{
"title": "Six responsibilities, one research workflow — Tools & experiences",
"text": "MODAVIS (2026). Six responsibilities, one research workflow — Tools & experiences. Research notebooks, September 2026 · Connected evidence edition. https://modavis.org/editions/2026-09-r6/software/architecture/",
"bibtex": "@misc{modavis_software_architecture_2026_09_r6,\n author = {{MODAVIS}},\n title = {Six responsibilities, one research workflow — Tools & experiences},\n year = {2026},\n month = {09},\n version = {2026-09-r6},\n url = {https://modavis.org/editions/2026-09-r6/software/architecture/},\n note = {Content SHA-256: aa00242a58bcf34d1e5ea8e577c9c1d549556c4bea925d8e06ab5edd00cbd4b6}\n}",
"url": "https://modavis.org/editions/2026-09-r6/software/architecture/",
"sha256": "aa00242a58bcf34d1e5ea8e577c9c1d549556c4bea925d8e06ab5edd00cbd4b6",
"edition": "2026-09-r6",
"date": "2026-09-26",
"canonicalPayload": "{\"book\":{\"number\":\"06\",\"slug\":\"software\",\"title\":\"Tools & experiences\"},\"chapter\":{\"blocks\":[{\"caption\":\"The six framework components divide responsibilities. This diagram is a conceptual overview, not an execution schedule.\",\"kind\":\"pipeline\",\"type\":\"figure\"},{\"headers\":[\"Component\",\"Responsibility\"],\"rows\":[[\"Initiator\",\"Shared schemas, migrations, integrity and data governance.\"],[\"Aggregator\",\"Approved source acquisition, transfer preparation and evidence retention.\"],[\"Processor\",\"Staging, normalization candidates, analysis and controlled canonical projection.\"],[\"Prospector\",\"Discovery of provisional literature, media and authority evidence.\"],[\"Navigator\",\"Source-aware public research projections and citations.\"],[\"Persistor\",\"Approved packaging, deposition and persistent identifiers.\"]],\"type\":\"table\"},{\"text\":\"The framework distinguishes original source evidence, inference, canonical representation, dissemination and preservation. A public view is a projection of a data state; browsing that view should not silently rewrite the underlying evidence.\",\"type\":\"p\"},{\"items\":[{\"href\":\"/notebooks/research/method/\",\"label\":\"Research method\"},{\"detail\":\"Research data, software and publications.\",\"href\":\"https://zenodo.org/communities/modavis/\",\"label\":\"MODAVIS on Zenodo\"}],\"type\":\"links\"},{\"id\":\"supporting-components\",\"text\":\"Supporting components\",\"type\":\"heading\"},{\"text\":\"Monitor observes status and freshness without changing research data or running jobs. The musiXplora Bridge imports and explores versioned person snapshots. These components complement the six core responsibilities.\",\"type\":\"p\"},{\"id\":\"statement-journey\",\"text\":\"Trace one statement through the framework\",\"type\":\"heading\"},{\"text\":\"Consider a source page that attributes a renovation to a named workshop. Acquisition first preserves the page and its source context. Processing can extract the statement, propose a normalized actor identity and attach a candidate role. A reviewed decision determines whether that interpretation becomes part of a canonical projection. Public exploration then presents the attributed result; release preparation fixes a publishable state.\",\"type\":\"p\"},{\"headers\":[\"Boundary\",\"What crosses it\"],\"rows\":[[\"Discovery → acquisition\",\"A candidate source or update proposal, not an accepted historical fact.\"],[\"Acquisition → processing\",\"Identified source evidence and documented transfer records.\"],[\"Processing → canonical projection\",\"An explicitly reviewed interpretation with its supporting evidence.\"],[\"Projection → public access\",\"A research view retaining attribution and qualifications.\"],[\"Approved output → preservation\",\"A checked, versioned package with persistent release information.\"]],\"type\":\"table\"},{\"boundary\":\"The example illustrates component responsibilities without attributing unverified processing history to the actual release.\",\"caseLabel\":\"Cuntz\",\"concept\":\"A processing candidate is a proposed interpretation of source material. Review and publication are additional decisions, with their own evidence and responsibilities.\",\"evidence\":\"Illustrative scenario\",\"expanded\":false,\"id\":\"cuntz-software-boundaries\",\"scenario\":\"Use the Cuntz workbook observation to imagine a new import workflow. This is an explanatory mapping to the Processor architecture, not a claim that the published Cuntz release was built by all current Processor modules.