Environmental Linked Features Interoperability Experiment (ELFIE)

For more information please contact innovation@ogc.org

The first phase of ELFIE has been completed. An engineering report summarizing the project’s findings can be found here. A summary of the IE outcomes can be found here.

Summary of first ELFIE Activity Plan

In recent years, environmental domain feature models have been established by a number of sub-domain groups. However, there is no best practice or standard methodology to encode documents containing links between and among domain features, such as a rivers, aquifers, or soils, and observational data about those features. A common approach to encoding such links is required to allow cross-domain and cross-system sharing and interoperability of such linked information.

The Environmental Linked Feature Interoperability Experiment (ELFIE) will bring interested stakeholders together around the shared goal of providing a best practice for encoding documents expressing links between hydrologic and related features from models including but not limited to HY_Features, GWML2, and SoilIEML as well as observations and related content from a variety of observations and monitoring standards. The IE will produce an OGC engineering report summarizing the overall cross-domain inter-standard findings and recommendations for a best practice and/or standard to follow. This report will include encoded examples that should feed future linked-data encoding implementation standard development in one or more related standards working groups. To keep scope limited, this IE will not test service interfaces or processing capabilities that would necessarily exist to create or consume the data and documents being tested. The IE is informed by a number of existing systems, examples included with use cases, that offer or could offer services that would benefit from a linked feature data-encoding best practice.

Objectives

The objectives of the ELFIE include:

Demonstrate the use of existing and pending OGC standards for the encoding of environmental observation data in an integrated dataset of features linked according to ReSTful and Linked Data principles.
Prepare an OGC engineering report summarizing the group’s findings with the intention of future development of the relevant policies, best practices or implementation standards.
Provide draft linked data encodings to be considered by relevant standards working groups.

Background

The ability to encode documents containing links between and among monitoring sites and environmental domain features in a common way will enable automation and lines of inquiry that are not possible without manual intervention today. Currently, people hand-curate links between these sorts of features in ad-hoc organization-specific ways that preclude use by others.

Recent development of feature (see section 2.5 of the OGC Reference Model) and observation models specific to a domain (field of study, see here for more) such as HY_Features, WaterML2, GroundwaterML2, and SoilIEML has created opportunities for unprecedented software reuse and cross-system information sharing within their respective domains. However, a reusable approach to encode documents that use these information models in cross-disciplinary applications, is not apparent given existing encodings, implementation standards, and best practices. This interoperability experiment seeks to establish and test a common approach for encoding documents that use these existing data models.

The linked open data paradigm presents numerous challenges that the community is working to solve. This IE seeks to identify and address issues related to how to encode linked-data documents using OGC feature and observation data models for entity and link types. The IE will not address issues related to the internet architecture required to support the use cases; although the engineering report will describe best practices adopted and issues encountered (ex: multiple representations/encodings of the same real world object and content negotiation). However, run-time handling of URL redirection, URL structure, and any temporal variation of link nature of existence will not be addressed directly in the IE.

Experiments

The ELFIE will test one hypothesis for each of the use cases described above: that existing practices and semi-automated linked data encodings can be used to encode documents that would satisfy the use cases above.

Technical Approach

The approach taken in ELFIE will be limited in contrast to recent IEs concerning environmental information models. ELFIE will focus on 1) linked data documents that contain collections of linked features and related observations and 2) feature or observation representation documents that contain links to related features or observations. With regard to the OGC Reference Model, the IE seeks to understand how features (geospatial information) can be linked to each other in fully dereferenced documents or by reference through use of HTTP URIs. ELFIE will not seek to solve problems regarding network architecture for resolving links or systems design and governance for applications that store and retrieving links or concept relationships.

Methodology

For each use case, experimental documents will be crafted to test the hypothesis. Successes, difficulties, and missing capabilities will be reported on. These results will be used to inform future work on standards and/or best practices.

Define specific technical criteria to test hypothesis
Refine use cases and select source data
Select and/or draft linked data encodings to be used
Encode documents for each use case
Evaluate documents against predefined criteria
Write engineering report

Demonstration

The encoded documents will be demonstrated through creation of rendered web pages containing the features linked by the documents. A potential implementation of this is through Markdown with embedded plots and/or maps generated by a script that is based on the linked data document. This demonstration will show how such encodings could be used to assemble such integrated resources automatically.