Upload
darleen-owen
View
217
Download
0
Tags:
Embed Size (px)
Citation preview
Collections revealed:the role and practical application of
collection descriptionsCIMI Forum,
National Museums of Scotland, Edinburgh, 31 October 2002
Pete Johnston
UKOLN, University of Bath
Bath, BA2 7AY
UKOLN is supported by:
[email protected]://www.ukoln.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
2
Collections revealed: the role and practical application of collection description
• The resource discovery context• Collections and services• Describing collections• Collection-level description in practice
• A case study: CLD at The Natural History Museum
• CIMI and CLD
CIMI Forum, Edinburgh, 31 October 2002
3
UKOLN & collection-level description
• Centre of expertise in digital information management
– Core funding from JISC and Re:source– Policy & advice
– Influencing policy– informing practice
– Research & development– Advancing knowledge– Wide range of areas
– Services – Information services (Ariadne, Cultivate Interactive)– Technical support/joint administration of Resource
Discovery Network (RDN)
http://www.ukoln.ac.uk/ http://www.ukoln.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
4
UKOLN & collection-level description
• JISC eLib programme– range of digital library development projects– integrated access to distributed resources– architectural studies (MODELS) – simple collection description schema
• RSLP Collection Description project– Analytical Model, RSLP CD Schema
• Collection Description Focus– funded by RSLP, JISC, British Library, Re:source– “improve consistency/compatibility of approaches”– awareness-raising, consensus-building– support to implementers– dissemination of good practice
CIMI Forum, Edinburgh, 31 October 2002
5
The resource discovery context
• Initiatives to improve access to services– political/economic drivers
• High-level strategies– UK “Modernising government” White paper– e-Government strategies
• Better connectivity – public library networking through “People’s
Network”– recognition of multiple channels of access
• Improved resource disclosure – e.g. Full Disclosure, Research Support Libraries
Programme, Archives Hub, Access 2 Archives, Scottish Archival Network etc
CIMI Forum, Edinburgh, 31 October 2002
6
The resource discovery context
• Collaborative approaches to service provision and resource management
• Move away from “silo mentality”• Towards more “joined-up” approaches
– Within museum domain– Across domains
• Embedded in strategic initiatives for museums
– Re:source (England) – Single Regional Agencies– “Preserving the Past for the Future”: Framework for
Collections Management– Cross-domain projects
CIMI Forum, Edinburgh, 31 October 2002
7
The resource discovery context
• Growth of digital content creation programmes
– making heritage (more) accessible– NOF-Digitise
– £50m content creation programme– Over 1 million digital objects– supporting strategy for social inclusion, lifelong
learning
– Culture Online– “to widen access to resources of arts/cultural sector
for purposes of learning and enjoyment”
• Digital collections not “stand-alone”– Value as reusable components
CIMI Forum, Edinburgh, 31 October 2002
8
The resource discovery context
• Functional separation of content provision and presentation
– user wants information relevant to task/activity – may see structural/organisational boundaries of
content providers as unimportant!
