DennisNicholsonGordo nDunsire

Slide1: 

Applied Collection Description in the SCONE Project Dennis Nicholson / Gordon Dunsire From Functional Granularity To People Interoperability

Slide2: 

Gordon Dunsire Collection Level Metadata and Functional Granularity

Overview (Part 1): 

Overview (Part 1) Collection Level Metadata Functional Granularity Cross-searching

SCONE: 

SCONE Scottish Collections Network Operational, but embryonic, service arising from the SCONE project (RSLP) Collections located in Scotland, and about Scottish topics 3500 collection level descriptions Range from official publications of Albania to the National Library

Slide5: 

Description Collection of official publications of Albania, as defined by the Scottish Working Group on Official Publications (SWOP). Notes Statistics. Type Collection.Library.Text.User … Location Strathclyde University. The Andersonian Library Curran Building 101 St James Road Glasgow (City of Glasgow, Scotland) G4 0NS Access Contact before a visit is not required. Item creators Albania Subjects Albania Part of Strathclyde University. Library SWOP collection Catalogues Strathclyde University WebPAC

Slide6: 

Description A small collection of 25 books from the library of John Buchan (1875-1940) … Size 25 v. + mss. … Administrative Date Transferred from To Terms history 1940 John Buchan … National Library … Bequest 1977 Tweedsmuir … National Library … Donation 1985 Tweedsmuir … National Library … Donation Location National Library of Scotland George IV Bridge … City centre location, Waverley rail station, bus … … Collectors John Buchan (Baron … [Collecting: closed] Item creators John Buchan (Baron Tweedsmuir) (1875-1940) Part of National Library of Scotland collection (main) Catalogues National Library of Scotland WebPAC

SCONE metadata: 

SCONE metadata Structure based on Heaney’s e-r model ‘Analytical model of collections and their catalogues’ RSLP guidelines 3+ entities Collections Agents (personal and corporate) Locations (physical and virtual) + Subjects now developed with HILT project Collections metadata not new, but not given focus until recently A, L, S familiar from item level metadata

Entity relationships: 

Entity relationships Location access and opening hours Agent-Administers-Location Accrual status and policy Agent-Collects-Collection History of custodianship Agent-Sells[Collection]To-Agent Agent-Delegates[Collection]To-Agent

Collections and sub-collections: 

Collections and sub-collections What is a collection? Heaney’s concept of ‘functional granularity’ “level of detail required [for] resources discovery or collection management” Defined by users, owners and managers Discovery Marketing Collaboration

SCONE findings: 

SCONE findings Multi-level granularity SWOP collections have 6 levels Most named special collections have a general (organisational) parent Mono-hierarchy with only one parent at each level Such a ‘backbone’ may form the basis of organisation-based collection services Regional or sectoral

Polyhierarchy: 

Polyhierarchy But distributed and ‘assembled’ collections require additional virtual parents Edward Clark Collection has Napier University Merchiston Learning Centre collection as physical parent And Bookhad project ‘collection’ as virtual parent And ‘Rare book collections in Scotland’ collection as virtual (potential) parent, etc.

Cross-searching: 

Cross-searching Collection-Collection relationships Parent/child Catalogue/Finding aid Other (splits) Standard collection name or other identifier? Pre-coordinate forms for other entities Name authority files (via extended warrant) Gazetteers Subjects!

Pre-coordination: 

Pre-coordination Local, service-specific pre-coordination needs to take account of the general All service-scope boundaries will be leaky Research collections held by public libraries Lifelong learning resources held by FE/HE Collections located in a region, and collections about a region (but located elsewhere) (SCONE)

Data management: 

Data management Functional granularity implies duplication of A, L, S entity metadata Volatility differs from item-level Persistent locations; dynamic administrators Range of agents wider Collectors, administrators as well as creators And collection entities likely to multiply Suggests need for coordination Scope and definitions Authoritative headings

Slide15: 

Dennis Nicholson (2) Collection Strength and People interoperability

Overview (part 2): 

