentwurf: DIVA --- Drone Integration via AAS.
  Earliest formal artefact for the DIVA project --- a DLR-internal, research-only consortium proposing to use Asset Administration Shell (AAS) Typ~3 as an \emphactive application protocol for decentralised UAS swarm coordination, transferring AAS technologies ``aus der Lieferkette (Aerospace-X) direkt in den operativen Flugbetrieb''. Names a five-institute DLR consortium: FT (Flugsystemtechnik, system architecture + UTM interface), KI (KI-Sicherheit, certifiable multi-agent AI), DW (Datenwissenschaften, AAS data modelling), MO (Instandhaltung und Modifikation, predictive maintenance + digital product passport), and BT/ZLP (Bauweisen und Strukturtechnologie, AAS-Repository-Infrastruktur \textscShepard). Four work-packages: protocol theory; submodel architecture (DroneIdentity, AirspaceSlot); ROS~2 / Gazebo 3--5 UAS swarm lab-demonstrator (``Delivery by Drone''); ML state-prognosis + predictive maintenance. Outcome target: open-access reference architecture + IDTA/EUROCAE standardisation contributions. Primary source for aidocs/strategy/101-diva-project-context.md. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: krebsDIVA2026
• Krebs, Florian (2025). Fähigkeiten-Inventur DLR-BT --- Standort Augsburg / ZLP.
  Comprehensive machine-portfolio inventory for DLR BT/ZLP Augsburg. Documents nine named cells/Anlagen (FPZ Fiberplacement-Zelle, TPZ Thermoplast-Zelle, IQZ Inline-Qualitätssicherungszelle, MFZ Multifunktionale Zelle, TEZ Technologie-Erprobungszelle, Heißpresse, Cutterzentrum, Wasserstrahlanlage, CoBots+AGV) and eight cross-cutting technologies (LEA Laser-Excited Acoustics, 3D-mapped Thermography, 3D Geometrie-Vermessung incl.\ Leica AT901, AFP on-the-fly 3D profilometer, autonome Produktion, Automated Defect Recognition, Soll-Ist-Vergleich, integrated digital-twin layer). Slide~22 (Panel Produktion process chain) names ``Datenerfassung, -speicherung und -auswertung, DigTwin'' as the underlying data layer --- the integration target Shepard serves. Slide~23 (Krebs F., 2025-02-21) positions Shepard explicitly as the cross-domain substrate spanning Structural mechanics / Materials / Design,Process / Production / Ground+Flight Test in the ``Digital integration by process chain spanning data management system'' band; this is the canonical thesis-substrate figure. Primary source for aidocs/strategy/100-shepard-bt-zlp-rollout-plan.md; one-line citation in aidocs/strategy/101-diva-project-context.md. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: krebsInventur2025
• Krebs, Florian and Glück, Roland and Lettowsky, Felix and Kaufmann, Patrick and others (2026). Semantically-Driven Data Analytics --- the ForInfPro use case.
  Position deck describing the ``Context Gap'' framing, ontology--Shepard interplay, and the ForInfPro composite-infusion process-control use case at DLR ZLP. Contains the canonical Shepard contextualisation diagram (Infusion / Pump / Sensor / Image container $\to$ Experiment data object) and the dry-spot reasoning chain (Phenomenon $\to$ Measurable $\to$ Physical quantity $\to$ Constraint). Used as primary source for aidocs/strategy/91-forinfpro-semantically-driven-analytics.md. Provenance: uploaded to AI working memory 2026-05-23.
  Citation key: krebsForInfPro2026
• Krebs, Florian (2026). Generative KI: Technologie, Ethik und Zukunft --- Wie LLMs (unser) Arbeiten und Denken (verändern).
  Author-led tutorial on LLM mechanics (tokenisation, embeddings, transformer attention, training phases), data limits (Chinchilla optimum, model collapse), the alignment problem, ethical risks for scientists (hallucination, bias, reproducibility), and the NIST AI RMF. Includes a closing ``AI Disclosure'' slide naming the SAIA/GWDG scientific cloud, Marp CLI, Docling-MCP, paper-search MCP, and web-search tools as the workflow used to produce the deck --- a primary-source statement of the author's transparent-AI-use posture. Used as primary source for aidocs/strategy/89-genai-methodology-and-reflexivity.md. Provenance: uploaded to AI working memory 2026-05-23.
  Citation key: krebsGenKi2026
• Krebs, Florian (2026). Humanoids 2026: Trends, Achievements, Gaps --- Implementation in Manufacturing.
  Market and capability survey of humanoid-robot deployments in manufacturing (Xiaomi, Tesla, BMW, Toyota, Amazon/GXO), with citations to Gitnux 2026, SVRC 2026, Lexology 2026, Roland Berger 2026, McKinsey 2025, Axis Intelligence 2025, Bloomberg 2026. Surveys ISO~10218, ISO/TS~15066, ANSI/A3~R15.06-2025, ISO~25785-1, ISO~13482 as the relevant safety-standards landscape. Tangential to Shepard's core thesis arc; logged in the thesis-memory artefact inventory but not the substrate of a standalone strategy doc. Provenance: uploaded to AI working memory 2026-05-23.
  Citation key: krebsHumanoids2026
• Bayerisches Fachpanel --- Aufbau und Nutzung von Datenräumen (2026). Ergebnis Fachpanel: Aufbau und Nutzung von Datenräumen --- 20.03.2026.
  Outcomes of a Bavarian expert panel on data spaces: SWOT analysis (strengths/weaknesses/opportunities/risks) plus six action fields (Gelebte Datenkultur; Expertise aktivieren; Zugang zur Infrastruktur; Souveränität & Resilienz; Regulatorik & Standards; Bayern agiert global), each with a vision statement. Regional/policy context for Shepard's positioning in the Bavarian dataspace ecosystem (cited as a regional-context aside in aidocs/strategy/87 §5). Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: bayernFachpanel2026
• Krebs, Florian (2025). Intro NFDI4Ing (1+2).
  NFDI4Ing introduction and Phase 2 (2025--2030) DLR-partner positioning. Names Measure F-1 (data integration guidelines; partners WZL, IPT, IoP, DLR) and Measure F-2 (Data Mesh to RDM adaption concept; partners WZL, IPT, TUB, DLR). Internal DLR partners listed: IW, TS, BT, RY, (VE) at 4 PM/a. Uploaded to AI working memory 2026-05-23. Primary source for aidocs/strategy/88-nfdi4ing-alignment.md.