\",\"sources\":[{\"detail\":\"Published manifest: identities, measurement observations, realization digests, profiles and rights. Exact examples checked against this release.\",\"href\":\"https://doi.org/10.5281/zenodo.22151203\",\"label\":\"Cuntz Positiv · VAO 0.5.0-rc.2\"},{\"detail\":\"Identity, state, configuration, evidence, assertions, context and provenance. Teaching examples explain its distinctions without inventing new normative properties.\",\"href\":\"https://doi.org/10.5281/zenodo.22126086\",\"label\":\"MODAVIS Ontology Network 0.1.0\"}],\"steps\":[{\"text\":\"Intake identifies the workbook bytes and source context. Extraction retains the literal cell, property code and unit rather than keeping only a convenient normalized number.\",\"title\":\"Preserve the source\"},{\"text\":\"A normalizer can propose a structured observation and a pipe identity. A reviewer checks the cell interpretation, the identity match and qualifications such as the timestamp note.\",\"title\":\"Inspect the candidate\"},{\"text\":\"An accepted projection keeps the evidence link and decision. A public view can be concise while a fixed release preserves the record needed to re-examine the claim.\",\"title\":\"Publish the accountable result\"}],\"takeaway\":\"Automation can propose a useful structure; an accountable workflow keeps the interpretation and its grounds inspectable.\",\"title\":\"Where does an extracted number become an accepted claim?\",\"type\":\"example\"},{\"id\":\"responsibilities\",\"text\":\"Why separate responsibilities?\",\"type\":\"heading\"},{\"text\":\"A component that finds a promising page does not also decide that its claims are correct. A viewer that displays a convenient summary does not silently rewrite its evidence. A status monitor does not repair a failed job by changing the research data. These boundaries make it easier to identify who owns a decision, what can be replayed and which part of a workflow needs attention.\",\"type\":\"p\"},{\"id\":\"not-just-a-chain\",\"text\":\"A framework is more than an execution chain\",\"type\":\"heading\"},{\"text\":\"Real research revisits earlier stages. A new source can reopen an interpretation; a validation failure can return a candidate for review; a publication decision can restrict a public subset without deleting retained evidence. The diagram describes responsibilities and exchanges rather than a promise that every item passes through one automatic, irreversible sequence.\",\"type\":\"p\"},{\"items\":[{\"href\":\"/notebooks/software/processing/\",\"label\":\"Follow the twelve Processor modules\"}],\"type\":\"links\"},{\"items\":[{\"href\":\"/notebooks/software/directory/#processor-modules\",\"label\":\"All twelve Processor modules\"},{\"href\":\"/notebooks/software/directory/\",\"label\":\"Applications, clients and research collections\"}],\"type\":\"links\"},{\"id\":\"discovery-and-publication\",\"text\":\"Discovery and publication have different contracts\",\"type\":\"heading\"},{\"text\":\"Prospector finds potential publications, media, authority links and updates, then hands proposals into a reviewable workflow. It does not make every discovered identity canonical. Aggregator acquires approved material and preserves its source context. Processor returns candidates and derived evidence. Navigator exposes a research view, while Persistor prepares approved publication packages and stable release information. Initiator maintains the common structural and governance foundations.\",\"type\":\"p\"},{\"text\":\"Monitor contributes a separate operational view of health and freshness. A healthy service does not establish that a historical assertion is correct, and a stale service does not invalidate all retained research evidence. Keeping operational status and scholarly judgment separate makes both easier to interpret.\",\"type\":\"p\"},{\"items\":[{\"detail\":\"Detailed processing responsibilities and their review boundaries.\",\"href\":\"/notebooks/software/modules/\",\"label\":\"Twelve module input/output guides\"}],\"type\":\"links\"}],\"intro\":\"A modular framework makes the route from source evidence to a citable release explicit.\",\"slug\":\"architecture\",\"sources\":[{\"detail\":\"Dissertation submitted to Universität Leipzig, 11 September 2026. Chapters 1–6 and 8; Appendix I.4 for the software and resource inventory. See the dedicated research guides for claim-specific section references. Page numbers refer to the printed manuscript pagination. The manuscript is not distributed by this website.\",\"label\":\"Dominik Ukolov · Musikinstrumente im virtuellen Raum (2026)\"},{\"detail\":\"Project component documentation inspected on 26 September 2026. Used for the division of responsibilities among Initiator, Aggregator, Processor, Navigator, Prospector, Persistor and supporting services.\",\"label\":\"MODAVIS Framework · component registry\"},{\"detail\":\"Project documentation inspected on 26 September 2026; package 0.2.0, an unpublished development candidate. Used for current module responsibilities and limitations, not as a claim of public package availability.\",\"label\":\"MODAVIS Processor · module guide and package README\"}],\"title\":\"Six responsibilities, one research workflow\"},\"date\":\"2026-09-26\",\"edition\":\"2026-09-r6\",\"figures\":{\"pipeline\":\"\"}}",
