Metadata elements

From Archivematica
Revision as of 13:16, 7 September 2010 by Evelyn McLellan (talk | contribs)
Jump to navigation Jump to search

Main Page > Development > Development documentation > Metadata elements

This page identifies a minimum set of metadata elements designed to ensure authenticity and interoperability of preserved objects and to facilitate their retrieval.

This process involves:

  1. Using the InterPARES Chain of Preservation (COP) model and the CoP/PREMIS crosswalk to identify required elements for objects preserved by Archivematica
  2. Analyzing existing metadata in the Archivematica AIP log files and METS.xml file in order to map them to METS and PREMIS elements (see Existing elements)
  3. Comparing 1) to 2) in order to determine what gaps exist in Archivematica
  4. Filling in the gaps - eg by modifying workflow to produce and/or capture missing elements
  5. Structuring the required elements into the Repository eXchange Package (RXP) specification
  6. Determining what metadata belongs in the DIP(s)



Events requiring metadata

Receive SIP (SIP gets placed in 1-receiveSIP)

Metadata for the SIP as a whole

Semantic component Sample value(s) Automated? Notes
2.1.1 eventIdentifierType Y
2.1.2 eventIdentifierValue Y
3.1.1 agentIdentifierType user account Y
3.1.2 agentIdentifierValue demo Y
3.1.1 agentIdentifierType workstation id Y
3.1.2 agentIdentifierValue archivematica-1 Y



Check checksums

Metadata for each file in the SIP

Semantic component Sample value(s) Automated? Notes
2.1.1 eventIdentifierType Y
2.1.2 eventIdentifierValue Y
2.2 eventType Y
2.3 eventDateTime Y
3.1.1 agentIdentifierType software Y
3.1.2 agentIdentifierValue MD5sum Y
2.5.1 eventOutcome Pass; fail Y
2.5.2 eventOutcomeDetail j6059_02.wav FAILED Y


Generate checksums

Metadata for each file in the SIP for which a checksum is generated by Archivematica

Semantic component Sample value(s) Automated? Notes
2.1.1 eventIdentifierType Y
2.1.2 eventIdentifierValue Y
2.2 eventType Y
2.3 eventDateTime Y
3.1.1 agentIdentifierType software Y
3.1.2 agentIdentifierValue MD5sum Y
1.5.2.1 messageDigestAlgorithm MD5 Y
1.5.2.2 messageDigest fa10ee76a575bafe43335abf6cd60bae Y
1.5.2.3 messageDigestOriginator City of Vancouver Y



Review SIP

Semantic component Sample value(s) Automated? Notes
2.1.1 eventIdentifierType Y
2.1.2 eventIdentifierValue Y
2.2 eventType Y
2.3 eventDateTime Y
3.1.1 agentIdentifierType user account Y
3.1.2 agentIdentifierValue demo Y
2.5.1 eventOutcome {pass; conditional pass} {pass; conditional pass} N If it fails, it doesn't move on to become an AIP, so failure is not an option
2.5.2 eventOutcomeDetail Some files missing; appraisal required Some files missing; appraisal required N This field is mandatory if eventOutcome = conditional pass


Quarantine SIP

-when it went in and when it came out

Unpack zipped files

-tool used, time unpacked, event outcome (successful?), map of zipped file to unzipped contents (map for each unzipped file + link to event)

Assign UUIDs

-the usual stuff, map from original name to UUID

Remove prohibited characters=

-the usual stuff, map from original name to sanitized name

Virus scan

-the usual stuff, result for each file (include eventOutcomeDetail to describe type of fail such as the type of malware found)

File characterization

-identification: format name, format version, registry name, registry key -validation: well formed? Valid?

