Semantic Technologies and Applications COMP5860M
John Stell, School of Computing
j.g.stell@leeds.ac.uk
Lecture Tuesday 5th May 2020 Metadata
1
Outline
Explain what “metadata” is Give examples of metadata The Dublin-core terms
2
What is Metadata?
Metadata means data about data
but that’s not very helpful – metadata is a rather vague term and we don’t talk about metametadata for data about metadata, let alone metametametadata etc
It’s like the class / instance distinction in ontology:
You can’t always tell if something ought to be a class or an instance – it will depend on the context / purpose of model
The data / metadata distinction is based on what is useful for modelling the world in a particular situation.
3
Simple example
Data: the text of a book in some electronic format
Metadata examples: title
author
genre (type of book e.g. fiction, textbook, etc) publication date
ISBN
Uses of this metadata:
Presenting the data
Connecting with other data Searching
4
Presenting the data
Display latest releases (Project Gutenberg)
5
Connecting with other data
other ways to get the data (Google Books)
6
Searching
using metadata to search (bookfinder.com)
This is physical books not electronic versions, but people still often say ‘metadata’
7
Types of Metadata
This is from [Ril17] (references are at end of slides)
8
Descriptive Metadata
This is from [Ril17]
9
Technical Metadata
This is from [Ril17]
10
Preservation Metadata
This is from [Ril17]
11
Rights Metadata
This is from [Ril17]
12
Structural Metadata
This is from [Ril17]
13
Markup Language Metadata
This is from [Ril17]
14
Metadata storage
We are mainly interested in metadata represented in rdf separately from the data itself
but in many cases metadata is embedded in data itself, for example in digital photos, similary in video and audio files, webpages etc
photo metadata presented in photo editing app
15
Metadata storage
webpage metadata in page html source Note the mention of dcterms here
16
Metadata Standards
Syntax
RDF for example
Controlled Vocabularies
Restrict terms used to describe data
Examples: Spotify genres, Internet MIME types, fields in bibtex
Content Standards
Communities often adopt conventions about how to describe things – style guides. Library cataloguing will follow a style; Referencing (Harvard etc) is another example: may specify punctuation, italics, etc
17
Spotify genres (a few) from http://everynoise.com/
18
Dublin Core Terms
The Dublin Core Metadata Element Set (DCMES) grew out of a 1995 meeting in Dublin, Ohio
15 elements today. See next slide.
Also known as “simple Dublin Core” or Dublin Core (DC). Widely adopted due to simplicity and wide application
Expanded version: DCTERMS www.dublincore.org/specifications/dublin-core/dcmi-terms/
19
Dublin Core Metadata Element Set
contributor coverage creator date description format identifier language publisher relation rights source subject title
type
20
Properties in the expanded set of terms
abstract, accessRights, accrualMethod, accrualPeriodicity, accrualPolicy, alternative, audience, available, bibliographicCitation, conformsTo, contributor, coverage, created, creator, date, dateAccepted, dateCopyrighted, dateSubmitted, description, educationLevel, extent, format, hasFormat, hasPart, hasVersion, identifier, . . .
21
Properties in the expanded set of terms
. . . instructionalMethod, isFormatOf, isPartOf, isReferencedBy, isReplacedBy, isRequiredBy, issued, isVersionOf, language, license, mediator, medium, modified, provenance, publisher, references, relation, replaces, requires, rights, rightsHolder, source, spatial, subject, tableOfContents, temporal, title, type, valid.
22
There are also Classes
Agent, AgentClass, BibliographicResource, FileFormat, Frequency, Jurisdiction, LicenseDocument, LinguisticSystem, Location, LocationPeriodOrJurisdiction, MediaType, MediaTypeOrExtent, MethodOfAccrual, MethodOfInstruction, PeriodOfTime, PhysicalMedium, PhysicalResource, Policy, ProvenanceStatement, RightsStatement, SizeOrDuration, Standard.
23
and a vocabulary for Types
Collection, Dataset, Event, Image, InteractiveResource, MovingImage, PhysicalObject, Service, Software, Sound, StillImage, Text
24
Dublin Core User Guide
The Publishing Metadata document here is very helpful and has lots of examples in RDF.
For the Final Assessment you will only need to use a small part
https://www.dublincore.org/resources/userguide/
25
dcterms sometimes requires non-literal values
For creator for example:
ex:myLecture dcterms:title “Lecture on Metadata” ;
dcterms:creator [ foaf:name “John Stell” ].
This rdf turtle example is a way of having a blank node in the rdf graph without giving it a name.
Meaning: the object here is a resource having the foaf:name property “John Stell”
Not allowed: (despite an example yesterday)
ex:myLecture dcterms:title “Lecture on Metadata” ;
dcterms:creator “John Stell”.
26
References
Jenn Riley.
Understanding Metadata. What is Metadata, and what is it for?
NISO Primer. National Information Standards Organization (NISO), www.niso.org, 2017.
https://groups.niso.org/apps/group_public/
download.php/17446/Understanding\%20Metadata.pdf
27