"payload": {
"edition": "2026-09-r6",
"date": "2026-09-26",
"book": {
"slug": "software",
"title": "Tools & experiences",
"number": "06"
},
"chapter": {
"slug": "architecture",
"title": "Six responsibilities, one research workflow",
"intro": "A modular framework makes the route from source evidence to a citable release explicit.",
"blocks": [
{
"type": "figure",
"kind": "pipeline",
"caption": "The six framework components divide responsibilities. This diagram is a conceptual overview, not an execution schedule."
},
{
"type": "table",
"headers": [
"Component",
"Responsibility"
],
"rows": [
[
"Initiator",
"Shared schemas, migrations, integrity and data governance."
],
[
"Aggregator",
"Approved source acquisition, transfer preparation and evidence retention."
],
[
"Processor",
"Staging, normalization candidates, analysis and controlled canonical projection."
],
[
"Prospector",
"Discovery of provisional literature, media and authority evidence."
],
[
"Navigator",
"Source-aware public research projections and citations."
],
[
"Persistor",
"Approved packaging, deposition and persistent identifiers."
]
]
},
{
"type": "p",
"text": "The framework distinguishes original source evidence, inference, canonical representation, dissemination and preservation. A public view is a projection of a data state; browsing that view should not silently rewrite the underlying evidence."
},
{
"type": "links",
"items": [
{
"label": "Research method",
"href": "/notebooks/research/method/"
},
{
"label": "MODAVIS on Zenodo",
"href": "https://zenodo.org/communities/modavis/",
"detail": "Research data, software and publications."
}
]
},
{
"type": "heading",
"id": "supporting-components",
"text": "Supporting components"
},
{
"type": "p",
"text": "Monitor observes status and freshness without changing research data or running jobs. The musiXplora Bridge imports and explores versioned person snapshots. These components complement the six core responsibilities."
},
{
"type": "heading",
"text": "Trace one statement through the framework",
"id": "statement-journey"
},
{
"type": "p",
"text": "Consider a source page that attributes a renovation to a named workshop. Acquisition first preserves the page and its source context. Processing can extract the statement, propose a normalized actor identity and attach a candidate role. A reviewed decision determines whether that interpretation becomes part of a canonical projection. Public exploration then presents the attributed result; release preparation fixes a publishable state."
},
{
"type": "table",
"headers": [
"Boundary",
"What crosses it"
],
"rows": [
[
"Discovery → acquisition",
"A candidate source or update proposal, not an accepted historical fact."
],
[
"Acquisition → processing",
"Identified source evidence and documented transfer records."
],
[
"Processing → canonical projection",
"An explicitly reviewed interpretation with its supporting evidence."
],
[
"Projection → public access",
"A research view retaining attribution and qualifications."
],
[
"Approved output → preservation",
"A checked, versioned package with persistent release information."
]
]
},
{
"type": "example",
"id": "cuntz-software-boundaries",
"title": "Where does an extracted number become an accepted claim?",
"concept": "A processing candidate is a proposed interpretation of source material. Review and publication are additional decisions, with their own evidence and responsibilities.",
"scenario": "Use the Cuntz workbook observation to imagine a new import workflow. This is an explanatory mapping to the Processor architecture, not a claim that the published Cuntz release was built by all current Processor modules.",
"steps": [
{
"title": "Preserve the source",
"text": "Intake identifies the workbook bytes and source context. Extraction retains the literal cell, property code and unit rather than keeping only a convenient normalized number."
