University College London · FACULTY OF INFORMATION TECHNOLOGY

INST0089 Chap.2 Metadata Properties and Contexts

- one subject, every graph, every model, every mark
6 Chapters3-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 2 of 6 · INST0089

Metadata Properties and Contexts

Metadata Properties and Contexts treats the property-value statement as the basic unit of structured description. A property supplies meaning, the value supplies content and a rule states how the value is formed. The date string 2026-05-14 becomes actionable only when the system knows whether it records creation, publication, modification or coverage.

Clear definitions, expected value types, repetition rules and controlled vocabularies allow records to be searched, sorted, compared and transformed. The chapter also explains why domain context matters. A cultural collection, medicine database, product catalogue and research dataset share some broad properties but require different evidence and actions.

Domain-specific fields are justified by real tasks rather than by a desire to capture everything. Machine processing strengthens actionability, yet technical structure cannot rescue an ambiguous concept. The final design must balance consistent values with the nuance needed for ethical and accurate description.

In this chapter

What this chapter covers

  • 01

    Metadata as structured encoded description

  • 02

    Property-value statements and semantics

  • 03

    Data dictionaries and value rules

  • 04

    Controlled values and free text

  • 05

    Domain-specific metadata requirements

  • 06

    Machine processing and user action

Worked example · free

Design fields for interview management

Q [8 marks]. A research project holds volunteer interviews. Staff track processing and consent, while future researchers search topics and request access. Design a compact metadata set that separates operational and public meanings. The marks here for field design are not a university marking scheme.
Domain field comparisonThe invented dataset shows how a shared title field is joined by different fields for a collection, a product and a research dataset.titlerightsmethodcondition
The radar shape compares four possible properties for an invented domain, illustrating that a shared title core does not remove the need for context-specific fields.
  • 2Define the interview as the record unit and list the internal and research tasks.
  • 2Select a shared core including title, identifier, date, language and rights.
  • 2Add interview-specific duration, consent, transcript relation and access properties.
  • 2Separate workflow status from access status and document value rules for both.
The resulting schema supports topic discovery and responsible access without confusing processing completion with availability. Each field has a definition, value rule and named task. The step weighting is original practice guidance.
Sia tip — Name the user operation beside a proposed property; remove or redefine the property if no operation depends on it.
Glossary

Key terms

Property-value pair
A descriptive statement in which a named property gives a value its meaning.
Data dictionary
Documentation defining properties, value types, repetition and formation rules.
Cardinality
A rule stating whether a property is required and how many times it may occur.
Actionability
The capacity of structured metadata to support a human or machine operation.
FAQ

Metadata Properties and Contexts FAQ

Why is a value meaningless without its property?

The same string can assert different facts. A year might be a creation date, publication date or coverage period, while a name might be a creator, subject or record author. The property and its definition identify the relationship between the value and the resource, allowing users and software to interpret it consistently.

When should I use a controlled vocabulary?

Use controlled terms when consistent matching, filtering or aggregation matters and an appropriate vocabulary represents the community and domain. Keep free text where explanation or nuance is required. Document the vocabulary version and allow a note or local term when the controlled option would erase an important distinction.

Can one Status field track everything?

A broad Status field often mixes workflow, access, rights review and physical condition. Those values lead to different actions and owners, so split them into defined properties with separate vocabularies. The change makes filters meaningful and prevents a public-access value from being grouped with an unrelated processing state.

Study strategy

Assessment move

Revise metadata by writing small data dictionaries rather than memorising isolated definitions. For each property, record its label, definition, value type, cardinality, permitted vocabulary, source and the action it supports. Create sample records that include repeated values, missing values and uncertain evidence, because easy complete records hide design weaknesses. Compare free text with controlled values.

Test whether users can search or group the records, then identify which nuance is lost when values are normalised. Do not assume a controlled list is neutral; ask whose language it reflects and whether the domain needs a local term or explanatory note. Practise separating concepts commonly collapsed into broad fields. Split processing status from access status, condition from rights and event date from deposit date.

Explain how the new properties change an operation such as filtering, request management or validation. Change the domain while keeping the descriptive purpose constant. A product, archival item and research dataset may all need identifiers and titles, but their distinctive attributes and risks differ. Justify each domain-specific property through a user task, not through habit. Finish with a portability test.

Give the dictionary and several records to someone outside the project and ask them to reproduce values without oral guidance. Record every question they ask. Questions about abbreviations, blank states, ambiguous roles or inconsistent dates identify the next documentation improvement. A reusable schema makes its assumptions visible and keeps meaning stable when the original designer is absent.

Review blank values as carefully as filled ones. Decide whether a blank means unknown, not applicable, withheld or pending, and avoid turning those states into one undifferentiated absence. Then perform a small aggregation, such as grouping records by access status or sorting by date. Inspect the outliers and trace them back to property definitions or entry rules.

This connects documentation directly to the behaviour of the dataset and shows where consistency breaks under real use. Record the corrective rule beside each outlier so the same ambiguity does not return during the next entry session.

A+Everything unlocked
Unlocks this Bible + all 32 of your subjects at University College London, plus 1,000+ more Bibles across the AskSia library.
Sia - your INST0089 tutor, unlimited, worked the way the exam marks it
The full 3-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works