  Citation key: nfdi4ingIntroKrebs2025
• Krebs, Florian (2025). HMC 2 Workpackages @ DLR ZLP.
  Helmholtz Metadata Collaboration Phase 2 work-package proposal naming three Shepard-bearing WPs: (1) Solutions for data export and import (RO-Crate, Shepard import/export functionality, PID integration), (2) Shepard semantic features extensions development (ontology-annotation usability), (3) Establish interoperability between DM solutions in DLR (Shepard, twinStash, Defacto, Inst.DLR; Databus + MOSS foundation). Uploaded to AI working memory 2026-05-23. Primary source for aidocs/strategy/90-hmc-phase-2-positioning.md.
  Citation key: hmcPhase2WpKrebs2025
• NFDI4Energy. Databus and MOSS Integration Guide.
  NFDI4Energy-published integration manual for DBpedia Databus + MOSS (Metadata Online Submission System). Databus is ``a digital factory platform for managing files online with stable IDs, high-quality metadata, powerful API and tools for building on data'' (\urlhttps://github.com/dbpedia/databus). Uploaded to AI working memory 2026-05-23. Cited in aidocs/strategy/90-hmc-phase-2-positioning.md \S 5 as the federation-substrate manual Shepard's WP-3 prototype would follow.
  Citation key: nfdi4energyDatabusMossGuide
• Krebs, Florian (2025). KUKA CNC Robotics.
  Author-led internal evaluation of the KUKA.CNC integration option (KUKA robots fronted by Siemens Sinumerik 840D control) versus the conventional KRL+RoboDK path the DLR ZLP centre operates today. Documents the X63 EMI interface (PROFINET IRT / ProfiSAFE), KSS 8.6+ requirement, lack of IIQKA.OS compatibility, the ``Sinumerik Run MyRobot / Machining with KUKA robots'' incompatibility with 840D Sl CNC-SW 4.9.x and Sinumerik ONE, and the medium-term compatibility risk posed by Siemens's ``Run My Robot / direct control'' (MTR) direction. Strategic conclusion: \emphCNC-as-compatibility-factor (Airbus, ASA, PAG) weighed against \char`\~\,€100k equipment cost, mandatory OLP, training, and the paradigm tension with autonomous-production research. Primary source for aidocs/strategy/87-dlr-zlp-positioning.md \S 3.1. Provenance: uploaded to AI working memory 2026-05-23.
  Citation key: krebsKukaCnc2025
• Krebs, Florian (2025). Integrating Metrology, Spatial Analyzer and Industry 4.0.
  Author-led survey of the DLR ZLP metrology platform (Spatial Analyzer Ultimate; 2$\times$~Leica AT~901~LR; T-Probe; T-MAC; T-SCAN~5) and its day-to-day application set (GD&T of finished parts/assemblies, process-specific measurements, surface-geometry reconstruction). The deck's load-bearing advanced application is in-situ elastic calibration of robots (IREC): a 18-DOF lumped kinetostatic model running on a Linux-Xenomai real-time PC corrects the KUKA KR~120 HA TCP via the RSI interface every 4~ms, based on ATI 6-axis F/T sensor readings. Published results on the MFFD double-curvature mould with 280~N lateral / 1100~N normal load cases: 89.30% reduction in mean TCP deviation, 70.19% reduction in maximum deviation, trajectory within the $\pm 0.3$~mm aerospace tolerance. Primary source for aidocs/strategy/87-dlr-zlp-positioning.md \S 3.2 and the design seed in aidocs/integrations/96-metrology-spatial-analyzer.md. Provenance: uploaded to AI working memory 2026-05-23.
  Citation key: krebsMetrologyI42025
• Welzmüller, Finn and Dannemann, Frank and Scharringhausen, Marco and Dannemann, Tanja (2025). Research Data Management for Space Missions: Practical Experiences and Lessons Learned.
  Single-slide A0-format poster presenting the PLUTO mission's RDM work and its use of Shepard as the technology stack. Authors are all from DLR Institute of Space Systems, Bremen. The poster cites Friedline et al.~2025 (HMC Conference shepard paper), Haase et al.~2021 (Zenodo shepard release), Aksteiner et al.~2024 (PLUTO mission paper), and Holz 2024 (Bremen master thesis on deployable flexible solar arrays for space). Documents Shepard's role: storage (data containers as partitions of dedicated DBs), organization (collections / data objects / references; Parent/Child for hierarchy; Predecessor/Successor for sequence), access (REST + frontend + Python library + IAM with API keys), and documentation (key-value metadata + ontology classes + built-in Lab Journal). The file was uploaded by Florian Krebs annotated with his comments --- the poster itself is not FK-authored. Bib key chosen to credit the actual authors (Welzmüller \emphet al.). Complements the existing welzmueller2024Pluto entry for the PLUTO mission paper. Uploaded to AI working memory 2026-05-23. Cited as primary source in the thesis-idea artefact inventory (project_thesis_idea.md).
  Citation key: welzmuellerForschungsdatenPoster2025
• DLR Zentrum für Leichtbauproduktionstechnologie (ZLP) Augsburg (2025). DLR ZLP Augsburg --- Data Driven Intelligence.
  Twenty-nine-slide ZLP Augsburg positioning deck framing the centre's mission as ``data-driven'' production technology. Slides 1--7 introduce the centre (mission, four research-group Venn diagram, valley-of-death framing, DIKW-pyramid epistemology, Industry 4.0 vs 5.0 comparison, manufacturing-vision 2030+). Slides 8--14 articulate the digital-strategy stack (Exchange / Standardisation / Analytics / Data management) and name Shepard explicitly as the ``integrated data management system''. Slide 9 is the canonical Shepard architecture diagram (Shopfloor + OPC UA / Edge Collector $\to$ REST API $\to$ graph-organisational-DB with Time series / GIT / Files / Structured-documents stores); slide 10 the (raw) data-capture interface view (PLC Cell, KRC4, TPS Controller, PLC MTLH with PN/EC/UA/FSD/ADS/REST protocol overlay); slide 11 the data-contextualization architecture (Context Provider + Data Aggregator + leading-component observation); slide 12 the MFFD AFP thermoplastic fibre-placement demonstrator showing the live ``shepard_cube2-IR Image'' figure; slide 14 the GTlab/ADAPT gas-turbine-laboratory use case (a second use-case beyond MFFD). Slides 15--17 + 22 cover the Asset Administration Shell (AAS / Plattform Industrie 4.0) substrate. Slides 18--24 lay out the Manufacturing-X / Catena-X / Aerospace-X federation landscape, with slide 23 ``First planned application E2E-Quality assurance'' showing the architectural intent EDC $\leftrightarrow$ AAS Repository $\leftrightarrow$ SHEPARD for cross-supply-chain quality control --- primary-source evidence Shepard is positioned as the substrate underneath the planned aerospace dataspace. Slides 25--29 outlook: REBAR (EASA Learning Assurance), ForInfPro ML control-loop, HoKiTEP mobile-robotics platform, DrapeBot HMI. Used as primary source for aidocs/strategy/87-dlr-zlp-positioning.md \S 8 and aidocs/strategy/94-federation-and-dataspaces.md. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: zlpDataDrivenIntelligence2025
• Krebs, Florian (2025). Aerospace-X --- for JAXA.