},
{
"title": "Inspect the candidate",
"text": "A normalizer can propose a structured observation and a pipe identity. A reviewer checks the cell interpretation, the identity match and qualifications such as the timestamp note."
},
{
"title": "Publish the accountable result",
"text": "An accepted projection keeps the evidence link and decision. A public view can be concise while a fixed release preserves the record needed to re-examine the claim."
}
],
"takeaway": "Automation can propose a useful structure; an accountable workflow keeps the interpretation and its grounds inspectable.",
"boundary": "The example illustrates component responsibilities without attributing unverified processing history to the actual release.",
"sources": [
{
"label": "Cuntz Positiv · VAO 0.5.0-rc.2",
"href": "https://doi.org/10.5281/zenodo.22151203",
"detail": "Published manifest: identities, measurement observations, realization digests, profiles and rights. Exact examples checked against this release."
},
{
"label": "MODAVIS Ontology Network 0.1.0",
"href": "https://doi.org/10.5281/zenodo.22126086",
"detail": "Identity, state, configuration, evidence, assertions, context and provenance. Teaching examples explain its distinctions without inventing new normative properties."
}
],
"expanded": false,
"caseLabel": "Cuntz",
"evidence": "Illustrative scenario"
},
{
"type": "heading",
"text": "Why separate responsibilities?",
"id": "responsibilities"
},
{
"type": "p",
"text": "A component that finds a promising page does not also decide that its claims are correct. A viewer that displays a convenient summary does not silently rewrite its evidence. A status monitor does not repair a failed job by changing the research data. These boundaries make it easier to identify who owns a decision, what can be replayed and which part of a workflow needs attention."
},
{
"type": "heading",
"text": "A framework is more than an execution chain",
"id": "not-just-a-chain"
},
{
"type": "p",
"text": "Real research revisits earlier stages. A new source can reopen an interpretation; a validation failure can return a candidate for review; a publication decision can restrict a public subset without deleting retained evidence. The diagram describes responsibilities and exchanges rather than a promise that every item passes through one automatic, irreversible sequence."
},
{
"type": "links",
"items": [
{
"label": "Follow the twelve Processor modules",
"href": "/notebooks/software/processing/"
}
]
},
{
"type": "links",
"items": [
{
"label": "All twelve Processor modules",
"href": "/notebooks/software/directory/#processor-modules"
},
{
"label": "Applications, clients and research collections",
"href": "/notebooks/software/directory/"
}
]
},
{
"type": "heading",
"id": "discovery-and-publication",
"text": "Discovery and publication have different contracts"
},
{
"type": "p",
"text": "Prospector finds potential publications, media, authority links and updates, then hands proposals into a reviewable workflow. It does not make every discovered identity canonical. Aggregator acquires approved material and preserves its source context. Processor returns candidates and derived evidence. Navigator exposes a research view, while Persistor prepares approved publication packages and stable release information. Initiator maintains the common structural and governance foundations."
},
{
"type": "p",
"text": "Monitor contributes a separate operational view of health and freshness. A healthy service does not establish that a historical assertion is correct, and a stale service does not invalidate all retained research evidence. Keeping operational status and scholarly judgment separate makes both easier to interpret."
},
{
"type": "links",
"items": [
{
"label": "Twelve module input/output guides",
"href": "/notebooks/software/modules/",
"detail": "Detailed processing responsibilities and their review boundaries."
}
]
}
],
"sources": [
{
"label": "Dominik Ukolov · Musikinstrumente im virtuellen Raum (2026)",
"detail": "Dissertation submitted to Universität Leipzig, 11 September 2026. Chapters 1–6 and 8; Appendix I.4 for the software and resource inventory. See the dedicated research guides for claim-specific section references. Page numbers refer to the printed manuscript pagination. The manuscript is not distributed by this website."
},
{
"label": "MODAVIS Framework · component registry",
"detail": "Project component documentation inspected on 26 September 2026. Used for the division of responsibilities among Initiator, Aggregator, Processor, Navigator, Prospector, Persistor and supporting services."
},
{
"label": "MODAVIS Processor · module guide and package README",
"detail": "Project documentation inspected on 26 September 2026; package 0.2.0, an unpublished development candidate. Used for current module responsibilities and limitations, not as a claim of public package availability."
}
]
},
"figures": {
"pipeline": ""
}
}
}