– content providers exposing content through multiple services, channels
– service providers “surfacing” content from multiple (distributed) sources
• … from web sites to “portals”– “A network service that provides a personalised,
single point of access to a range of heterogeneous network services, local and remote, structured and unstructured”
– Powell 2002
CIMI Forum, Edinburgh, 31 October 2002
9
Historic Environment Information Resources Portal (HEIRPORT)
http://ads.ahds.ac.uk/heirport/
Historic Environment Information Resources Portal (HEIRPORT)
http://ads.ahds.ac.uk/heirport/
Cross-searching metadata databases of ADS, RCAHMS, SCRAN, Portable Antiquities
Z39.50 search/retrieval protocol
Dublin Core (in XML)
Cross-searching metadata databases of ADS, RCAHMS, SCRAN, Portable Antiquities
Z39.50 search/retrieval protocol
Dublin Core (in XML)
CIMI Forum, Edinburgh, 31 October 2002
10
The resource discovery context
• Wide implementation of XML-based standards/specifications for data exchange
• Open Archives Initiative Protocol for Metadata Harvesting Version 2.0
– roots in scholarly publication community, but deployed more widely
– Version 2.0 as stable release– lightweight protocol enabling sharing of metadata
records – using XML over HTTP– alternative to distributed search model– Must provide simple/unqualified Dublin Core record
(OAI provides XML Schema)– May provide other metadata formats (in XML)
CIMI Forum, Edinburgh, 31 October 2002
11
OAIRepository
A
OAIRepository
B
Harvestvia OAI PMH
Portalsite C
Website A
Collection of digital
metadata records
Collection of physical
items
Website BCollection of
digitalmetadata records
Collection of digitalitems
OAIRepository
X
CIMI Forum, Edinburgh, 31 October 2002
12
Social Science Information Gatewayhttp://www.sosig.ac.uk/
Social Science Information Gatewayhttp://www.sosig.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
13
CIMI Forum, Edinburgh, 31 October 2002
14
Resource Discovery Network Resource Finder http://www.rdn.ac.uk/
Resource Discovery Network Resource Finder http://www.rdn.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
15
CIMI Forum, Edinburgh, 31 October 2002
16
The resource discovery context
• CIMI Dublin Core testbed– Evaluate use of DCMES for describing range of
museum objects– 185,000 simple DC records from 10 contributors
• CIMI Metadata Harvesting using OAI PMH– Development of repository– Harvesting metadata– Making aggregated metadata available for harvest
• CIMI XML Schema for SPECTRUM– Recently made available for review (Oct 2002)– Semantics based on mda SPECTRUM standard– Enables encoding of rich descriptive information– Objects, related people, places, events
CIMI Forum, Edinburgh, 31 October 2002
17
The resource discovery context
• Access… • Integration…• Collaboration…. • “Metadata is data in the system which allows
people and machines to do smarter things”
“Interoperability as recombinant potential”(Dempsey, 2002)
• The whole is more than the sum of the parts
Collections and services
CIMI Forum, Edinburgh, 31 October 2002
19
What is a collection?
• Collection– “an aggregation of items”
• Aggregations of, e.g.– natural objects: fossils, mineral samples…– created objects: artefacts, documents, records…– digital resources: documents, images, multimedia
objects, data, software…– digital surrogates of physical objects: documents,
images…– metadata: catalogue records, item descriptions,
collection-level descriptions (!)…
CIMI Forum, Edinburgh, 31 October 2002
20
What is a collection?
• Various criteria for aggregation, e.g.– By location– By type/form of item– By provenance of item– By source/ownership of item– By nature of item content– ….
• Any number of items• Permanent, temporary• Discrete, distributed• Collections created with intent/purpose
– “consciously formed”– collection development policies
CIMI Forum, Edinburgh, 31 October 2002
21
What is a collection?
• Functional granularity– “[T]here is no structure inherent in the model that
requires or predisposes a particular level of aggregation. The institution should base its choices on its own pragmatic grounds, such as the level of detail required to make explicit those elements of the Collection-Description that the institution deems to be useful or necessary for the purposes of resource discovery or collection management. i.e. institutions should adopt a functional granularity approach.”
– Heaney, 2000
CIMI Forum, Edinburgh, 31 October 2002
22
Collections and services
• Collections are made available to users through services
• Museum collections – collections of physical objects/items– collections of digital objects/items– collections of metadata records
– describing physical objects
– describing digital objects
• Museum services – physical services– digital/network services
CIMI Forum, Edinburgh, 31 October 2002
23
Physicalservice
Physicallocation
Physical services make physical collections available at physical locations
Collection of physical
items
CIMI Forum, Edinburgh, 31 October 2002
24
Collection of digitalitems
Website
Networkservice
Digitallocation
Network services make digital collections available at digital locations
CIMI Forum, Edinburgh, 31 October 2002
25
Physicalservice
Catalogueinterface
Digitallocation Network
service
Digital catalogue (Collection of digitalmetadata records)
Physicallocation
Physical services make physical collections available at physical locations
Collection of physical
items
CIMI Forum, Edinburgh, 31 October 2002
26
Digitallocation
Collection of digital
metadata records
Collection of digitalitems
Website
Networkservice
Digitallocation
Network services make digital collections available at digital locations
CIMI Forum, Edinburgh, 31 October 2002
27
User wants to know…
• Which collections are relevant to their requirement?
