Significant Properties Definition (CPP-022)

CPP-IdentifierCPP-022
CPP-LabelSignificant Properties Definition
AuthorBertrand Caron
ContributorsKris Dekeyser
EvaluatorsFelix Burger, Maria Benauer, Franziska Schwab
Change historyComments
Version 1.0.0 - 2025-08-29Milestone version
Version 1.1.0 - 2026-04-10Migration to XML

1. Description of the CPP

The TDA defines significant properties for sets of Information Objects (i.e., properties that it commits to preserve over the long term through preservation actions and rendition).

Inputs and outputs

Input(s)
Documentation/guidance
Significant properties policy from equivalent TDAs
Community and stakeholders needs
Output(s)
Documentation/guidance
Significant properties policy for a given Object type
Significant properties detection method

Definition and scope

Significant properties is a widely-adopted concept in the analog and digital preservation sectors that can be defined as “essential characteristics of a digital object which must be preserved over time for the digital object to remain accessible and meaningful” (definition taken from the Digital Preservation Coalition, British Library, and JISC joint workshop, “What to preserve? Significant Properties of Digital Objects”). It acknowledges that digital Objects - just like analog Objects - are subject to intentional or unintentional changes and that memory organisations should identify which of their properties are significant, and should therefore be preserved by committing efforts and means. For a more in-depth presentation of the concept of significant properties, see Montague, Lynne, and Eleonora Nicchiarelli. “The Concept of Significant Properties.” The National Archives, Österreichischen Nationalbibliothek, May 26, 2010. Available at https://www.planets-project.eu/docs/reports/Planets_PC3-D23A_TheConceptOfSignificantProperties.pdf .

Significant properties are declared for a certain Object type, although this categorisation will vary. For example, one organisation can declare significant properties for all its still images, while another will specify different significant properties for digital photographs, graphic design works and screenshots. Significant properties are generally classified into five top-level categories:

  • Content: the main signal(s) conveyed by the Object (e.g. textual, visual, audiovisual, etc.);
  • Context: the information documenting the circumstances of creation and history of the Object;
  • Appearance: the visual aspects of the Object (e.g. fonts, layouts, colours, etc.);
  • Structure: the organisation of the content (e.g. pages, sections, ordering, etc.);
  • Behaviour: all features beyond simple displaying (e.g. links, actions, navigation, etc.)

In addition to these categories, significant properties may be qualified by the possibility to be determined in an automated way, or only by manual observation. Significant properties that can be determined in an automated way are called “extractable”. For example, the visual aspect of source and target PDFs can be easily compared by means of tools like diffpdf. On the other hand, those significant properties that only a human action can assess are called “observational”. The step-by-step description below focuses on extractable properties. In cases where observational properties are considered significant by the TDA, it must define a manual determination method that is incorporated in its processes.

Guidance can be provided to memory organisations on which properties of certain Objects are generally considered significant, but the final decision of adopting these or not is still the responsibility of the TDA. Thus, the definition of significant properties is an institutional policy. Significant properties might even vary within the same organisation from one project to another. For example, the history of changes recorded as XMP internal Metadata in field xmpMM:History may be considered crucial for born-digital digital photographs, but negligible for album thumbnails.

The significant properties policy should be produced or at least confirmed by the TDA management, since it will influence the usability that the Objects will support. For this reason, significant properties are likely to be more high-level, organisational properties that will need to be mapped to technical extractable properties. For example, for a significant property “audio signal quality”, the technical properties associated will be sampling frequency, bit depth, etc. These properties might change (e.g. if a conversion from audio signal in DSD64 1 bit / 2.8224 MHz is made to PCM 24 bits / 96kHz) but their evolution should be controlled to ensure the audio signal quality is preserved.

In some cases, significant properties will be dictated by stakeholders (e.g. Producer, Consumer, etc.) - see the use case “Page Layout as Significant Property of Reflowable EPUBs” below. Community Watch (CPP-018) must therefore ensure that its significant properties policy is aligned with the needs of its designated community.