  Author-led briefing on the German Manufacturing-X / Aerospace-X programme for the Japanese Aerospace Exploration Agency (JAXA). Core claim: adapt the emerging Catena-X / Tractus-X automotive dataspace ecosystem to aeronautics needs (slide 3 cites Airbus's ``70% reuse, 30% adaptation'' positioning). Slide 5 names the Catena-X stack (AAS as digital-twin technological base; Eclipse Tractus-X open-source code base; estimated $\sim$26 million AAS in the ecosystem demonstrator). Slide 6 lists the four Manufacturing-X priority use cases (Demand-and-capacity management, CO\textsubscript2 emissions / PCF, Circularity / R-strategies, End-to-End quality management). Slide 7 sketches the cross-tier Aerospace-X data space for safety-critical aero-engine components (production monitoring + early warning, root-cause analysis from MRO findings, containment + countermeasure across Tier-n $\to$ Engine OEM $\to$ Aircraft OEM $\to$ Airline). Slide 9 lays out the Manufacturing-X flagship-project function set (Design, Development, Do-how, Control, Communication, Support, Coordination). Slide 10 closes with the synchronisation-project framing (``a standardized global data exchange based on European values''). Used as primary source for aidocs/strategy/94-federation-and-dataspaces.md. Provenance: uploaded to AI working memory 2026-05-23 by F.~Krebs.
  Citation key: krebsAerospaceXJaxa2025
• Krebs, Florian (2026). Federated Shepard & Pub-Service architecture sketches.
  Two architecture sketches by F.~Krebs articulating the cross-instance integration paths Shepard must support. Sketch A (``Federated Shepard architecture''): a process-chain figure showing Simulation/Auslegung (as-designed) $\to$ Bauteilherstellung $\to$ NDT Prüfung (Laser US) $\to$ Re-Simulation (as-built) $\to$ Maintenance, with a side-branch for Destructive Testing. Each step labelled with DLR organisational roles (SY/VPH, SY/BT, BT/MO?, VPH, MO). A bottom horizontal bar labelled ``Common Data Infrastructure (shepard)'' connects bidirectionally to every process step. Footer note records the two integration paths explicitly: \emph``2 Wege: Export Import zur Datenweitergabe (mittels RO-Crate) / Federated shepard service (viel Aufwand, höheres Risiko)'' --- the lower-risk RO-Crate export/import path, vs.\ the higher-effort federated Shepard service path. Sketch B (``Pub-Service publication architecture''): a whiteboard photograph showing Shepard with ``Sammlung V1'' $\to$ ``meta.json'' (with a DL-Lin\ldots\ annotation) $\to$ ``Pub Service'' branching to Databus / S3 / Invenio; md5-checksum annotation on the collection node; ``FDM Institution'' header on the right margin. Concrete shape of the publication pipeline contemplated under HMC Phase 2 WP-1 (PID + export functionality). Both sketches are publication-quality candidates for the thesis architecture chapter; the physical originals (file uploads of the sketch and the whiteboard photograph) need to be re-uploaded for PDF embedding. Cited in aidocs/strategy/94-federation-and-dataspaces.md \S 4 (Sketch A) and aidocs/strategy/90-hmc-phase-2-positioning.md \S 3 (Sketch B). Provenance: received in AI session chat 2026-05-23.
  Citation key: krebsFederationSketches2026
• Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V. (2024). Handlungsempfehlungen zum Einsatz von Künstlicher Intelligenz.
  Helmholtz-binding AI recommendations across 18 member centres. Five risk categories (4.1 data-protection / leak; 4.2 copyright/IP; 4.3 misinformation/scientific integrity; 4.4 bias from training data; 4.5 violation of AI-specific regulation incl.\ EU AI Act). Good-practice baselines: review AI-generated content; document AI use transparently; responsible non-use for sensitive data; prefer European-hosted services. Authorship reserved for natural persons. Primary source for aidocs/strategy/93-management-context-and-compliance.md \S 2 (AI-governance framework) and \S 6.A audit lens. Uploaded to AI working memory 2026-05-23.
  Citation key: helmholtzKiEmpfehlungen2024
• Deutsches Zentrum für Luft- und Raumfahrt (DLR). Unser Leitbild zu Künstlicher Intelligenz.
  DLR's official AI mission statement. Headline claim: ``Wir sind Vorreiter in der Nutzung von und Forschung an KI-basierten Lösungen. Wir setzen KI selbstverständlich und verantwortungsbewusst ein.'' Ten values: Aufgeschlossenheit, Bewusstsein, Menschenorientierung, Kompetenz, Ethik, Rechtssicherheit/Compliance, Effizienz, Souveränität, Transparenz, Alleinstellung. Cited in aidocs/strategy/93-management-context-and-compliance.md \S 2 + \S 6.B and aidocs/strategy/89-genai-methodology-and-reflexivity.md (three-way alignment).
  Citation key: dlrKiLeitbild
• Köster, F. and Raulf, A. (2024). Leitfaden zur Einführung und Nutzung KI-basierter Lösungen im DLR.
  Operational AI-use guidance for DLR, translating the dlrKiLeitbild into a checklist. Builds on EU AI HLEG ethics guidelines, federal AI strategy, OECD AI principles. Seven value-pairs: Menschenzentrierung/Gemeinwohl, Fairness/Diskriminierungsfreiheit, Transparenz/Erklärbarkeit, Privatsphäre/Schutz, Sicherheit/Robustheit, ökologische Nachhaltigkeit, Beherrschbarkeit/Verantwortung. \S 2.3 names EU AI Act, DSGVO, Digital Services Act, Data Act, Data Governance Act as applicable. \S 3 contains the operational checklist. Primary source for aidocs/strategy/93-management-context-and-compliance.md \S 2 + \S 6.B audit lens.
