Roadmap
What exists, what is proposed, and what each stage must show
A present-tense statement here describes an implemented capability. Everything else is a plan, a question, or a research direction. No dates, staffing, or funding commitments have been made; none are shown.
Where the publication stands
5of 10 essays drafted
5essays as briefs
7of 20 monographs drafted
13monographs as briefs
50lexicon entries in draft
0items reviewed
A brief is an assignment, not a manuscript. A draft is unreviewed. The target inventory of 10 essays, 20 monographs, and 50 concept entries stays visible until it is complete and reviewed. See the essays and monographs indexes for each item's state.
Governing principles
- P001 Build the living prototype before bureaucracy. Everything else must bow to that.
- P002 Unfinished in quantity, complete in soul.
- P003 Stop and reflect before expanding.
Stages
Roadmap stages follow the content creation guidance (section 4, roadmap item structure) and the website requirements PLAT01 to PLAT08. No dates, staffing, or funding commitments were agreed; none are shown.
W0Living publication (this website)Current
- Current state
- Implemented as a private prototype: home, about, five sections, essay and monograph indexes, a 50-entry lexicon, search with type and section filters, this roadmap, and a consent-based updates interest list. Manuscripts are a mix of AI-assisted drafts and writing briefs; nothing is yet source-checked or reviewed.
- Intended capability
- A usable editorial structure that connects essays, monographs, and concepts, with honest content states, so that the intellectual program can be read and reviewed before any platform is built.
- Dependencies
- Editorial ownership and named reviewers
- Source checking of the priority manuscripts (E01, E03, E07, E09 and their core monographs)
- Demonstration gate
- A reader can move from an essay to a concept to a monograph and back to a section; every item discloses its state; the production build and content integrity checks pass.
- Unresolved question
- Who owns editorial review, and when does the first reviewed manuscript set justify W1 (public access, domain configuration, newsletter operations)?
- Basis
- R§42 to 46; requirements SCP01 to SCP05
W1Editorial launch readinessPlanned
- Current state
- Not started.
- Intended capability
- Priority manuscripts source-checked and approved by named reviewers; contributor identity and responsibility recorded; newsletter sender, delivery service, retention, and unsubscribe handling established; public access and the nuclex.com domain confirmed separately.
- Dependencies
- Reviewed manuscripts
- Designated newsletter operator and delivery provider
- Domain control
- Demonstration gate
- The first approved essays and monographs are public under the canonical domain with author and review records complete.
- Unresolved question
- Whether a headless CMS is needed before editors who do not use version control join the project.
- Basis
- Requirements section 2 and ED04; D016
V1Technical V1: SNOMED CT-anchored terminology foundationPlanned
- Current state
- Prototype only. Two workbench prototypes run as a private deployment at nuclexc.unithera.app: NucLex-storm on Snowstorm and NucLex-owl on Snow Owl, each with Elasticsearch, a reasoner and an authoring layer, holding illustrative sample content. No SNOMED CT content is loaded, no namespace has been issued, and no public terminology endpoint is offered.
- Intended capability
- A terminology server (Snowstorm with Elasticsearch was proposed) serving a NucLex SNOMED CT extension and a FHIR terminology API; an OWL and OBO-style source repository maintained in Git; versioned CodeSystem, ValueSet, and ConceptMap publication; public documentation alongside restricted access to licensed SNOMED CT content. Seed content: radiopharmaceuticals, radionuclides, nuclear medicine procedures, theranostic treatments.
- Dependencies
- Licensed SNOMED CT source access (US Edition RF2 through UMLS)
- Deployment resources (AWS was a hosting proposal)
- Seed modeling rules verified against SNOMED International editorial guidance
- Demonstration gate
- Working SNOMED CT and NucLex terminology operations ($lookup, $validate-code, $expand) on a versioned deployment with authentication.
- Unresolved question
- The official SNOMED CT authoring process and extension rules still need authoritative research; the originating discussion left that question incomplete.
- Basis
- R§1 to 2, 50; PLAT01
V1AAgentic curation layerPlanned
- Current state
- Not built.
- Intended capability
- Literature surveillance, term extraction, suggested definitions, relations, and mappings, evidence and provenance packaging, candidate queues, model and prompt version records, and reviewer packets. AI proposes; humans approve.