– subject/coverage of items?– type?– legal status?– conditions of access/use?– etc
• What services make those collections available?
– location?– access?– etc
CIMI Forum, Edinburgh, 31 October 2002
28
OAIrepository
Harvestvia OAI-
PMH
Z39.50target Search/retrieve
via Z39.50
Website
Collection of digital
metadata records
SOAPreceiver
operationsvia SOAP
unstructured network service
structured network service
Collections of digital metadata records made available through multiple network services
CIMI Forum, Edinburgh, 31 October 2002
29
“User” wants to know…
• Which collections are relevant to their requirement?
• What services make those collections available?
• “User” may be human researcher or software tool
CIMI Forum, Edinburgh, 31 October 2002
30
The problem
“We’ve created this incredible constellation of collections, of pools of information accessible through the Net. And people can’t find which pool to look in”
– (Lynch, 2002)
Describing collections
CIMI Forum, Edinburgh, 31 October 2002
32
Collection description
• Analytic– info about items in collection and their content
• Indexing– info derived from items in collection
• Hierarchic– info about collection as whole, and about items and
their content (and relationships between items and whole)
• Unitary– info about collection as whole, not about items– “collection-level description”
– (typology from Heaney 2000)
CIMI Forum, Edinburgh, 31 October 2002
33
Why collection-level description?
• Enable collection provider to– disclose information about collections to potential
users– overview of otherwise uncatalogued items
– summary where item-level detail inappropriate/unavailable
– manage collections– audit/review holdings internally
– manage in collaboration with other providers
– assess priorities for item-level cataloguing
– inform strategic planning – institutional, cross-institutional, regional, sectoral,
national….
CIMI Forum, Edinburgh, 31 October 2002
34
Why collection-level description?
• Enable user to – discover/locate collections– select collections to explore/search on basis of
summary description e.g.– content/coverage, access conditions, resource type
– compare collections as broadly similar objects (even where items heterogeneous)
– understand conditions of access & use– interpret collections
– provenance, context, relationships
• Enable software agents to – select collections to search on behalf of user
– e.g. on basis of profile/preferences
– control searches across collections
CIMI Forum, Edinburgh, 31 October 2002
35
What information should collection-level description provide?
IMLS on describing collections…• “Collections should be described so that a
user can discover important characteristics of the collection, including scope, format, restrictions on access, ownership, and any information significant for determining the collection’s authenticity, integrity and interpretation.”
IMLS Framework of Guidance for Building Good Digital Collections
CIMI Forum, Edinburgh, 31 October 2002
36
Approaches to CLD in archives
• “Collections” defined by provenance of (unique, physical) items
– records of organisation or individual– principle that value of individual record derives
from context, relationships
• (Mainly) physical collections made available by physical services
• Archival description– emphasis on “multi-level” resource description– hierarchical collection description– well-established standards e.g. ISAD(G), EAD
CIMI Forum, Edinburgh, 31 October 2002
37
Approaches to CLD in archives
• Encoded Archival Description (EAD)– SGML/XML Document Type Definition for
encoding archival finding aids– Owned/maintained by Society of American
Archivists/Library of Congress – Hierarchical finding-aids– “Structured document” approach– Designed to encompass broad range of existing
practice
• Basis of established services– Online Archive of California– RLG Archival Resources– UK Public Record Office, Archives Hub, A2A etc
http://www.loc.gov/ead/ http://www.loc.gov/ead/
CIMI Forum, Edinburgh, 31 October 2002
38
Approaches to CLD in libraries
• Focus on description of (non-unique, physical) item
– well-established standards (MARC, AACR2)– shared cataloguing – emphasis on discovery
• Collections defined by– location– subject
• Until recently, CLD informal, unstructured– some use of MARC for CLD (especially in USA)– deployment of RSLP CD schema by RSLP projects