  Citation key: dlrKiLeitfaden2024
• Schlauch, T. and Druskat, S. and Steude, B. (2025). Rahmenrichtlinie Software-Engineering. Deutsches Zentrum für Luft- und Raumfahrt (DLR). DLR-Managementsystem Verfahrensanweisung -- fachspezifisch Qualität
  DLR-wide software-engineering framework directive. Issue 6 (2025-07-01) introduces ``Forschungssoftware'' as an explicit category with Anwendungsklasse 1 as the floor for research software. Defines four Anwendungsklassen (0--3) plus accountability chain (DLR-Vorstand $\to$ CIO $\to$ Software-Engineering-Beauftragter SEB $\to$ Leitung der Organisationseinheit $\to$ Software-Engineering-Ansprechperson SEA $\to$ Verantwortliche Person für die Software). \S 4.2.4 mandates documentation (Name, Zweck, Verantwortliche Person, Anwendungsklasse, Reifegrad) for AK 2/3 and AK 1 with long-term Weiterentwicklungsperspektive. Mitgeltende Unterlagen include QMH-DLR-VA014 Open Science and QMH-DLR-VA015 Forschungsdatenmanagement. Primary source for aidocs/strategy/93-management-context-and-compliance.md \S 3 + \S 6.C audit lens.
  Citation key: dlrRahmenrichtlinieSE2025
• IT-LTG, DLR (2019). Cloud-Strategie des DLR: Rahmenbedingungen für die strategische Nutzung von IT-basierten Cloud-Diensten. Deutsches Zentrum für Luft- und Raumfahrt (DLR).
  DLR-wide cloud-services authorisation framework. Shifts from ``Cloud-Verbot'' to Chancen-Risiken-Abwägung posture. \S 5.3 defines four risk tiers (Niedrig / Mittel / Hoch / Sehr hoch): ``Sehr hoch'' = external cloud forbidden; ``Hoch'' = Einzelfallprüfung + IT-Sicherheitsbeauftragter approval required; clouds outside Germany ``often unsuitable'' for Hoch tier. CIO + IT-Sicherheitsbeauftragter hold Veto-Recht. \S 8 mandates Exit-Strategie + container-Lösungen. \S 9 central cloud-services register. \S 10 export-control rules. Primary source for aidocs/strategy/93-management-context-and-compliance.md \S 4 + \S 6.D audit lens.
  Citation key: dlrCloudStrategie2019
• IT-LTG, DLR (2019). Anhang zur Cloud-Strategie des DLR: Nutzungsszenarien. Deutsches Zentrum für Luft- und Raumfahrt (DLR).
  Annex to the DLR Cloud-Strategie with concrete usage scenarios. \S 4.1 Prozessierung öffentlicher Daten; \S 4.2 Externe Speicher/Datenaustausch (Dropbox / OneDrive / AWS); \S 4.3 Dedizierte externe Wissenschafts-Clouds (DFN, EOSC) -- the category Shepard's nuclide demo falls in; \S 4.4 Externe Kommunikation/Kollaboration; \S 4.5 Externe Rechenkapazität -- mandates Exitstrategie; \S 4.7 Externe Software-Entwicklungsplattformen (GitHub); \S 4.8 Externe Web-Portale; \S 5 interne Cloud-Dienste. Cited in aidocs/strategy/93-management-context-and-compliance.md \S 4.
  Citation key: dlrCloudStrategieAnhang2019
• Deutsches Zentrum für Luft- und Raumfahrt (DLR) (2026). Datenstrategie des DLR.
  DLR-wide data strategy, v60 working draft. Vision: ``DLR als datengetriebenes Innovationszentrum -- Pionier für zukunftsweisende Forschung und kontinuierlichen Fortschritt.'' Four objectives: (Ziel 1) integrierte Dateninfrastruktur + -plattformen (Datenökosystem); (Ziel 2) interdisziplinäre Forschung + Innovation fördern; (Ziel 3) Entscheidungsprozesse durch unabhängige prädiktive Analysen unterstützen; (Ziel 4) Datenkultur etablieren + Datenkompetenz erhöhen. Maßnahmen include: föderierte Speicherarchitektur (Data Hub); FAIR-Prinzipien verbindlich; kontrollierte Vokabulare/Ontologien; Data-Steward-Netzwerk; Politikberatung. Cross-references DSGVO, SatDSiG, Exportkontrolle, NDA. \S 5 Empfehlungen-Maßnahmen list and overall accountability/governance pattern still under editorial debate. Primary source for aidocs/strategy/93-management-context-and-compliance.md \S 1 (fifth institutional layer) and \S 5 (funding/evaluation cadence). Cite future versions as v$X$ in the same key once finalised.
  Citation key: dlrDatenstrategieV60_2026
• Krebs, Florian (2010). Steuerungskonzept der MFZ-Anlage.
  Earliest primary source in the predecessor-systems lineage at DLR ZLP Augsburg. Specifies the automation control concept for the Multifunktionale Fertigungszelle (Multifunctional Production Cell). Read for design intent, it articulates 14 commitments that map directly onto Shepard surfaces a decade-plus later: Konsequente Vermeidung von Einzel-/Insel-L\"osungen (avoidance of point/island solutions), Modulare Architektur f\"ur schrittweise Weiterentwicklung (modular architecture for incremental evolution), Generische Schnittstelle f\"ur Messdatenerfassung (generic measurement-data interface -- the PayloadKind / PayloadStorage SPI ancestor), Automatische Dokumentation von Prozessdaten (automatic process-data documentation -- the W3C-PROV Activity stream ancestor), Detektion von Abweichungen zum Kalibrier-Prozess (deviation detection from calibration baseline -- anomaly-detection ancestor), Zentrale Datenhaltung auf Datenbank (central DB-backed storage), Replay-F\"ahigkeit in Simulation zur Analyse und Optimierung (replay capability for analysis and optimisation), Anlagenkonfigurationsmanagement (plant configuration management -- the admin-configurable knobs ancestor, cf. A3b / N1c2 / UH1a), Offene Schnittstellen: AutomationML, XML (open interfaces), Drahtloses virtuelles SmartPAD + Mobile Bedienpanels (wireless shop-floor terminals), Integration des NDT Frameworks ($\to$ PI-NDT) (NDT as first-class). Primary source for aidocs/strategy/86-shepard-predecessor-systems.md \S 1.5 (pre-PRAESTO design thinking already articulated in 2010).
