Give relationships a precise meaning
A semantic model helps different collections describe what they know without flattening their differences.
The MODAVIS Ontology Network is a modular model for evidence-backed, time-aware knowledge about musical instruments and their representations. It is instrument-neutral, with a dedicated specialization for pipe organs.
Identity comes first
An instrument is distinct from a state of that instrument, a digital surrogate, a playable virtual instrument, a software product and a package. Explicit identities make it possible to connect these things without treating them as interchangeable.
| Area | Examples |
|---|---|
| Objects and context | Instruments, components, configurations, agents, places and uncertain time. |
| Knowledge and history | Evidence, assertions, provenance and documented events. |
| Digital representations | Media, bitstreams, audio, tuning and virtual instruments. |
| Domain and interpretation | Organ specialization, heritage contexts and reviewed vocabulary. |
Cuntz · Released record#
Is the 3D model the Cuntz organ?
An ontology is an explicit agreement about the kinds of things we describe and the relationships between them. Its first useful distinction here is between an instrument and a representation of it.
Work through the example 3 STEPS
The physical Cuntz positive organ is identified as mXp 4010243. The release also contains a geometric model and audio files. They concern the instrument, but each has its own identity.
Name the physical subject
Use the instrument identity for claims about the organ itself. Its record can link to the museum and to documentary evidence.
Name each representation
Give a model, a recording and a playable software product separate identities. A relationship such as “represents this instrument” connects them without declaring them the same thing.
Follow a change
Compressing the model changes its digital realization. It does not rebuild the museum instrument. Repairing a physical component is a different event and needs its own evidence.
BASIS FOR THIS EXAMPLE
- Cuntz Positiv · VAO 0.5.0-rc.2 ↗
Published manifest: identities, measurement observations, realization digests, profiles and rights. Exact examples checked against this release.
- MODAVIS Ontology Network 0.1.0 ↗
Identity, state, configuration, evidence, assertions, context and provenance. Teaching examples explain its distinctions without inventing new normative properties.
The displayed model is a derived visualization with an unverified physical scale. Being about the Cuntz organ does not make every visible detail a measured historical fact.
The useful distinction. An identity answers “which thing?”; a representation relationship answers “what is this about?”.
One organ, several identities
An organ can retain its identity through a documented alteration while its configuration changes. A photograph of that organ, a sampled virtual instrument and a downloadable package are related things with their own identities. Treating all of them as one object makes it difficult to say which date, measurement, right or claim belongs where.
| Identity | A statement that belongs here |
|---|---|
| Physical instrument | An attributed maker or a documented location history. |
| State or configuration | A particular arrangement of components within a historical scope. |
| Representation | A photograph, model or other surrogate and the subject it represents. |
| Playable virtual instrument | Control and sound behaviour supported by a particular implementation. |
| Release or package | Versioned content, files, dependencies and distribution information. |
A modular language
The network’s core and instrument modules establish general identities and configurations. Evidence, assertion and provenance connect statements to their grounds and editorial history. Events describe historical occurrences, while media, audio and virtual-instrument modules address digital representations. The organ module specialises the generic instrument model; vocabulary, context and heritage modules supply further qualified relationships.
This separation lets a collection use instrument-neutral concepts without adopting every organ-specific detail. It also keeps the ontology’s role clear: the model describes meaning and constraints, while a VAO carrier packages exact realizations under an exchange contract. The ontology does not need to import the VAO container to describe an instrument.
Start modelling with a question
Ask a concrete question before choosing classes and properties: “Which source supports this stop assignment for this configuration?” is more useful than “How do I describe the whole organ?” The question exposes the identities, time scope and evidence links that must survive exchange. Use the release’s interpretation guide and constraints to check those distinctions.
Cuntz · Illustrative scenario#
What does a knowledge graph add to a catalogue entry?
A graph stores named things and explicit relationships. RDF expresses a statement as a subject, a relationship and an object; many such statements can share the same subject.
Work through the example 3 STEPS
Consider the Cuntz pipe represented as rank:ged:pipe:key-036. A measurement, a source workbook and an activity can all refer to it without being squeezed into one text field.
Start with a plain sentence
“This observation concerns that pipe.” Here the observation is the subject, the feature-of-interest relationship connects it to the pipe, and the pipe is the object.
Connect the evidence
The observation points to the measurement activity and raw-result workbook realization. The activity names Philipp Hosbach as its agent. These links answer different questions: what was observed, where it came from and who is credited.
Ask a connected question
A reader can follow the pipe to an observation, then to the workbook cell, unit and processing history. That chain is more useful than merely finding the same word in several descriptions.
BASIS FOR THIS EXAMPLE
- Cuntz Positiv · VAO 0.5.0-rc.2 ↗
Published manifest: identities, measurement observations, realization digests, profiles and rights. Exact examples checked against this release.
- MODAVIS Ontology Network 0.1.0 ↗
Identity, state, configuration, evidence, assertions, context and provenance. Teaching examples explain its distinctions without inventing new normative properties.
These sentences explain graph structure. They are not a copy-and-paste RDF serialization or a claim that English relationship labels are normative ontology terms.
The useful distinction. A graph makes relationships addressable and traversable; it does not make the linked assertions automatically true.
Use the distinctions on a concrete record
Start with one claim about one instrument. Identify the physical object, relevant state, source fragment and assertion scope before adding broader links. If a playable virtual instrument is involved, give its product, version and package their own identities. If a vocabulary supplies a classification, retain whether it is source-asserted, contextual or reviewed. These distinctions make an integrated record richer without making its interpretation artificially certain.
Sources & further reading
- MODAVIS Ontology Network 0.1.0
The published semantic model.
- MODAVIS Ontology Network 0.1.0
Release README, fourteen module descriptions and SHACL validation contract.
- OMARO 0.1.0
Release README: source-asserted classification, contextual assertions, reviewed projections and documented SQLite queries.
Page editions 2026-09-r6 ↓
Read a fixed snapshot of this chapter, or return to the current notebook.
Edition 2026-09-r6 · SHA-256
788f4c31343ePage metadata ↗