CIMI Forum, Edinburgh, 31 October 2002
39
Approaches to CLD in libraries
• RSLP Collection Description project• Michael Heaney, An Analytical Model of
Collections and their Catalogues– Entity-Relationship model– Implementation independent– Based mainly on library/archival view of ‘collection’– but intended to be applicable across wide range of
collection types
• Functionally (IFLA FRBR) concerned with : – Finding (provide access points for discovery)– Identifying (describe sufficiently well for user
interpretation)
http://www.ukoln.ac.uk/metadata/rslp/ http://www.ukoln.ac.uk/metadata/rslp/
CIMI Forum, Edinburgh, 31 October 2002
40
Approaches to CLD in libraries
• RSLP Collection Description schema– Andy Powell (UKOLN) – structured set of metadata attributes– some simplification of model– description of subset of entities in model– attributes based on Dublin Core Element Set
where possible
• RSLP CD schema supports creation of “unitary” collection description
– simple high-level description– not a substitute for richer CLD schema
• RSLP CD instance – set of linked descriptions of several resources
CIMI Forum, Edinburgh, 31 October 2002
41
RSLP CD Model and Schema (simplified)
ContentCreatorcreates
Collector
Owner
collects
owns
Administratoradministers
ItemProducerproduces
is-embodied-in
Collection
is-gathered-into
Location
is-located-in
CIMI Forum, Edinburgh, 31 October 2002
42
Approaches to CLD for digital resources
• Some description of aggregates of resources– use of general metadata schemas (e.g. DC, GILS)– application-specific, protocol-specific approaches
• Evolution of approaches to creating digital collections
– “proof of concept” (technological focus?)– greater attention to custodianship, use– focus on integration, reuse, interoperability,
sustainability– (Cole 2002, Besser 2002)
• Integration requires shared conventions for talking about collections
– growing interest in collection-level metadata
CIMI Forum, Edinburgh, 31 October 2002
43
Approaches to CLD in museums
• Focus on description of (unique, physical) object
– for management more than discovery?
• But concept of “collection” is present• Various criteria
– type/form of item– subject– ownership/source– audience/purpose
• Some CLD (maybe not called CLD…!)– e.g. guides to holdings, directories, gateways
CIMI Forum, Edinburgh, 31 October 2002
44
Approaches to CLD in museums
• CIDOC Conceptual Reference Model draft v3.3.2, 31 July 2002
– E78 Collection– Subclass of: Physical Man-Made Stuff – Scope Note: This entity describes an aggregate of
items, which is maintained by an Actor following a plan of cultural relevance over time. Things may be added or taken out of a collection in pursuit of this plan. A collection is designed for a certain public, and the conservation of the collected items is normally catered for.
– (emphasis added)
CIMI Forum, Edinburgh, 31 October 2002
45
FENSCOREhttp://fenscore.man.ac.uk/
FENSCOREhttp://fenscore.man.ac.uk/
set up in early 1980'sCLDs for UK biological collections, other than those in the NHM
set up in early 1980'sCLDs for UK biological collections, other than those in the NHM
CIMI Forum, Edinburgh, 31 October 2002
46
Cornucopiahttp://www.cornucopia.org.uk/
Cornucopiahttp://www.cornucopia.org.uk/
Service provided by Re:source
1800 institution-level records, collection-level data from regional “collection mapping” exercises
Current work to expand coverage, enhance interoperability
Service provided by Re:source
1800 institution-level records, collection-level data from regional “collection mapping” exercises
Current work to expand coverage, enhance interoperability
CIMI Forum, Edinburgh, 31 October 2002
47
24 Hour Museumhttp://www.24hourmuseum.org.uk/
24 Hour Museumhttp://www.24hourmuseum.org.uk/
Launched by mda/Campaign for Museums, now independent charity with funding from DCMS/Re:source
Institution records, plus events/articles/educational resources; virtual exhibits
Potential for embedding CLDs from Cornucopia
Launched by mda/Campaign for Museums, now independent charity with funding from DCMS/Re:source
Institution records, plus events/articles/educational resources; virtual exhibits
Potential for embedding CLDs from Cornucopia
CIMI Forum, Edinburgh, 31 October 2002
48
Approaches to CLD in museums
• Project/service specific approaches– Informal?– Limited standardisation?
