1 of 24

1
Digitometric Services forOpen Archives
Environments

• Tim Brody
• Simon Kampa, Stevan Harnad,
• Les Carr, Steve Hitchcock
• tdb01r,srk,harnad,lac,sh94r_at_ecs.soton.ac.uk

University of Southampton, Intelligence, Agents,
Multimedia Group
2
Open Archives Initiative
3
OAI Data ModelResources/Items/Records
resource
All available (meta)data about the resource
Item OAI identifier
item
Dublin CoreMetadata
MARC Metadata
???XML
records
record metadata identifier datestamp
4
Protocol Responses
5
Protocol
HTTP URL Requests
Service Provider
Data Provider
XML Responses
Identify
1
Collection-level Description
ListRecords?metadataPrefixxyz
2
All repository xyz records
ListRecords?from2003-04-02
3
All repository xyz records since 2003-04-02
6
Other Commands

• ListIdentifiers
• Return only the identifier/datestamp/set
  membership
• ListMetadataFormats
• Return the available data formats
• ListSets
• Return the set structure (if there is one)
• GetRecord
• Return a record given by OAI identifier

7
Interest in OAI

• 111 registered OAI repositories
• Many unregistered (e.g. all GNU EPrints.org and
  DSpace archives)
• 4,500,000 public records
• http//arc.cs.odu.edu/
• NSDL project, UKs JISC Information Environment
• OLAC (language community built on OAI)

8
Why OAI?

1. Mandated Dublin Core allows the quick
   establishment of basic services and tools
2. Simple and metadata-neutral protocol allows more
   interesting possibilities (without breaking 1.)
   and extensions

9
Adding Caching to OAI-PMH
10
Celestial (OAI Cache)

• Developed to maintain a local metadata copy
• Avoid repeated, large harvests during development
• Provides an abstraction over multiple OAI
  versions
• (hence acts as a gateway to older
  implementations)
• Useful for testing OAI implementations
  improving performance
• Using XSLT provides a Web interface to OAI
• Provides redundancy

11
(No Transcript)
12
Citebase Search Data Model
e-Services
13
Content

• 250,000 full-text resources
• 240,000 of which arXiv.org
• 6 million references
• 29 mean refs/paper (therefore failed to extract
  references for 18 of papers)
• (n.b. modal refs is 19)
• 1 million references linked internally to the
  full-text (15)

14
(No Transcript)
15
Citebase Search
16
Citebase SearchNavigation by Citation Links
Article withreference list
Future
Referencelink
Related
Current Article
Co-cited
Past
17
Citebase Search
cites
cites
18
Citebase Search
cites
cites
19
Citebase Search
Co-cited
20
Read/Cite Cycle
21
Digitometric Services for OAI

• Tools for visualising research metadata
• Builds an analysis service on Citebase
• Knowledge mapping (co-authors, co-citation, etc.)

22
Co-Citation Network
23
Full Co-Citation Map
24
Digitometric Services forOpen Archives
Environments

• http//www.openarchives.org/
• http//opcit.eprints.org/
• http//citebase.eprints.org/
• http//www.eprints.org/
• http//www.hyphen.info/
• AKT Project (knowledge)

Thank you for listening! Tim Brody