- Dependencies
- Technical V1
- A human review interface and named reviewers
- Demonstration gate
- A candidate term with its evidence packet is reviewed and either accepted, revised, or rejected with a recorded rationale.
- Unresolved question
- How to record model and prompt versions so that a proposal can be audited later.
- Basis
- R§10, 32; PLAT02
V2Curated sources and concept composerPlanned
- Current state
- Not built. No connectors to external authorities exist.
- Intended capability
- The three pillars: the SNOMED CT extension; a curated source-authority registry naming the preferred external source by domain (PubChem, FDA GSRS, RxNorm, UniProt, NCI Thesaurus, LOINC, RadLex); and a composer that builds compound expressions from external concepts with strong provenance, backed by resolver connectors and an SSSOM-style mapping store. Compound expressions become official NucLex concepts only after formal review.
- Dependencies
- Technical V1
- Working resolver connectors with source versioning
- Demonstration gate
- A compound concept composed from two external authorities is retrieved with its provenance and mapping record, and is visibly marked as unreviewed until approved.
- Unresolved question
- Mapping confidence and evidence conventions before any merged claim is shown.
- Basis
- R§3 to 4; PLAT03
V3Standards governance workflowPlanned
- Current state
- Partly prototyped. The workbench prototypes NucLex-storm and NucLex-owl (private deployment) implement authoring, internal review by a second person, signed approval, promotion, versioned release candidates and a hash-chained decision log. The submission portal, triage, expert review and public docket are not built. W0 content uses a lightweight brief, draft, source check, domain review, editorial approval workflow in version control.
- Intended capability
- Proposal, triage, authoring, internal review, expert review, public comment or hearing, reconciliation, approval, publication, and maintenance or deprecation, with a submission portal, authoring workbench, public docket, reviewer assignment, decision logs, release candidates, and consensus mechanisms under accountable roles.
- Dependencies
- V2
- Accountable contributor identities (ORCID, roles, permissions)
- Funding for a standards operation
- Demonstration gate
- A proposed concept passes a visible, tracked process from submission to publication with every decision recorded.
- Unresolved question
- How much of a formal standards body the project needs before external adoption justifies it.
- Basis
- R§5 to 6, 38 to 39; PLAT04
V4Public ontology servicePlanned
- Current state
- Not built. No sample endpoint on this site is live.
- Intended capability
- Ontology as a service: versioned FHIR terminology APIs, a native REST API with OpenAPI documentation, an MCP server for AI agents, a developer portal with quick starts and examples, developer keys and rate limits, versioned bulk downloads, and defined service objectives, across public, registered-developer, and trusted-partner tiers.
- Dependencies
- V3
- Operational resources for service-level objectives, monitoring, and security
- Demonstration gate
- An external application retrieves a versioned NucLex concept through the documented API with authentication and rate limiting.
- Unresolved question
- Which tiers and objectives are affordable before grant funding.
- Basis
- R§7, 26; PLAT05
V4+Interactional knowledge for modelsPlanned
- Current state
- Not built. Described in the monograph on interactional ontology as a design direction.
- Intended capability
- Model entities, interactions, parameters, and evidence records carrying context, units, uncertainty, and provenance; digital-twin APIs and MCP tools; investigation of SBML, CellML, and OMEX interoperability; atlas and evidence connectors (HuBMAP, Human Protein Atlas, cBioPortal, NCI GDC, CPTAC, DepMap) and an exploration workbench once source connectors and query infrastructure exist.
- Dependencies
- V4
- Reviewed parameter and interaction records
- Entity resolution and source versioning across atlases
- Demonstration gate
- A prototype model retrieves a reviewed parameter or interaction from NucLex with its provenance, not merely a diagram.
- Unresolved question
- Which interaction representation to adopt or extend (PSI-MI, BioPAX, SBML) rather than invent.
- Basis
- R§8 to 9, 35 to 37; PLAT06 to PLAT08
Deliberately not built now
- Paywall or entitlement system
- Full licensing and reuse machinery
- Full standards body
- Heavy integrations and production ontology APIs
- Atlas ingestion and simulation execution
- Automated outward social publishing
These remain roadmap subjects where they were discussed. Formal rights and reuse machinery was explicitly deferred until the project has funding. The essay From living publication to interactional knowledge (a brief) will develop the evidence that should justify each stage.