  Citation key: krebsMfzSteuerung2010
• Krebs, Florian (2010). Themen Robotik für Faserverbundfertigung.
  Companion slot-deck to krebsMfzSteuerung2010. Sparse substantive content (the deck functions as a section-header placeholder for the research-themes-around-the-MFZ context). Cited for completeness in aidocs/strategy/86-shepard-predecessor-systems.md \S 1.5 as evidence that the 2010 design-thinking layer extended beyond pure cell-control concerns into the robotics-for-composites research agenda.
  Citation key: krebsRobotikThemen2010
• Krebs, Florian (2012). ``Wanted'' office-door poster.
  Single-page satirical ``Wanted'' notice that hung on Krebs's office door at DLR ZLP Augsburg. Text in full: ``WANTED. EVIL GENIUS seeks minions to sacrifice their lives in world domination attempt. Must be prepared to work 24-7 for psychopath for close to no pay. Messy death inevitable but costumes and laser death rays provided. NO Weirdos.'' Cited at aidocs/strategy/89-genai-methodology-and-reflexivity.md \S 8.5 as a small biographical primary source: the satire works only if the author is self-aware about the PI/collaborator asymmetry it parodies, which calibrates the register of the author's working voice fourteen years before the same author articulated, on a public deck, a thoughtful position on the human-machine asymmetry in AI-assisted science. Footnote-class source; not load-bearing.
  Citation key: wantedFlo2012
• Krebs, Florian (2025). ODIX -- Milestone Report: ``first prototype finished and evaluated''.
  Cross-cutting milestone report covering all four ODIX work-packages (HAPs): HAP 1 -- information extraction from semi-structured technical documents (DW, LLM-based PDF $\to$ JSON with HITL correction, GUI usable, 2--3\,min latency per PDF, output to move from JSON to RDF in next deliverable); HAP 2 -- Shepard storage management with metadata schemas + timeseries support (BT, Krebs's group; semantic annotation + interval-metric calculation shipped; component-to-component automation pending); HAP 3 -- exploration UI for the extracted KG (DW; categories: Machines / Procedures / Planned Procedures / Samples; planned-vs-set-vs-measured comparison; CI on GitLab Pages with GitLab-issue feedback loop); HAP 4 -- textual evaluation documentation (SY; per-component evaluation summaries from WF/FM and MO). Two thesis-load-bearing findings: (a) HAP 2's documented gap (``connection of components not yet automated'') is exactly the gap that platform-level SHACL / TPL design closes; (b) HAP 1's pivot from direct PDF $\to$ RDF to JSON-as-intermediate, with planned move to RDF output, confirms the platform's posture of first-class shapes as the substrate-side answer. Primary source for aidocs/semantics/95-shacl-templates-and-individuals.md \S 14ee (ODIX Milestone 2 -- what the four work-packages actually shipped).
  Citation key: odixMilestone2_2025
• Helmholtz-Gemeinschaft Deutscher Forschungszentren (2025). Forschungsbereichsübergreifende forschungspolitische Ziele für die PoF V (2028--2034). Helmholtz-Gemeinschaft Deutscher Forschungszentren.
  Forschungspolitische Ziele (research-policy targets) of the Helmholtz Association for the fifth Programme-Oriented Funding period PoF V (2028--2034). Binding cross-association targets that apply to ZLP Augsburg through its HGF-affiliated centre DLR. Three Datenmanagement obligations directly relevant to Shepard: (a) ``the definition of a joint and cross-cutting architecture concept for the data management infrastructure (data acquisition, data analysis / aggregation, data provision per FAIR principles, and IP management)''; (b) NFDI alignment required, not optional -- ``the fundamental inclusion of work and recommendations already produced by NFDI''; (c) Datenresilienz -- ``the safeguarding of data resilience to guarantee the availability, integrity and access to data even in the case of unforeseen events, disruptions or failures''. Sibling cross-association targets: Pakt für Forschung und Innovation continuity through PoF V; Nachhaltigkeit (joint sustainability strategy + climate-neutral operation of research infrastructures). Primary source for aidocs/strategy/87-dlr-zlp-positioning.md \S 7.5 (the federal-funding-cycle frame -- HGF research-policy targets for PoF V).
  Citation key: hgfFoPoZi5_2025
• Krebs, Florian (2020). CUBE iDMS --- Final Presentation.
  Load-bearing primary source for aidocs/strategy/86 \S 4. Authoritative for: system name (``CUBE iDMS'', slide 14), architecture (slide 14: graph DB + per-data-type substrate, ``no one-size-fits all''), technology stack (slide 15: Neo4j + InfluxDB + MongoDB + Tomcat/Jersey + Flask + docker containers), data model (slide 17: Project / Experiment / Step / Artifact), vision (slide 10: ``one data base -- many digital twins''), data-pipeline model (slide 12: three input modes), Neo4j rationale (slide 16: schema-light iterative model growth), security model (slide 18: Keycloak + API-key JWTs at bt-au-keycloak.intra.dlr.de). Slide footers carry ``Datenmanagement > Florian Krebs 12.12.2017'' on multiple slides, evidence the 2020 deck is a refinement of the December 2017 vision (see \S 3.6.4 of \S 86).
  Citation key: krebsIdms2020
• Krebs, Florian (2017). Industrie 4.0 @ ZLP --- Prozessdatenerfassungssystem (ProDES) v1.
  Earliest ProDES iteration. Sole-author Krebs. Cited at aidocs/strategy/86 \S 3.6.1. Carries the ``many islands'' diagnosis of the state of the art at ZLP (PRAESTO + KIBID as isolated platforms), the layered basis architecture (Datenerfassung $\to$ Datenorganisation & Datenablage $\to$ Datenzugriff $\to$ Analyse $\to$ Management & Administration), and the NoSQL technology survey (Document / Column / Key-Value / Graph categories named with CouchDB, MongoDB, hBase, Cassandra, Neo4j, OrientDB).
  Citation key: krebsProDES2017a
• Nuschele, Stefan and Schuster, Alfons and Krebs, Florian (2017). Industrie 4.0 @ ZLP --- Prozessdatenerfassungssystem (ProDES) v2.