Process description

Trigger event(s)

Trigger EventCPP-identifier
A new Object type is planned to be ingested by the TDA

Step-by-step description

NoSupplierInputStepsOutputCustomer
 sequence
1CPP-018 (Community Watch)Community needsStudy Community needs and derive required Object propertyOrganisational high-level property
2Organisational high-level propertyMap the organisational high-level property to one or several concrete extractable property(ies)Associated concrete extractable property(ies)
3Gather a test set of a given Object typeTest set
4Associated concrete extractable property(ies)Select a tool capable of extracting the concrete property(ies)Extractor tool
5Test setRun the candidate extractor tool on the test setCandidate extractor tool output
Candidate extractor tool
6Candidate extractor tool outputDetermine whether the significant property is identifiable in the candidate extractor tool outputSignificant properties detection method confirmed: extractor tool chosen (step 7)
No detection method confirmed (resume steps 4-6)
7Chosen extractor tool outputDetermine the interpretation of the extractor tool output (path to the property, error tolerance, etc.)Significant properties detection method: path or expression to extract the significant properties from the tool’s outputCPP-009 (Metadata Extraction)

Rationale(s) and worst case(s)

RationaleImpact of inaction or failure of the process
If the TDA aims to do semantic / logical preservation, give access to its holdings to end users and create copies out of the original Objects, defining the significant properties of its Objects to be able to assess the success or failure of all preservation and access actions is required.Without a significant properties policy and methods to detect that they are maintained in any modified preservation or derivative copy, the TDA may produce unnoticed faulty copies or render the Objects incorrectly.

2. Dependencies and relationships with other CPPs

Dependencies

CPP-IDCPP-TitleRelationship description
CPP-018Community WatchThe Significant Properties Definition is fundamental to ensure that information and features supported by the Objects are those that the TDA’s designated community is expecting.

Other relations

RelationCPP-IDCPP-TitleRelationship description
Affinity withCPP-019Data Quality AssessmentAs Data Quality Assessment identifies quality properties whose value will determine whether the Objects are ingested or not, these quality properties will likely be also considered significant by the TDA.
Affinity withCPP-016Metadata Ingest and ManagementA TDA defines significant properties for digital Objects. These are then translated to Technical metadata that is ingested. The significant properties definition process also influences which Technical metadata standards are applied.
Required byCPP-013Object Management ReportingTo report on the characteristics of Objects for deeper analysis, the significant properties must have been defined in the first place.
Required byCPP-014File MigrationFile Migration implies the production of a Representation supposed to act as a preservation copy and must rely on significant properties to determine its success or failure.
Required byCPP-015Emulation and Rendering ToolsRendering should be evaluated based on significant properties. The faulty rendering of an Object because of unsuitable hardware or software may go unnoticed and produce wrong interpretations from the TDA’s community.
Required byCPP-026File NormalisationLike Format migration, Normalisation should be evaluated based on significant propertie
Required byCPP-027File RepairModifying the File’s Bitstream in order to correct faulty structures can cause unexpected changes to the rendering and behaviour of an Object. Thus, file repair should also be evaluated based on significant properties.
Required byCPP-028Creation of DerivativesAlthough derivatives may not be subject to the same requirements as preservation Representations, a TDA might consider providing access through copies that do not convey significant properties as a failure.

4. Reference implementations

Use cases

Page Layout as Significant Property of Reflowable EPUBs

Institutional background
InstitutionEbooks publishers / digital books service provider, FR
Hyperlink
Description
Trigger eventUsers of reflowable EPUBs may change their visual aspect by zooming or switching to another font.
Problem statementIn 2009-2010, French publishers were implementing workflows to produce digital books. Initially, they were concerned by the users’ ability to change the digital books’ visual aspect and page layout. They were tempted to use exclusively the EPUB 3 fixed-layout format to prevent users from doing so.
Proposed solutionAn EPUB production service provider had to convince publishers of the benefits of reflowable EPUBs. The publishers had to accept that reflowable ebooks' visual aspect and page layout were not significant.In this example, a stakeholder - a service provider, followed by an industry sector - determined explicitly that a property of a specific Object (digital books in reflowable EPUB format) should not be considered significant. TDAs should then acknowledge this and make no attempt at preserving this property.

Publicly available documentation

InstitutionOrganisation typeLanguageHyperlink
TIB – Leibniz Information Centre for Science and Technology and University Library, DENational library
Non-commercial digital preservation service
Research infrastructure
Research performing organisation
English https://wiki.tib.eu/confluence/spaces/lza/pages/93608961/Significant+Properties
CSC – IT Center for Science Ltd., FINon-commercial digital preservation serviceEnglishhttps://urn.fi/urn:nbn:fi-fe2025040925236
(section 6.2.1)
Bibliothèque nationale de France, FRNon-commercial digital preservation serviceFrenchhttps://bnf.hal.science/hal-03374030v2
(sections titled “Propriétés…”)
National Archives and Records Administration, USNational archiveEnglish https://github.com/usnationalarchives/digital-preservation/tree/master/Digital_Preservation_Record_Categories