Appraise SIP

-usual event stuff -event outcome (no files removed; some files removed) -list of files removed

Normalization to preservation formats

-everything already in the table plus identification information: format name, format version, registry name, registry key

Normalization to access formats


Mandatory PREMIS elements (mandatory semantic units + mandatory components)

Entity Semantic unit Semantic component Present in Archivematica?
Object 1.1 objectIdentifier 1.1.1 objectIdentifierType No
Object 1.1 objectIdentifier 1.1.2 objectIdentifierValue Yes
Object 1.2 objectCategory none No
Object 1.5 objectCharacteristics 1.5.1 Composition level No
Object 1.5.4 objectCharacteristics/format Either 1.5.4.1 formatDesignation or 1.5.4.2 formatRegistry must be used
  • 1.5.4.1.1 formatName Yes
  • 1.5.4.2.1 formatRegistryName No
  • 1.5.4.2.2 formatRegistryKey Yes
Object 1.7 Storage Either 1.7.1 contentLocation or 1.7.2 storageMedium must be used. However, "if the preservation repository uses the objectIdentifier as a handle for retrieving data, contentLocation is implicit and does not need to be recorded." No, but retrieval may be managed through UUIDs.
Event 2.1 eventIdentifer 2.1.1 eventIdentifierType No
Event 2.1 eventIdentifer 2.1.2 eventIdentifierValue No
Event 2.2 eventType none Partial
Event 2.3 eventDateTime none Partial
Agent 3.1 agentIdentifier 3.1.1 agentIdentifierType No
Agent 3.1 agentIdentifier 3.1.2 agentIdentifierValue No


PREMIS elements relating to derived objects

Since AIPs are constructed from both original and normalized files, we need to determine what PREMIS elements should be used to describe the normalized files and their relationship to the originals.

Original file metadata

Entity Semantic unit Semantic component Example
Object 1.10 relationship 1.10.1 relationship type derivation
Object 1.10 relationship 1.10.2 relationshipSubType is source of
Object 1.10.3 relatedObjectIdentification 1.10.3.1 relatedObjectIdentifierType UUID
Object 1.10.3 relatedObjectIdentification 1.10.3.2 relatedObjectIdentifierValue (UUID of the normalized file)
Event 2.1 eventIdentifer 2.1.1 eventIdentifierType
Event 2.1 eventIdentifer 2.1.2 eventIdentifierValue
Event 2.2 eventType none Normalization
Event 2.3 eventDateTime none 2010:05:19 00:49:15+00:00
Event 2.5 eventOutcomeInformation 2.5.1 eventOutcome Processing completed
Event 2.5.2 eventOutcomeDetail 2.5.2.1 eventOutcomeDetailNote Output #0, wav, to '/tmp/MultimediaSIP-9ece5881-640e-4bdc-9863-4ff50046a0bd/objects/sample.wav': Stream #0.0: Audio: pcm_s16le, 8000 Hz, stereo, s16, 256 kb/s
Agent 3.1 agentIdentifier 3.1.1 agentIdentifierType
Agent 3.1 agentIdentifier 3.1.2 agentIdentifierValue
Agent 3.2 agentName none FFmpeg version SVN-r19352-4:0.5+svn20090706-2ubuntu2.2


Normalized file metadata

Entity Semantic unit Semantic component Example
Object 1.1 objectIdentifier 1.1.1 objectIdentifierType UUID
Object 1.1 objectIdentifier 1.1.2 objectIdentifierValue 270bd067-0483-4c5f-bdec-f2cbd6e651aa
Object 1.10 relationship 1.10.1 relationship type derivation
Object 1.10 relationship 1.10.2 relationshipSubType has source
Object 1.10.3 relatedObjectIdentification 1.10.3.1 relatedObjectIdentifierType UUID
Object 1.10.3 relatedObjectIdentification 1.10.3.2 relatedObjectIdentifierValue (UUID of the original file)
Object 1.10.4 relatedEventIdentification 1.10.4.1 relatedEventIdentifierType
Object 1.10.4 relatedEventIdentification 1.10.4.2 relatedEventIdentifierValue
Object 1.5.2 fixity 1.5.2.1 messageDigestAlgorithm MD5
Object 1.5.2 fixity 1.5.2.2 messageDigest 537e0206ae83f815e4fg5f28464f6rt7