Presentation given by Krysztof Janowicz and Pascal Hitzler in the afternoon Architecture Forum Session on Day 1, June 24, at the EarthCube All-Hands Meeting.
AHM 2014: OceanLink, Smart Data versus Smart Applications
1. Why
OceanLink: Smart Data Versus Smart
Applications
Krzysztof Janowicz
STKO Lab University of California, Santa Barbara, USA
Pascal Hitzler
DaSe Lab Wright State University, Datyon, USA
EarthCube All-Hands Meeting
EarthCube Architecture Forum
June 2014
Smart Data Versus Smart Applications Janowicz and Hitzler
2. Why
What kind of architecture specification do you have?
OceanLink relies on the Semantic Web and Linkd Data; strictly
speaking this is not an architecture.
Data component: Data is translated into RDF (Resource Description
Framework), semantically lifted, and published as 5-star Linked Data.
Schema component: OceanLink wants to foster discoverability and
interoperability without restricting heterogeneity and thus does not
use classical static data models. Instead, it relies on Ontology Design
Patterns (ODP) and data-driven, application-centric ontologies that
use these patterns.
Service component: OceanLink data is made discoverable and
queryable via a SPARQL Endpoint. Inferencing is supported via the
used ODPs and standard Semantic Web reasoning services. A user
interface for data seeking and exploration is provided via a faceted
browsing interface.
Smart Data Versus Smart Applications Janowicz and Hitzler
3. Why
How is it being used?
In OceanLink Semantic Web technologies & ontologies are used to
Ease the publication of data (so far BCO-DMO and R2R)
Improve the retrieval of data beyond keyword search
Deploy ODP to be used by other EarthCube repositories
Establish links between repositories (planned)
Compress data based on background ontologies (planned)
Support simple inferencing based on the ODP
Smart Data Versus Smart Applications Janowicz and Hitzler
4. Why
Why is it valuable?
Two key insights and paradigm shifts
1 Enable the creation of smart data in contrast to smart applications.
Instead of developing increasingly complex software, the business
logic should be moved to the (meta)data. Smart data will make all
future applications more usable, flexible, and robust, while smarter
applications fail to improve data along the same dimensions.
2 Cultural, conceptual, and infrastructural heterogeneities must be
respected in order to maintain different perspectives and differing
priorities and thus foster inclusivity in the EarthCube endeavor.
Heterogeneity is a resource and should not be resolved.
The Semantic Web and Linked Data were developed with those ideas and
Web-scalability in mind. Longevity is ensured via an early, open, and rigid
standardization process by the W3C.
Smart Data Versus Smart Applications Janowicz and Hitzler
5. Why
What worked, and what did not?
Worked
Creation of ontology design patterns
Creation of Linked Data
Faceted browsing interface
Did not work
Too early to tell
Some data types/formats will be difficult to transfer to Linked Data
Smart Data Versus Smart Applications Janowicz and Hitzler