CIDOC-CRM Intro

...or metadata for CHC

Simple/flat datastructure

{
  "dc:identifier": "5386315a-3eb6-4722-8f21-5a456acd6180",
  "dc:title": "Title",
  "dc:subject": ["Subject1", "subject2"],
  "dc:creator": "Tarje Sælen Lavik",
  "dc:created": "2021-01-01"
}

https://www.dublincore.org/specifications/dublin-core/dces/

Richer datastructure

{
  "@id": "5386315a-3eb6-4722-8f21-5a456acd6180",
  "@type": "dct:BibliographicResource",
  "dct:identifier": "5386315a-3eb6-4722-8f21-5a456acd6180",
  "dct:title": "Title",
  "dct:creator": [
    {
      "@id": "ca837fdd-4cfa-4f0f-8885-33319403842a",
      "@type": "dct:Agent",
      "label": "Tarje Sælen Lavik"
    }
  ]
}

https://www.dublincore.org/specifications/dublin-core/dcmi-terms/

What is CIDOC-CRM?

  • Evidence based ontology originating from the museum sector
  • A conceptual reference model
  • Event based
  • 1:1 or one thing at a time
  • Extendable

Evidence based ontology

  • [From the start...] Only real usecases can be modeled
  • Data from loads of museums and systems analysed
  • The goal was to create a common datamodell for exchange

A conceptual reference model

  • From the start... not a realised datamodel
  • A guide for designing a system or own datamodel

Event based

  • Events is the shared structured between specialised subsystems
  • Stuff exists and something happened to this stuff
  • Metadata about a book are often one or more simple properties, but is a series of distinct events happening to the book in CIDOC-CRM

1:1 or one thing at a time

  • One instance must only contain properties about that instance
  • "Unpacking" data, hidden data structures must be made explicit
{
  "dct:title": "Title",
  "ex:author": "Tarje Sælen Lavik",
  "ex:authorBirthdate": "1877-23-57"
}
{
  "dct:title": "Title",
  "ex:author": [{
    "type": "Person",
    "ex:name": "Tarje Sælen Lavik",
    "ex:birth": {
      "type": "Birth",
      "ex:date": "1877-23-57"  
    }
  }]
}

Extendable

  • Extended with specialised models for digitization, archeology, scientific reasoning +++
  • Not well documented
  • Hard to use and find extensions

Basics

Typing

Object

Photograph

hasType

Evil typing

Object

Photograph

depicts

mode of depiction

Material vs the immaterial

Man-made object

Linguistic object

carries

Long cut

Object

Photograph

assigned

Attribute Assignment

was attributed by

carried out by

Person

has type

Class hierarchy

Man-made object

Persistent item

Entity

Thing

Man-made thing

Physical man-made thing

Legal thing

Property hierarchy

Man-made object

Persistent item

Entity

Thing

Man-made thing

Physical man-made thing

is identified by

---

has dimension

has title

depicts

---

Class hierarchy

Material vs the immaterial

Made object vs Linguistic object

isSubjectOf

image/svg+xml
  1. canvas
  2. canvas
  3. canvas

carries

Property hierarchy

  • P11 had participant

    • P14 carried out by

      • P22 transferred title to
      • P23 transferred title from
      • P28 custody surrendered by
      • P29 custody received by
      • P10 is participation of
      • R43 carried out by
      • R44 carried out by

End product

Evils of datamodelling

  • Classitis
  • Propspropsporpspsops

LOUD

Linked Open Usable Data

  • CIDOC-CRM is complex!
  • CIDOC-CRM is too far removed from real world needs
  • CIDOC-CRM is not a standard
  • ...but very precise and inspiring

LOUD

linked.art

LOUD @ UiB-UB

=

JSON-LD + IIIF

+

CIDOC-CRM Patterns

CIDOC-CRM introduction

By Tarje Lavik

CIDOC-CRM introduction

  • 270

More from Tarje Lavik