  Two-month expansion of krebsProDES2017a adding Nuschele (PRAESTO author, \S 2 of \S 86) and Schuster as co-authors. First primary-source evidence that the PRAESTO-era authoring community and the iDMS-era architect co-authored a transitional design. Slide 27 (``Ein System -- mehrere Sichten'') names the three-views model (Prozessketten / Prozesse / Ressourcen) that surfaces verbatim in the December 2017 Vision deck and the 2020 iDMS final presentation. Slide 36 schedules ``MS1 12/17: Evaluation KiBiD abgeschlossen'' -- dating the KIBID-evaluation deadline six months in advance of the December Vision synthesis. Slide 38 (``Mögliche Komponenten eines IDMS am ZLP'') coins the ``iDMS'' name six months before the Vision deck. Cited at aidocs/strategy/86 \S 3.6.1.
  Citation key: krebsProDES2017b
• Krebs, Florian (2017). Status Prozessleitsystem.
  Cited at aidocs/strategy/86 \S 3.6.2. Names the orchestration goal as ``modularisation of automation solutions (Vision: Plug & Automate)'' with cross-process information exchange as concrete first step. Slide 10 frames the first iDMS-era orchestrator as ``einfache Schrittkettensteuerung einer SPS'' -- the PLC-step-sequence-into-software-orchestrator lineage that Shepard's importer-plugin family inherits. Slide 11 (``Status Big-Data-Adapter'') reports ``Aufzeichnung von Roboterdaten mittels KiBiD erfolgreich'' and names Christian R\"ahtz, Mark Willmeroth on the integration workstream -- the earliest primary-source appearance of Willmeroth (later 2021 Shepard co-author per haaseShepard2021) in the predecessor record.
  Citation key: krebsStatusProzessleitsystem2017
• Krebs, Florian (2017). NSR Prozessleitsystem Architektur (NSR Orchestrator Konzept).
  Load-bearing primary source for aidocs/strategy/86 \S 3.6.2. Slide 3 (``IPRO: Aufbau'') names the IPRO architecture as ``4 Interface-Knoten, 3 Virtuelle Knoten, Kommunikation über OPC UA, Datenerfassung in KiBiD über HTTP Schnittstelle'' -- the first primary-source corroboration in the chapter that KIBID was the IPRO data-acquisition layer, named inside an architecture diagram by Krebs himself (until this deck, the KIBID$\to$IPRO connection was only hostname-level evidence in source code). Slide 5 (``Architektur: OPC UA Prozesleitsystem'') depicts the OPC-UA-router-fronting-heterogeneous-device-fleet pattern (KUKA + Siemens IOT 2040 + Festo CEC + Siemens S7-300) with ``Python implementation for all clients'' -- the architectural ancestor of shepard-timeseries-collector.
  Citation key: krebsNSROrchestrator2017
• Krebs, Florian (2017). Industry 4.0 defined (curated reading of Design News article).
  Cited at aidocs/strategy/86 \S 3.6.3. Not original design work but a vocabulary-adoption pass -- documents the moment Digital Twin, Cyber-Physical Systems, M2M and IoT vocabulary entered the iDMS design discourse from external trade press. Slide 11's ``Digital Twin: Virtual Products in Virtual Production'' definition becomes the ``one data base -- many digital twins'' framing in krebsIdms2020 slide 10. Methodologically narrower evidentiary role than the surrounding decks (reading-pass, not design moment).
  Citation key: krebsI4Definition2017
• Krebs, Florian (2017). NSR Architektur (working deck).
  Cited at aidocs/strategy/86 \S 3.6.4. Substantive prose content thin (deck retained template Lorem-Ipsum boilerplate around figure placeholders), but the \emphdate and filename establish that an NSR-architecture working deck was iterated on 2017-12-04 -- the day before the Vision-Datenmanagement deck. Bracket the December architecture-vision-deliverable week.
  Citation key: krebsNSRArchitektur2017
• Krebs, Florian and Nuschele, Stefan (2017). Vision: Datenmanagement.
  Load-bearing primary source for aidocs/strategy/86 \S 3.6.4. The architectural blueprint that the 2020 iDMS final presentation (krebsIdms2020) refines, not synthesises. Slide 9 (``Mögliche Architektur eines integrierten Datenmanagementsystems'') is the direct architectural ancestor of krebsIdms2020 slide 14 (graph-DB for relations + per-type substrates + ``kein one-size-fits all'' -- verbatim). Slide 10 (``Exkurs: Graphorientierte Datenbanken'') carries the schema-light justification (``Das Datenmodell muss nicht vollständig sein, und kann ständig um neue Beziehungen erweitert werden'') that becomes krebsIdms2020 slide 16 and Shepard's SHACL-driven model evolution. Slide 19 (``Weitergehende Fragestellungen'') articulates the open-source commitment (``Open Source Ansatz'', ``Umsetzung eines funktionalen Prototypen anschließend Transfer an und Upscaling durch Partner'') and Helmholtz Open Science alignment four years before Shepard appeared on GitLab. Footer date 12.12.2017 is exactly the KIBID-evaluation deadline that krebsProDES2017b slide 36 set six months in advance.
  Citation key: krebsVisionDatamanagement2017
• Krebs, Florian (2018). SOA MES --- Service-oriented architecture for a Manufacturing Execution System.
  Cited at aidocs/strategy/86 \S 3.6.5 as an adjacent conceptual artefact that cannot be ordered by its own metadata. ``Everything is a service'' framing with three participant classes (physical / virtual / worker-ERP) and a four-step task-execution model. Slide 6's Basis-Architektur places \emphSOA MES alongside \emphProzess- und QS-Datensammlung (CUBE, KiBiD?) -- the CUBE acronym appears at slide-label level with \emphKiBiD? carrying a question mark. Workshop-organiser slide names Tobias Haase (first-named 2021 Shepard Zenodo author per haaseShepard2021) and Krebs as co-organisers -- the earliest primary source in \S 86 linking Haase to Krebs's working community.
  Citation key: krebsSoaMes
• Schönheits, Manfred and Krebs, Florian (2017). Robotics: Human-Robot-Interaction (Workshop DLR@UBC, 06.12.17--08.12.17).