Overview (part 2) Scotland and Collection strength: Conspectus, Collaborative Collecting, User Navigation Known problems and SCONE solutions: Long term/new systems: automated CS indices Meanwhile: SCAMP-mediated constrained professional judgement for legacy metadata Linked by ‘people interoperability’ Importance of human level processes: Lessons from CAIRNS and SCONE A CoSMiC example…

Collection Strength in Scotland: 

Collection Strength in Scotland A Scottish Conspectus, used for: Collaborative Collecting: SCURL, Conspectus, CCD policies User Navigation Research Collections Online CAIRNS distributed catalogue and dynamic landscaping SCONE to examine known problems and propose solutions (amongst other things)

Known Problems: 

Known Problems A selective list: Lack of objectivity and consistency, so navigational information could be better Labour intensive Subject scheme incompatibilities between: Service schemes/Conspectus; Services Snapshots - data not current No information on coverage overlap Misleading: small collections with unique items Data insufficient for deep resource sharing Granularity level less than user’s query How do user terminologies map to schemes?

SCONE Solutions (2): 

SCONE Solutions (2) Long term/new systems: automated CS indices Systems to include this item level metadata: DDC number to a reasonable granularity level An agreed subject term for the number (HILT) MEG educational level Unique object identifier Data on format (e.g. electronic only) Charging policies Service exclusion policy Language(s) of content Each will build CS index containing these elements for cross-searching

SCONE Solutions (2): 

SCONE Solutions (2) Solves a number of the known problems: Relatively objective Relatively consistent Able to indicate overlap (unique identifier) Shows small collections with unique items Up to the minute currency Single subject scheme at CS and item level across all services Support for deeper resource sharing CS description to low granularity levels used by users Marginal cost if built into system Leaves user terminologies mapping to HILT

SCONE Solutions (1): 

SCONE Solutions (1) Automated CS indices will work for new systems and in the long term legacy metadata systems Meanwhile: SCAMP-mediated constrained professional judgement for legacy metadata systems (Objectivity via CCD, user needs, agreed methodologies, peer review) Not as good as automated but an improvement on current situation: Should be more objective Should be more consistent Less labour intensive (but still poorer results than automated approach with more effort)

SCONE Solutions (1): 

SCONE Solutions (1) Can show small collections with unique items if special provision agreed Unable to indicate overlap Better currency but still a snapshot Subjects inter-compatibility problem remains Data insufficient for deep resource sharing CS description granularity levels still poor Doesn’t solve user terminologies to schemes problem (HILT could solve but at higher cost than automated approach) Added advantage – a human level process place leading to in time to fully automated approach People interoperability…

Why Human Processes Matter: 

Why Human Processes Matter Lessons from CAIRNS and SCONE: Users increasingly use/need distributed resources, finding tools so co-operation now essential as well as desirable: Distributed networked collections need collaborative management Coherent distributed virtual ‘libraries’ won’t just happen – we must co-operate to manage retrieval/user environments People interoperability a pre-requisite of technical and metadata interoperability Design people into the system Small is beautiful where people are the key

A CoSMiC example…: 

A CoSMiC example… Building people into the (SCONE) system: CoSMiC, SCAMP, a Co-operative Infrastructure A quality service needs good inter-compatible metadata maintained by collection managers Harvesting whatever turns up won’t do the job To deal with ‘people factor’, need processes to : Actively monitor problems and promote progress Agree common terminological and other metadata standards, apply them in a standard way, input them in the same form; ensure adherence through central input with local output; currency control… Make any necessary standards changes jointly Train to think globally before acting locally

Thank you!: 

Thank you! http://scone.strath.ac.uk/ http://scone.strath.ac.uk/service/index.cfm http://scone.strath.ac.uk/scamp/index.html http://cosmic.cdlr.strath.ac.uk/ http://hilt.cdlr.strath.ac.uk/ d.m.nicholson@strath.ac.uk g.dunsire@napier.ac.uk