7AAVMAR2 Chap.3 Metadata, FAIR and CARE
Metadata, FAIR and CARE
Records as structured decisions
Week 3 shows how objects and finds become structured data through descriptive and administrative metadata, classification and controlled vocabularies, and how those systems carry institutional assumptions.
This chapter separates object records, image metadata and public labels and maps the seminar's standards to the decisions they support.
Two sets of principles
It then normalises catalogue dates without erasing the source and reads FAIR beside CARE, using the seminar's Hawaiian kapa exercise.
Why this matters for the assessments
Every option in the group brief begins with a spreadsheet of catalogue data, and the selection and description heading asks you to explain which parts of the source records you retain and how you mark your own changes.
The individual piece rewards responsible treatment of provenance, representation and consent.
Metadata decisions are where both are won or lost, because a field you merge, rename or omit changes what visitors can find and whose account they read.
Examples to explore
The session lists Japan Search, which connects national collections through shared metadata; the Towards a National Collection discovery projects linking dispersed UK holdings; the Chengdu Plain Archaeological Survey as an openly published dataset; the Dublin Core user guide; and the FAIR and CARE principles from their publishing bodies.
What this chapter covers
- 01
Descriptive and administrative metadata
- 02
Object, image and label records
- 03
Spectrum, Dublin Core, CCO and Getty
- 04
Normalising dates
- 05
FAIR principles
- 06
CARE principles
Worked example · free
Merge two catalogue terms responsibly
- 2Keep both original terms in their own columns with each institution's identifier and link.
- 2Add a shared term only if a controlled vocabulary justifies it, and cite the vocabulary used.
- 2Add a notes column explaining the decision and any uncertainty in identification.
Key terms
- Dublin Core
- Dublin Core is a set of shared descriptive fields that helps bring records from different sources into one working dataset.
- Spectrum Standard
- Spectrum is the Collections Trust standard for managing and documenting museum collections, including identifiers, sources, rights and changes.
- Normalisation
- Normalisation adds a consistent, comparable value beside each original catalogue value so records can be sorted or searched together.
- Administrative Metadata
- Administrative metadata records information used to manage an item, such as identifiers, rights, acquisition and changes to the record.
Metadata, FAIR and CARE FAQ
How do FAIR and CARE differ?
FAIR tests whether data can be found, accessed, combined and reused. CARE, from Indigenous data governance, tests benefit to the collective, authority to control, responsibility and ethics. A project can satisfy one while failing the other.
Can students apply Traditional Knowledge Labels?
The seminar states that Traditional Knowledge Labels are applied by communities, not assigned by students. Point to Local Contexts as a framework and describe protocols rather than claiming to attach a label.
Why keep the original catalogue wording?
The proposal brief asks you to keep the institution's wording and data visibly separate from your own corrections and groupings. Keeping the original preserves traceability and shows how the record itself frames the object.
Which standard should we use for our working spreadsheet?
The seminar matches tools to decisions: Spectrum for what to retain or link, Dublin Core terms for shared fields, CCO for consistent descriptions and dates, and Getty AAT or TGN for justified shared terms, always keeping the institution's original wording alongside.
Why are catalogues described as embodying power?
The Week 3 session argues that apparently technical systems carry institutional assumptions about how objects are classified and named. Choices about categories and vocabulary decide which perspectives become searchable and which remain hidden in a collection.
How should uncertain dates be recorded?
Keep the catalogue's wording, add numeric start and end fields where the source gives a range, and mark qualifiers such as circa in their own column. A notes field should state that the range expresses uncertainty rather than an exact date.
Assessment move
Take three rows from one of the project spreadsheets and list every field they share. For each, decide whether you would keep it, normalise it in a new column or leave it out, and name the standard that supports the choice. Then write one sentence for each CARE question about the same rows.
Before your group meets, prepare a short list of fields you propose to normalise and the original columns you will keep, so the discussion starts from evidence rather than preference. Compare your list with a partner's and resolve any field you treated differently.