Linked Open Data
and
Scientific knowledge
management systems
Gregor Hagedorn, Museum für Naturkunde
Berlin © U.Kils, CC BY-SA 3.0; from Wikimedia Commons
1. Scientist reads in the Web browser
2. Data- federation /-indexing
Links are not yet operational
GBIF Datenfederation
Data from Thailand?
Fungal Cultures
CBS
Herbarium München
Insect Collection
MfN
Bird data Germany
after Dave Remsen, GBIF, TDWG 2011
Bird data Austria
Insect Collection NHM, UK
© i4Life
Some linking (R. Page)
Linking is usually DB-internal = local
This is inefficient.
- Stable Links - Return data
Identifiers
Specimen Collection Datenbase
Botanical Nomenclatural
Database
Classical Identifiers
Specimen Collection Datenbase
Taxon = urn:uuid: 6e8bc430-9c3a-
11d9-9669-0800200c9a66
Botanical Nomenclatural
Database
Taxon = urn:uuid: 6e8bc430-9c3a-
11d9-9669-0800200c9a66
Classical Identifiers
Specimen Collection Datenbase
Taxon = urn:uuid: 6e8bc430-9c3a-
11d9-9669-0800200c9a66
Botanical Nomenclatural
Database
Taxon = urn:uuid: 6e8bc430-9c3a-
11d9-9669-0800200c9a66
Classical Identifiers If this
found:
And this
found:
Then relation
detected:
Specimen Collection Datenbase
Taxon = http://
id.pesi.org/tax/6e8bc430-9c3a-
11d9-9669-0800200c9a66
Botanical Nomenclatural
Database
@ http:// id.pesi.org/tax/6
e8bc430-9c3a-11d9-9669-
0800200c9a66
Semantic Web
Specimen Collection Datenbase
Taxon = http://
id.pesi.org/tax/6e8bc430-9c3a-
11d9-9669-0800200c9a66
Botanical Nomenclatural
Database
@ http:// id.pesi.org/tax/6
e8bc430-9c3a-11d9-9669-
0800200c9a66
Semantic Web If this
found: Then
relation derefenced
Specimen Collection Datenbase
Taxon = http://
id.pesi.org/tax/6e8bc430-9c3a-
11d9-9669-0800200c9a66
Botanical Nomenclatural
Database
@ http:// id.pesi.org/tax/6
e8bc430-9c3a-11d9-9669-
0800200c9a66
Semantic Web
Micro-citation of data!
Semantic Web
B C
Thing
A D E
= typed links (properties)
Thing
Thing
Thing
Thing
Thing Thing
Thing
Thing
Thing
Linked Open Data Cloud (LOD 2011)
Linked Open Data Cloud (LOD 2011)
Why Linked Open Data?
– Distributed Web Model • using W3C standards (xml, rdf, owl)
• Machine usable data (automatic analysis & reasoning)
• Web pages at same identifier (content negotiation)
Why Linked Open Data?
– Distributed Web Model • using W3C standards (xml, rdf, owl)
• Machine usable data (automatic analysis & reasoning)
• Web pages at same identifier (content negotiation)
– Anyone can say anything about anything, anywhere • Usages that the data providers never anticipated • Third parties connect concepts between data sets • Particular needs contribute to global achievement
Why Linked Open Data?
– Distributed Web Model • using W3C standards (xml, rdf, owl)
• Machine usable data (automatic analysis & reasoning)
• Web pages at same identifier (content negotiation)
– Anyone can say anything about anything, anywhere • Usages that the data providers never anticipated • Third parties connect concepts between data sets • Particular needs contribute to global achievement
– Flexible to adapt to almost any form of data – Information managed at source plus annotated globally
Why Linked Open Data?
– Distributed Web Model • using W3C standards (xml, rdf, owl)
• Machine usable data (automatic analysis & reasoning)
• Web pages at same identifier (content negotiation)
– Anyone can say anything about anything, anywhere • Usages that the data providers never anticipated • Third parties connect concepts between data sets • Particular needs contribute to global achievement
– Flexible to adapt to almost any form of data – Information managed at source plus annotated globally – Queries and other analysis can combine arbitrary sets of
data, anywhere and owned by anyone – Common and diverse vocabularies can be used together
and related to each other (creativity, science!)
Strategy: 1. Stable Identifiers Now (Semantic Web compatible, http-dereferenceable)
2. Semantic Web Later ...
Identifier patterns?
–http://objects. myorg.edu/id/1C4EDC178 AD79DD7F1A5AB856E8C5BCA#treatment
or –http://concepts.myorg.edu/id/123#treatment or –http://id.plazi.org/specimen/123#label
Stability is a management
decision!
Scientific knowledge
management systems?
Discoverable Distributed Dissent & Discourse
Fine-granular knowledge curation?
LOD is not solution to all.
Infrastructure to build on top of it
Collaboration must become more efficient