  12-slide deck on the same robot-cell testbed that becomes the VAM-HRI 2018 paper (dlrSchoenheitsVamhri2018) three months later. Slide~1 dates the talk explicitly: ``Workshop DLR@UBC 06.12.17 -- 08.12.17''. Slide~3 (Objectives): ``Applying HRI concepts for (big) industrial robots'' and ``Create basis / starting point for future work''. Slides~4--7 introduce the four core concepts: UWB localisation (Pozyx), Augmented Reality (Hololens), OPC-UA, V-REP co-simulation. Slide~10 emphasises a peer-to-peer distributed-system stance: ``No master or central node $\ne$ ROS running master control logic''. Slide~11 lists the same four voice commands (``Stop'', ``Look at me'', ``Come here'', ``Go home'') and notes the safety-critical convention ``Stop command is always required to proceed''. Slide~12 (Future work): ``multi-laser sensor setup'', ``expansion to large scale scenarios'', ``automatic collision-free path planning''. Chronological observation: this presentation was given on the same day as 20171205_Vision_Datenmanagement.pptx (prior batch) -- DLR@UBC 2017-12-05 produced both the institute's HRI demo \emphand the institute's data-management vision presentation. Accessed 2026-05-23.
  Citation key: dlrSchoenheitsKrebsHriWorkshop2017
• Krebs, Florian (2025). Intro DaMaST --- Data Management for Space Transport.
  Primary source for aidocs/strategy/95-damast-space-transport-data-management.md. Names DLR space-transport-cluster Vorhaben 2025 (use-case selection: STORT + LUMEN; user-requirement interviews across AS/BT/RA/RB/RY institutes; 2025 proposal failure attributed to ``komplexe Interaktionen'' + ``offene Fragen bzgl. Governance''; extension to Vorhaben 2026) and Vorhaben 2026 (goal: technology demonstrator DataHub = shepard + databus + MOSS; proposal target $\sim$6/26, full proposal $\sim$10/26). Slide-5 alignment statement: ``1:1 was in NFDI (und auch HMC) aufgebaut werden soll, aber anderer Use-Case''. Slide-6 ZLP rationale: establishing Shepard in DLR Raumtransport area is a stated strategic objective.
  Citation key: krebsDamastIntro2025
• DLR ZLP Augsburg (2022). Erweiterungsszenario ZLP Augsburg --- KI-Produktionsnetzwerk / HoKiTeP.
  Primary source for aidocs/strategy/96-hokitep-zlp-extension-and-joint-usecase.md \S 1. Names HoKiTeP-the-cell (\emphHoch-Kinematisierte Technologie-Entwicklungs-Plattform) as the new building extension's flagship cell type: industrial robots on linear axes + autonomous mobile units cooperating \emphim Team for Lot-Size-1 / Changeover manufacturing, chainable into existing robot cells for multi-step processes. Funding context: regional Bavarian \emphKI-Produktionsnetzwerk initiative, $\sim$\,\texteuro8\,M from Freistaat Bayern; extension also adds Co-Working-Space + F\"ugelabor + workshops + storage. Slide 4 threads the project into seven strategy layers (DLR-Strategie 2030, Luftfahrt-Leitkonzepte, HGF POF-4, Bayerische Luftfahrtstrategie 2030, BDLI Technologiestrategie 2020, BMWi I4.0 Leitbild 2030, EU Clean Aviation SRIA, EASA AI Roadmap). Slide 5 future use-cases: panel assembly, fuselage-barrel assembly, cabin-element integration, flying-taxi assembly.
  Citation key: hokitepPresentation2022
• Krebs, Florian (2025). ZLP joint use-case (HoKiTeP HAP 2 + 13 sibling projects).
  Primary source for aidocs/strategy/96-hokitep-zlp-extension-and-joint-usecase.md \S 2. Maps a Halle-43 production-hall process chain (CFRP skin manufacture $\rightarrow$ QS $\rightarrow$ Stringer/Spante $\rightarrow$ NDT $\rightarrow$ Assembly\,&\,System-Installation $\rightarrow$ QS $\rightarrow$ final NDT, target part a small \emphvorausger\"ustetes Demo-Schale enabling iteration) to fourteen named project funding lines: ZLP, NFDI4Ing 2, HMC 2, ASPIRO, HOKITEP HAP 2, HEMERA, HERA, INTEGRA, DICADEMORE, ODIX/ForInfPro, GREATER, FRAME H2, HERFUSE, DaMaST. Load-bearing assignment (slide 3): the data-management & analytics layer of the joint use-case is assigned to NFDI4Ing 2, HMC 2, ODIX/ForInfPro, and DaMaST --- precisely the four Shepard-bearing projects this library already integrates. Focus areas: process chaining (à la Protec NSR), digitale Durchg\"angigkeit, geometric precision (``No Shim''). Slides 4--6 (Balkenplan / Zulieferungen Projekte) are present but largely image-based; project-scheduling detail not extracted.
  Citation key: hokitepZlpUseCase2025
• Haase, Tobias (2022). Workshop -- Datenmanagement mit dem iDMS / Architecture Overview: Welcome to shepard.
  Primary source for aidocs/strategy/86 \S 4.6. Direct iDMS$\rightarrow$Shepard bridge artefact in Tobias Haase's hand: the file's original title is an iDMS workshop deck (2021-02) but the first slide already reads ``Architecture Overview: Welcome to shepard'' --- caught mid-stride in the naming transition (last-modified 2022-03-18, four months after the 2021-12 Zenodo Shepard release [@haaseShepard2021]). Five load-bearing content claims: (1)~the acronym \emphstorage for heterogeneous product and research data is fixed by 2021-02; (2)~the 2020 iDMS four-payload architecture (Time series + GIT + Files + Structured Documents over a graph DB with REST API + Web Interface) survives verbatim into the deck's slide 1; (3)~the Project/Experiment/Step entity model of [@krebsIdms2020] is re-rendered as Collection/DataObject/DataObjectReference/Dataset (slide 2); (4)~Containers as ``enclosed boxes'' with per-container permission boundary (slide 3); (5)~explicit API-as-primary-contract claim --- Web-UI is the recommended-against path, Python/Java/C++ client libraries are the recommended interaction shape (slide 4). Slide 5 names ``Currently plans for machine-actionable RDM-policies to automatically validate input. Also versioning capabilities are being looked at in planned projects'' --- both items now operationalised in this fork (SHACL substrate + VersionableEntity).
  Citation key: haaseShepardArchOverview2022
• Krebs, Florian (2023). Shepard SPATIAL --- adding support for ``spatial data''.
