Metadata serving several masters Malcolm Todd malcolm toddnationalarchives

Metadata: serving several masters Malcolm Todd malcolm. todd@nationalarchives. gov. uk © Crown Copyright, 2003

Main drivers • • • Resource discovery e-Govt interoperability Records management Digital preservation …. these are not mutually exclusive categories………… © Crown Copyright, 2003

RD / interoperability • Openness (searching) • System talking to system through configuration to support a schema (usually in XML) rather than fundamental design architecture • Legacy systems • Implementation issues © Crown Copyright, 2003

E-Govt interoperability • More on this tomorrow…… • What is the point of some data being comprehensible by other systems? (could be misleading or risky……. ) • Business case ROI doubtful • Very substantial investment (users. . ) • Very long range benefits © Crown Copyright, 2003

Domains/environments • • • Digital preservation Records management Resource description Specialist environments Different, & differences cause problems…. . • QU: Is ours a specialist environment? Discuss © Crown Copyright, 2003

DC compliant RD metadata Records management (administrative / process) metadata Object level technical metadata RD Layer RM layer DP layer Resolving the domains (clashes) e. g. date, format, ©etc. (some overlaps helpful…. ) Crown Copyright, 2003

Digital preservation • Generally (but not exclusively) at the digital object level • ‘Making explicit in the metadata what was implicit in the original environment’ • e. g. ‘ole’ linkage (© MS) • Facilitate emulation, migration, reconstruction or inference of these properties • Support emulation, migration, management of the preservation system © Crown Copyright, 2003

Records management • Structures • Linkage of objects • Context (BCS, ‘documents forming a record’) • Levels of resource • Evidence, accountability • Meaning © Crown Copyright, 2003

Resource description [1] • • Standard description Flat model: web page paradigm Retrieval / discovery Much of most prominent ‘Metadata’ territory occupied before archivists got to it • Misunderstandings arising & persisting © Crown Copyright, 2003

Resource description [2] • Some prominent initiatives, esp. DCMI (more tomorrow……) • Rules, paradigms, assumptions • ‘Administrative metadata’ dismissal but this is all-important to us • Semantic problems with incorporating ‘specialist’ metadata • e. g. coverage spatial / temporal © Crown Copyright, 2003

Resource description [3] • Issues include: • Archival records • Aggregations of records • Documents [? ] • May or may not equate to digital objects • Capture mechanisms • Source of MD • User compliance • User definition unacceptable • Rules (fixity, encoding) • RD metadata quite inadequate for robust ERM • Google© et al. issue © Crown Copyright, 2003

Resource description [4] Summary to date: “Highly important to work with other professionals in this area, but how do we do this whilst achieving what we need for our archival domain? ” © Crown Copyright, 2003

How to comply with DCMI for a more complex, ‘specialised’ paradigm? • Comply broadly • Comply more fully • Adjust obligation levels • Adjust semantics • Refinements (qualifiers) • Don’t comply • Comply with RD elements • Accept differences • Sub-elements level • Grouping=element © Crown Copyright, 2003

E-gov initiatives • • • Mi. REG [not Mo. REQ compatible……] DC-GOV [more tomorrow……. ] AGLS, GILS e-GMS Varying degrees of compatibility © Crown Copyright, 2003

Problems • DCMI ‘Format’ element • ‘Type’ • ‘Mime’ list encoding scheme obsolete • ‘Administrative’ metadata • ‘Dumbing down’ principle © Crown Copyright, 2003

Are we ‘qualified’? • Qualified (extended) DC • Refinements / qualifiers of main DC elements • Observing ‘dumbing down’ principle • Some DC elements unusable in our domain but something close required • Semantics breaking down (clashes) © Crown Copyright, 2003

How to comply with DCMI for a more complex, ‘specialised’ paradigm? • Comply broadly • Comply more fully • Adjust obligation levels • Adjust semantics • Refinements (qualifiers) • Don’t comply • Comply with RD elements • Accept differences • Sub-elements level • Grouping=element • XML representations will make the deceit clear…… © Crown Copyright, 2003

‘Elements’ • What is an ‘element’? • A ‘grouping of “subelements”’ that are our real elements is the most honest answer. Discuss! • Little actually happens at the element level (apart from some basic description) • Even that description sometimes has its problems (e. g. dumbing down: dates) • Semantics stretched in ways unpalatable to purists (including us): e. g. Rights, Type • Some of our sub]elements behave to very strangely, e. g. Aggregation (ISO 15489) • Rules for subelements have to be in separate part of [looser] schema © Crown Copyright, 2003

Approaches / ‘solutions’ • Carl Lagoze (‘Warwick container architecture’), Cornell University • Referenced by ISO TD work • Different metadata for different domain, same resource • May become more influential • Problems: user time, compliance • ? accept that purist compliance is undesirable and impossible? © Crown Copyright, 2003

Resource description [5] • Repurposing of metadata (e. g. for archival purposes)…… • Descriptive • Every resource requires [valid] description at whatever level • Schemas, not a single schema • Linkage between levels: recursion / IDs • Validation problems © Crown Copyright, 2003

Archival description [1] • Automation (clear those backlogs!) • Robustness • Evidence / diplomatics • Authority • Enrichment • Previous description needs to be retained anyway (reversibility) • Easier for documents than other e-records (!) • Consequences of clashes at design stage © Crown Copyright, 2003

Archival description [2] • • • Websites? ; Non-hierarchical assemblies? ISAD [G] Exotic formats ? Are these problems inevitable anyway? Will it be ‘good enough’? Do we have resources to do anything better? • See Essays in Honour of Michael Cook, LUCAS, July 2003 (Roper, M / Caya, M) • Sheppard, E. & West, V. ; Are ISAD(G) and ISO 15489 compatible? RMJ, Vol 13 (2 parts, issues 1 & 2) © Crown Copyright, 2003

Archival description [3] • Automation of ingestion into preservation systems • Automation of management of those systems: Migration of platforms, possibly formats (checksums, etc. ) • Will initial description stand for the entire continuum? ; e. g. capture (end user / creator), active, semi active business use • Archival (other) re-use • Content management paradigm[? ] © Crown Copyright, 2003

Conclusions • Much to offer all of these environments • Many significant & detailed differences • Some persistent misunderstandings, esp of underlying and continuing needs of government for archives and records management • Terminology, definitions (keep talking!) • DCMI useful but not as dominant as appears • Our ‘Specialism’ needs to become mainstream……. © Crown Copyright, 2003

Mainstreaming: how? • Implementation • Automation • Archives and records management will prove to be the principle implementation tool of metadata standards in government © Crown Copyright, 2003

http: //www. pro. gov. uk/recordsmanagement/e records/2002 reqs/2002 metadatafinal. pdf http: //www. govtalk. gov. uk (schemas area) Thank you. Questions? © Crown Copyright, 2003