Enhanced Publications - Semantic Web in Libraries

advertisement
Enhanced Publications, Linked Data –
Experiences From ECO4R
This talk
• About „Enhanced Publications“
– Links between Publications and underlying
data significant for research conduct
• About „Compound Objects“
– Focused, activity for Germany
– DFG project ECO4R led by the hbz Cologne
– Using semantic web as pragmatic tool
2
Enhanced Publications
3
links between publications and data exist
4
however, no systematic or machine re-usable links…
5
same article, even more enriched with links
6
Leading to ontology-based data
7
…. with even more data attached
8
another example, linking data to publication
9
but no reference to data in publication
10
and in the repository world? Also supplements.
11
and partly complex structures – not represented appropriately…
12
a repository-based publishing system with supplements…
13
and complex internal structures
14
Enhanced Publications
• Precious, existing patterns of scholarly
resources to be ‚liberated‘ for re-use
• Currently no way to systematically
represent these patterns for re-use
• Expressing „expectations“ of humans
– Concepts!!!
15
Compound Objects
16
Semantic Web / LOD a possible way to
represent enhanced publications?
• Yes
– machine readable web-representation of
compound objects
• No
– data missing
– not designed for human concepts
– service providers are missing
17
Eco4R – an exploration activity
• How to support existing concepts in human reasoning?
• How to make opaque or unstructured content in existing
repositories (rather than catalogues) visible and re-usable?
• focus Germany (OpenAIRE and Netherlands not shown here)
• Which existing vocabularies and ontologies can be re-used?
• What could be linked to other existing LOD data sources?
See also: „Report on Enhancing Interoperability
between existing Open Access Publication Infrastructures“
18
Eco4R Approach
Services
Machines
Enhanced
Publications
Data
19
Humans
Pragmatic Data Model Approach in
describing „Enhanced Publications“
• Based on OAI-ORE
• To semantically describe the aggregative publication entities
• Finding a compromise “Frbr Aligned Bibliographic Ontology”
• Extends the FRBR entities (Work, Expression, Manifestation,
Item) with more specific classes (e.g. JournalArticle,
WebPage, SupplementaryInformationFile)
• Complement to other ontologies in SPAR (Semantic
Publishing and Referencing), which can describe a whole
publishing workflow
20
Data Model Instance
21
Implementations done in ECO4R – I
•
Support of the data model by plugins in two fundamentally
different repository platforms:
•
OPUS – most used platform in Germany, simple architecture
with numerous limitations in file-handling
•
•
Fedora – rather complex system, fine granularity in access
modes, file management, versioning, audits
•
22
The 1st ORE support for this platform
Refactoring of existing ORE implementations, compliant with the
latest specification
Implementations done in ECO4R – II
•
•
Plugins neither need modifications of the internal repository
data model nor the repository user-interface
•
A pragmatic way for making repositories linked data
compliant
•
Authors of publications are not bothered
Overlay Journal – prototype as a prove of concept
•
23
Research project: operation not part of the plan
Overlay Journal as proof of concept
1. OAI-PMH Harvesting of ORE
ResourceMaps
2. Persisting in a Triple-Store
3. Processing for the Overlay
Journal (Broker)
•
Storing results in MySQL
4. Visualization for end-users
5. SPARQL interface for service
providers
24
Work in progress
25
Re-Use Aspect demonstrated
• Placeholder for User-Interface Screenshot
Work in progress
26
Discussion
27
Benefits for Linked Data compliant
Aggregators
• Descriptions of (Enhanced) Publications are compliant with
web standards
• Linking and enrichment by using terminology services, for
e.g.
•
DDC, PND, Organizations (lobid.org), Projects (rkb-explorer)
• Reliable links to e.g. full-text and datasets, when publication
entities are semantically described
• Improved retrieval interfaces by provision of a SPARQL
endpoint
28
KO criteria for enforcing
Linked Data Services
• Interoperability: after the experiments and prototypes – which
ontologies, vocabularies will survive ?
• What about the maintenance of all the ontologies, vocabularies ?
• Availability, Reliability and Quality of terminology services,
SPARQL-Endpoints etc.
• A fundamental requirement regarding the complex linking
character of the Semantic Web
29
Enhanced Publications as LOD – a possible
way with ‚concepts‘ are starting point
• Yes
– machine readable web-representation of
relations << ORE / FABIO
• No
– data missing << Repository PlugIns
– not designed for human concepts << Enhanced pub.s
as starting point
– service providers are missing << Demonstrator
30
Plans
• Open Source Release of the OAI-ORE repository plugins
• Further developments for Dspace, Eprints?
• Using the CARPET platform & DINI/OA-Netzwerk as hosts to
further discuss and circulate the “Enhanced Publication”
paradigm in Germany
• Conceptualization (SKOS?)
• linking to Enhanced Publication Developments in the NL
• embedding in in EU (OpenAIREplus, euroCRIS)
• Building aggregators (BASE?)
31
Thanks For Your Attention
• The Project Team Members
• Anouar Boulal, Jochen Schirrwagen, Martin Iordanidis,
Andres Quast, Jan Schnasse, Friedrich Summann
• Web-site: http://www.eco4r.org
• Wiki: https://trac.eco4r.org
32
Download