Ariadne: Interoperability
- 1. ARIADNE is funded by the European Commission's Seventh Framework Programme
Interoperability
Holly Wright
- 2. Overview
• Why do we care about interoperability? Isn’t it
enough to just make our data available for re-use?
• Once you go through all the hard work of
preserving and disseminating your data, a whole
world opens up!
• By combining your data with the data of others in
new ways, we can create new knowledge and
understanding that was not possible before.
- 4. Metadata
• Metadata sits at the heart of interoperability. Without
good metadata, interoperability isn’t possible.
• Metadata should, whenever possible be based on
standards
• It often takes some time and research to determine
what the appropriate standards are for your metadata
• Very often the metadata created for a project, may not
conform to a standard, and mapping has to take place.
This isn’t a bad thing, but if metadata is already
standards compliant (all or in part), its much less work!
- 5. Mapping
• Use of standards-based ontologies, thesauri and
controlled vocabularies can help your data become
interoperable.
• The most widely used ontology in Cultural Heritage is
the CIDOC CRM, which is an ISO standard.
• Not specific to any one domain, so no terminology or
relationships are presented that are specific to
archaeology.
• Extensions have been developed like the CRM-EH and
CRMarcheo
- 6. Mapping
• Working with ontologies can require a lot of work and
expert knowledge, which may or may not be necessary
• Using thesauri; lists of agreed upon terms with simple,
hierarchical relationships is often all that is needed
• Even using controlled vocabularies, where you just map
to a list of terms can be an easy path to interoperability
• Example is the SENESCHAL project, which brought
together archaeology vocabularies and thesauri used by
the national agencies for England, Scotland and Wales,
which can now be used as standards
- 7. Mapping
• ARIADNE is using the CIDOC CRM and experimenting
with the new CRMarchaeo, but for archaeological
subjects has chosen to map to the Getty Art &
Architecture Thesaurus (AAT) as a central spine
• Mappings are made in the partner’s native language
(and English if desired)
• This means you can search for a subject in Hungarian,
and get results in German (English is just the glue)
- 8. Geo data
• Lots of work has been done with making
archaeological data interoperable with regard to
place
• One of the best examples is the Pelagios project
– Links online resources to the historic past, primarily
within the classical world, meant to be primarily machine
readable
– New initiative called Peripleo, which provides a map
interface
- 9. Temporal Data
• Most of the work has been done on what and
where, as its (relatively) easy
• By far the most difficult aspect of making
archaeological data interoperable is dealing with
WHEN
• When is always dependent on where (bronze age is
different depending on where you are in the world
• CRMarchaeo has tried to deal with this using a
concept called ‘space-time’ volumes
• PeriodO is using ‘assertions’ to build consensus
around temporal terms
- 10. Different Approaches
• Some approaches for making data interoperable use
a top-down approach
– Using a controlled vocabulary to which everyone agrees
to map their data
– Mapping to an ontology like the CIDOC CRM
• Some approaches for making data interoperable use
a bottom-up approach
– Using a variety of sources showing where an
archaeological place is mentioned in a text to create a
research resource
– Using the assertions used by different people in different
place to build up assertions about archaeological time
periods
- 11. Linked Open Data
• Most of what underlies all of the examples shown is
based on technologies and concepts that use
Linked Data, preferable Linked Open Data (LOD)
• Linked Data is a very different way of organising
data
• Rather than using relational tables found in most
traditional databases, it uses a graph data
structure, which has no hierarchy.
- 12. Linked Open Data
• Everything is built using a subject-predicate-
object relationship that can linked in any
direction, pulled apart and recombined in
any direction.
• It allows the use of inference to leap across
concepts.
• We even use it as important part of our CMS.
- 13. Portals
• Interoperable data can be brought together
and searched from a single interface, often a
portal. This is known as a federated query.
• One of the earliest examples of portal was
part of the ARENA2 project.
• Most of the previous examples are portals,
using data from a variety of sources which
has been made interoperable.
- 14. Acknowledgements
ARIADNE is a project funded by the European Commission under the Community’s
Seventh Framework Programme, contract no. FP7-INFRASTRUCTURES-2012-1-
313193.