Metadata Model (MJW Fedora Archive)
Audience: Library integrators (primary) · Curators (secondary) · Technical (secondary)
Use this doc when you need to: understand exactly which predicates/fields are written to Fedora during ingest and how to map them into other systems.
Related:
- Integrator onboarding: Library Access Guide
- Curation rules: Curator's Guide
This project stores descriptive metadata in Fedora using Dublin Core Terms (DCTERMS) plus a small MJW project namespace for fields that are not covered cleanly by DCTERMS.
The Workers primarily write metadata during ingest in the submit-ingest Worker.
Prefixes / Namespaces
dcterms:→http://purl.org/dc/terms/mjw:→http://michaeljwright.com.au/ns#
Fedora also adds system-managed predicates (created/modified/by, etc.) which should be treated as read-only.
Dublin Core Terms (DCTERMS) used
Each row lists the tag/prefix used in JSON-LD/Turtle and the canonical URI.
| Tag | URI | Meaning / How we use it |
|---|---|---|
dcterms:identifier |
http://purl.org/dc/terms/identifier |
Stable Catalog ID (e.g., MJW-P-1987-042). This is the citation-safe identifier and should not be repurposed for display titles. |
dcterms:title |
http://purl.org/dc/terms/title |
Human-readable title for the item. Collected per-file in ingest (defaults to filename if blank). |
dcterms:creator |
http://purl.org/dc/terms/creator |
Creator/author (currently set to “Michael J Wright” for items ingested via the Worker). |
dcterms:created |
http://purl.org/dc/terms/created |
Item date created. Supports full date (YYYY-MM-DD) when supplied; otherwise falls back to year (YYYY). |
dcterms:type |
http://purl.org/dc/terms/type |
High-level content type label (Painting/Drawing/Sculpture/Photograph/Poem/Notebook). |
dcterms:rights |
http://purl.org/dc/terms/rights |
Rights statement / copyright notice. |
dcterms:description |
http://purl.org/dc/terms/description |
Description/notes (bulk default with optional per-item override). |
dcterms:subject |
http://purl.org/dc/terms/subject |
Keywords/subjects (stored as multiple values; entered comma-separated in ingest). |
dcterms:medium |
http://purl.org/dc/terms/medium |
Medium/material/format (e.g., “Oil on canvas”). |
dcterms:extent |
http://purl.org/dc/terms/extent |
Size/dimensions (free text, e.g., “76 x 101 cm”). |
dcterms:isPartOf |
http://purl.org/dc/terms/isPartOf |
Series/collection membership name (free text “Series/Collection Name” from ingest). |
dcterms:source |
http://purl.org/dc/terms/source |
Provenance source reference (typically a SharePoint-relative path + standardized filename) when provided. |
dcterms:temporal |
http://purl.org/dc/terms/temporal |
Collection/batch date range. Stored as either a single date (YYYY-MM-DD) or an interval (YYYY-MM-DD/YYYY-MM-DD). |
dcterms:spatial |
http://purl.org/dc/terms/spatial |
Reserved for future use (place/location). |
MJW project namespace (mjw:*)
These fields exist to keep the model strict and machine-friendly where DCTERMS would be ambiguous or too coarse.
| Tag | URI | Meaning / How we use it |
|---|---|---|
mjw:condition |
http://michaeljwright.com.au/ns#condition |
Condition statement (Excellent/Very good/Good/Fair/Poor). |
mjw:sourceFolder |
http://michaeljwright.com.au/ns#sourceFolder |
SharePoint-relative folder path captured at ingest time (no host/scheme). |
mjw:sourceFilename |
http://michaeljwright.com.au/ns#sourceFilename |
Original filename from upload. |
mjw:standardizedFilename |
http://michaeljwright.com.au/ns#standardizedFilename |
Standardized filename used for the Fedora binary (e.g., {CATALOG_ID}_web.jpg). |
mjw:collectionStart |
http://michaeljwright.com.au/ns#collectionStart |
Structured start date for the batch/collection (YYYY-MM-DD). |
mjw:collectionEnd |
http://michaeljwright.com.au/ns#collectionEnd |
Structured end date for the batch/collection (YYYY-MM-DD, defaults to start date when omitted). |
Filename conventions (binary naming)
Ingest standardizes the Fedora binary filename based on the Curator’s Guide naming reference:
- Visual previews:
{CATALOG_ID}_web.<ext>(with.jpegnormalized to.jpg) - Poems:
{CATALOG_ID}.txt(or preserves the uploaded extension when present)
Notebook page naming patterns (e.g., _p001_...) are not yet captured because page numbering is not currently collected by the ingest UI.
Notes on stability
- The Fedora container slug / URL and
dcterms:identifierare treated as stable once created. - When backfilling historical items, prefer assigning the correct Catalog IDs from an external register (paste into the ingest UI) rather than trying to derive year ordinals automatically from upload order.