Enabling Linked Open Data to accomplish the envisaged Semantic Web

Most of the data are in formats other than RDF.
Not all media can be available as Linked Open Data.
Nor ever all media will be represented as Linked Open Data.

RML provides a solution to model domain-level knowledge
in a scalable, integrable and interoperable fashion
by semantically representing using the RDF framework,
data derived from multiple heterogeneous sources.

Factors for enhanced mappings and primary interlinking

Uniform mapping definitions
Mapping definition independent of the references to the input data.

Reusable mapping definitions
Same mapping definitions to be (re)used across different sources

Interoperable mapping definitions
Mapping definitions to be reused across different implementations for different source formats.
Reduced implementation and learning costs.

Robust cross-references
Uniquely define the pattern that generates a resource.
Refer to its definition every other time this resource is mapped (enrichement).
Modifications to the patterns or data values are propagated, making the interlinking more robust.

Primary Interlinking on mapping level
Cross-references among data in different input sources to be defined already on mapping level.

Scalable and extendable mapping definitions
Reference to the data extracts and the mapping definitions in a distinct fashion.
Define in a generic way what can be for all possible different input sources and scales over what cannot.

Performing mappings with RML

Limitations of current mapping methodologies

Mapping on a per-source basis
A single source is mapped every time and only relationships between data within this single source appear.

Mapping on a per-format basis
Only support mappings from a certain source format.
Require to install, learn, use and maintain different tools for each case separately.

Independent mapping
Mappings are defined independently, disregarding its possible prior definitions and links to other resources.

Manual alignment
Manual alignement when same resources appear in different input sources.
The resources URI patterns definition are redefined and/or replicated every time.
Duplicates are generated and appear even within a publisher's own dataset.

Mapping and Interlinking as consecutive steps.
Relationships between resources are defined after the data is mapped and published.

Limited reuse of mapping definitions
There is no standard mapping formalization for other source formats apart from relational databases.

Non-interoperable mapping definitions
Mapping definitions are tied to the implementation.
Mapping definitions extraction, process and reuse across different implementations is prevented.

current mapping methods