• However, growing interest in– repurposing management data to support disclosure– sharing data across systems
• Requires harmonisation of approaches– some use of Encoded Archival Description DTD
– Museums in the Online Archive of California– The Natural History Museum
– some use of RSLP CD schema– Crossroads (West Midlands), Find It In London
– building consensus on community schema e.g. BioCase
Collection-level descriptionin practice
CIMI Forum, Edinburgh, 31 October 2002
50
Collection-level description & Research Support Libraries Programme
• Support for academic researchers– disclosure of collections– discovery of/access to collections– collaborative management of collections
• Collections in RSLP– projects describing primarily collections of
physical items (library/archive)– projects also describing digital catalogues (which
describe physical items) – collections of metadata records
http://www.rslp.ac.uk/ http://www.rslp.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
51
Collection-level description & Research Support Libraries Programme
• Adoption of – RSLP CD Schema and/or – ISAD(G)/EAD– Some project-specific schemas
• Web-based discovery services built on databases of CLDs
– subject/discipline-based e.g.– Backstage: Theatre Studies– CASBAH: Caribbean Studies/Black & Asian History – Cecilia: Music collections– … and many more…
– national/regional
• Presently no cross-programme search/aggregation service
CIMI Forum, Edinburgh, 31 October 2002
52
Scottish Collections Network Extensions (SCONE)http://scone.strath.ac.uk/
Scottish Collections Network Extensions (SCONE)http://scone.strath.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
53
Research and Special Collections Available Locally (RASCAL)http://www.rascal.ac.uk/
Research and Special Collections Available Locally (RASCAL)http://www.rascal.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
54
Collections Waleshttp://www.mappingwales.ac.uk/
Collections Waleshttp://www.mappingwales.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
55
Collection-level description & NOF-digitise
• Projects creating digital collections– collections of digital resources– collections of metadata records
• CLD as mechanism for disclosure/discovery– NOF-digi technical standards recommend use of
RSLP CD Schema– Use in NOF-digitise portal– (Re) use in other services/environments
• Initial portal not providing item-level search– but simple DC item-level metadata available– potential for e.g. OAI-based approach
CIMI Forum, Edinburgh, 31 October 2002
56
Collection-level description & the JISC Information Environment
• Content made available as collections– various content providers
• Physical collections– of physical resources (e.g. books, journals)
• Digital collections– of digital resources (texts, images, multimedia
objects, software, datasets, “learning objects” etc) – of digital metadata records
– describing physical items, digital items, physical collections
– metadata record contains identifier/locator of resource
• Users access content through services
http://www.ukoln.ac.uk/distributed-systems/jisc-ie/arch/ http://www.ukoln.ac.uk/distributed-systems/jisc-ie/arch/
CIMI Forum, Edinburgh, 31 October 2002
57
End-user needs to join services together manually - as well as learning multiple user interfaces
End-user
Currently…. Content providers
Authentication
Authorisation
Shared services
Web Web Web WebWeb
CIMI Forum, Edinburgh, 31 October 2002
58
Using Collections in the JISC Information Environment
• HTML Web sites– Aimed at human reader not software tool– Different user interfaces, different metadata
schemas– Researcher “joins up” services manually
• The portal solution– task/user-centred– single point of access to range of heterogeneous
network services
• The “IE service registry”– Database of collection-level descriptions,
service descriptions
CIMI Forum, Edinburgh, 31 October 2002
59
Broker/Aggregator
Portal PortalPortal
End-user
Collection Desc
Service Desc
Shared services
Service registry
Broker/Aggregator
Content providersThe vision….
User Profiles
End-user is “automatically” presented with relevant resources
Web
Fusionlayer
Presentationlayer
Provisionlayer
CIMI Forum, Edinburgh, 31 October 2002
60
Collection-level description & BioCASE
• BioCASE = Biological Collections Access Service for Europe
– Partnership of 31 countries funded by European Commission under Framework V
• Biological collections large– Item-level description time-consuming
• BioCASE XML Schemas for– Unit-level (specimens/observations) metadata– Collection-level metadata
http://www.biocase.org/ http://www.biocase.org/
CIMI Forum, Edinburgh, 31 October 2002
61
Collection-level description & BioCASE
• National nodes collecting information about biological collections in that country
• The Natural History Museum is UK National Node
– Pool records from – NHM CLD project,
– FENSCORE,
– other organisations
• Searches on BioCASE portal will return– unit-level records where available from
participating databases– collection-level records
CIMI Forum, Edinburgh, 31 October 2002
62
Surveying the landscape
• Collections as broadly similar objects– across domains– even if constituent items, approaches to
management etc. differ– easier to harmonise collection-level metadata?