  Primary source for aidocs/data/81-spatial-data-binding.md \S 8a. Design ancestor of plugin-spatial (aidocs/16 task #79) and of the current spatial DataBinding design. Five claims that survive in present design: (1)~Spatial as a first-class kind alongside time-series / GIT / files / structured documents (slide-2 architecture diagram, slide-3 four-lane extension shape: Frontend / API Extension / Processing / Data storage); (2)~Payload schema (slide 4): mandatory \emphcoordinates (x,y,z) + \emphmeasurement data (scalar; 1D/nD array; serialised JSON); optional \emphtimestamp, \emphorientation (Euler angles), \emphoperational metadata (coordinate system, track/run number, layer); (3)~Three operation classes (slide 5): POST /store for live capture, POST /import for a-posteriori capture, GET /select for retrieval with k-nearest / bounding-sphere / bounding-box / by-timestamp / by-metadata queries + downsampling + coordinate-transform; (4)~Container as the permission boundary (slide 10) analogous to timeseries containers; (5)~Spatio-temporal as the outlook layer (slide 8 ``Join classic timeseries and spatial data''; slide 9 ``Outlook: Spatio-temporal data''). Slide 4 even sketches the URL shape: col/do T123[scanDir:right]/spatialRef $\rightarrow$ payload.
  Citation key: krebsShepardSpatialConcept2023
• Gülhan, Ali and Hargarten, Dorian Amadeus and Zurkaulen, Maximilian and Klingenberg, Florian and Siebe, Frank and Reimer, Thomas and Di Martino, Giuseppe (2022). Objectives and Achievements of the Hypersonic Flight Experiment STORT.
  DLR eLib 189284. STORT = Schlüsseltechnologien für hochenergetische Rückkehrflüge der Trägerstufen --- DLR-led hypersonic sounding-rocket experiment. Three-stage configuration, achieved Mach 8+. Lead investigator Gülhan (ORCID 0000-0003-4905-5881). Partner institutes: DLR Institute for Aerodynamics and Flow Technology + Mobile Rocket Base. Funding: HGF Space Programme / DLR Space Focus. Cited from aidocs/strategy/95 \S 1 as the primary STORT objectives publication; STORT is the second of two DaMaST Vorhaben use-cases (alongside LUMEN).
  Citation key: guelhan_2022_stort_iac
• Gülhan, Ali and others (2023). Selected results of the hypersonic flight experiment STORT. Acta Astronautica.
  DLR eLib 200758. Peer-reviewed comprehensive-results paper for the STORT hypersonic flight experiment. Cited from aidocs/strategy/95 \S 1 as the STORT load-bearing peer-reviewed publication of record.
  Citation key: guelhan_2023_stort_acta
• Reimer, Thomas and Di Martino, Giuseppe and Dauth, Mario and Baier, Hans-Joachim and Gülhan, Ali and Klingenberg, Florian and Hargarten, Dorian Amadeus (2025). STORT Hypersonic Flight Experiment Forebody Temperature Results.
  DLR eLib 218011. AIAA SciTech presentation of STORT forebody temperature instrumentation results. Cited from aidocs/strategy/95 \S 1 as the recent thermal-management measurements publication.
  Citation key: reimer_2025_stort_aiaa
• Di Martino, Giuseppe and Reimer, Thomas and Baier, Hans-Joachim and Dauth, Mario and Hargarten, Dorian Amadeus and Gülhan, Ali (2026). Thermo-Mechanical Design of the C/C-SiC-Based Thermal Protection Structure for the Forebody of the Hypersonic Sounding Rocket STORT.
  DLR eLib 223910. Thermo-mechanical design publication for STORT's C/C-SiC thermal protection structure. Cited from aidocs/strategy/95 \S 1 as the engineering-detail publication on the CMC thermal-protection system used in STORT.
  Citation key: dimartino_2026_stort_tps
• Traudt, Tobias and Deeken, Jan Christian and Oschwald, Michael and Schlechtriem, Stefan (2019). Liquid Upper Stage Demonstrator Engine (LUMEN): Status of the Project.
  DLR eLib 131429. Foundational LUMEN project-status paper. DLR Institute of Space Propulsion (RA-Lampoldshausen). Cited from aidocs/strategy/95 \S 1 as the LUMEN project-overview publication of record.
  Citation key: traudt_2019_lumen_status
• Riccius, Jörg R. and Zametaev, Evgeny B. and Gulczynski, Mateusz Tadeusz and Dos Santos Hahn, Robson (2022). Numerical turbine blade fatigue analysis with partial admission effects.
  DLR eLib 189488. LUMEN turbine-blade fatigue analysis with partial-admission effects --- engineering-detail publication on the LUMEN test bed turbopump. Cited from aidocs/strategy/95 \S 1 as the LUMEN turbine-fatigue methodology paper.
  Citation key: riccius_2022_lumen_turbine
• Gulczynski, Mateusz Tadeusz and Riccius, Jörg R. and Zametaev, Evgeny B. and Dos Santos Hahn, Robson and Deeken, Jan Christian and Waxenegger-Wilfing, Günther and Oschwald, Michael (2023). Turbine blade fatigue life investigation for reusable liquid rocket engines.
  DLR eLib 194305. LUMEN turbine-blade fatigue-life investigation in the reusable-engine context. Cited from aidocs/strategy/95 \S 1 as the LUMEN reusable-engine-lifetime publication.
  Citation key: gulczynski_2023_lumen_fatigue
• Hardi, Justin S. and Traudt, Tobias and Börner, Michael and Armbruster, Wolfgang and Bee, Lukas and Dos Santos Hahn, Robson and Dresia, Kai and Kurudzija, Erkan and Groll, Robert and Müller, Marcel and Klein, Sebastian and Suslov, Dmitry and Haemisch, Jan and Deeken, Jan Christian and Schlechtriem, Stefan (2025). Recent improvements in the LUMEN test bed for oxygen-methane rocket engine technologies.
  DLR eLib 218583. Most recent LUMEN test-bed publication; 14-author paper documenting recent improvements for oxygen-methane rocket-engine technologies. Cited from aidocs/strategy/95 \S 1 as the current-state LUMEN publication; the depth of the author list signals the scale of the active research team.
  Citation key: hardi_2025_lumen_testbed

How to cite

• In a docs/* page: link to the anchor on this page using the citation key, e.g. [Wilkinson et al. (2016)](/bibliography#ref-wilkinsonFair2016).
• In an aidocs/* note: add a footnote-style reference using the citation key and add an entry to references.bib if it isn’t there yet.
• In a downstream LaTeX document: clone the repo and run \bibliography{docs/_data/references} — the keys are stable.
• Citing Shepard itself: use the CITATION.cff at the repo root. GitHub’s “Cite this repository” UI consumes it automatically.

Snapshot date: 2026-05-22.