• Disclosure via collection-level description achievable
– where item-level description not feasible/appropriate
• CLD not a substitute for item-level description– complementing item-level discovery – enabling item-level discovery by providing gateway
to item-level metadata
CIMI Forum, Edinburgh, 31 October 2002
63
Surveying the landscape
• CLDs as resource discovery metadata• CLDs support “survey of information
landscape”– “to identify areas rather than specific features - to
identify rainforest rather than to retrieve an analysis of the canopy fauna of the Amazon basin”
(Heaney, 2000)
– detailed item-level “map” not appropriate in all contexts
• The “navigator” of the landscape may be a human researcher or a software tool
CIMI Forum, Edinburgh, 31 October 2002
64
CIMI and CD Focus survey
• Collaboration on survey questionnaire, Sep/Oct 2001
– current activity in CLD– standards/specifications used for CLD
– schemas
– terminologies, thesauri etc
– how schemas/specs used/deployed– approaches/technologies/tools– problems
• Broad target audience– international– not limited to museums
CIMI Forum, Edinburgh, 31 October 2002
65
CIMI and CD Focus survey
• 106 respondents• Mostly working on library/archive collections• Emphasis on resource disclosure/discovery• Outside archival community, little consensus
on descriptive schemas– local/project-/community-specific schemas– extensions to standard schemas– different schemas for different classes of resource– legacy (unstructured?) CLDs
• Only 22 describing museum collections– only 8 exclusively museum collections
CIMI Forum, Edinburgh, 31 October 2002
66
CIMI and CD Focus survey
Simple DC (6),
Qualified DC (4),
RSLP CD (5),
ISADG (5),
EAD (3),
SPECTRUM (2),
Project-specific (9),
Other (9) : subset of ISADG (1), legacy data (1), national standard (1), MARC (2), may develop (1), VRA 3.0 (1), AMICO (1) , project-specific (1)
Simple DC (2),
Qualified DC (3),
RSLP CD (1),
SPECTRUM (2),
Project-specific (4),
Other (4) : national standard (1), VRA 3.0 (1), AMICO (1) , project-specific (1)
Schemas in use (in 22) Schemas in use (in 8)
CIMI Forum, Edinburgh, 31 October 2002
67
Issues, challenges….
• What is a collection?– “even-ness” across contexts, domains?– does it matter?
• Absence of cross-domain consensus on schemas for CLD
• Access points for CLD– what is the “subject” of a collection?– terminologies
– intra-domain/cross-domain
• Relationships between items and collections• Relationships between services and
collections
CIMI Forum, Edinburgh, 31 October 2002
68
Issues, challenges….
• Can single CLD support multiple functions?– Collection management and resource discovery?– Audience, purpose– Language: vernacular v technical/specialist
• Uncertainty of value of CLD, compared to item-level description?
– resource managers, resource users– perception of dumb-down?– varying perceptions within museum sector?
CIMI Forum, Edinburgh, 31 October 2002
69
Summary
• Access, integration, interoperability– across institutional, domain boundaries
• Collections as broadly similar objects• CLD as resource discovery metadata• CLD as contextual metadata• CLD supporting a “survey of the landscape”
– By human user, by software tool
• CLD as complement to item-level description• CLD as gateway to richer description• CLD as achievable mechanism for resource
disclosure
CIMI Forum, Edinburgh, 31 October 2002
70
More information….
http://www.cimi.org/http://www.cimi.org/
http://www.ukoln.ac.uk/http://www.ukoln.ac.uk/
http://www.nhm.ac.uk/http://www.nhm.ac.uk/
CIMI Forum, Edinburgh, 31 October 2002
71
Acknowledgements
UKOLN is funded by Resource: the Council for Museums, Archives and Libraries, the Joint Information Systems Committee (JISC) of the UK higher and further education funding councils, as well as by project funding from the JISC and the European Union. UKOLN also receives support from the University of Bath where it is based.
http://www.ukoln.ac.uk/