Skip to content

Original COR Proposal

John Graybeal edited this page Jul 10, 2015 · 1 revision

This was the original proposal for the ESIP Community Ontology Repository.

A Community Ontology Repository Testbed: Dev

Based on MMI’s Ontology Registry and Repository
John Graybeal, Marine Metadata Interoperability (MMI) Project
12 May 2015

a. Project Description

We propose to develop and deploy an ESIP-dedicated and ESIP-branded “Community Ontology Repository” (COR) pilot instance, to be used in a semantic technologies evaluation study.

In past engagements with the Semantic Web Cluster, managers of the MMI Ontology Registry and Repository (ORR) have noted differences between the existing ESIP deployment of a BioPortal-based ontology repository and MMI’s BioPortal-based ontology repository. More recently, discussions in the EarthCube community mentioned ESIP as a possible semantic registry and repository host, and the Open Geospatial Consortium has called a Geospatial Ontology Summit to focus on geospatial semantics issues. Given steadily increasing adoption of MMI’s ontology repository as well as the broader repository interest outside of ESIP, we believe it is timely to formally evaluate the ontology repository needs of the community.

We propose to make a dedicated instance of the MMI repository available to the ESIP community, and to identify and document the process of deploying the system in the ESIP context. Working with the Semantic Cluster during the FY2015 period, we will initiate the processes required to document the needs of the wider earth science community. We propose this deployment in order that ESIP can achieve the following goals:

  • Give non-experts the opportunity to experiment with semantic resources and explore semantic concepts using their own vocabularies and semantic content;
  • Give the ESIP user communities the opportunity to consider what features and functionality they find valuable in such repositories;
  • Give the ESIP semantic and technical communities the opportunity to evaluate the two repositories on a side-by-side basis, using criteria to be developed; and
  • Give ESIP the opportunity to consider its possible roles, opportunities, and costs in providing semantic services to the larger earth science information community.

We propose to complete the installation, configuration, and initial checkout of ESIP’s MMI repository instance by September 30, 2015. This will initiate the opportunities described above. However, because an extended evaluation period will offer the most advantage to ESIP and the broader earth science community, we will propose for FY2016 to leave this instance operational, continuing to support it for the purposes outlined above, until an extended evaluation is complete.

b. Project Work Plan, Timeline, and Budget

Work Plan

Upon approval of this proposal, we would review the ORR instance deployment procedures to identify any concerns, finalize an agreement on target URI paths with ESIP, and work with ESIP to acquire an appropriate AWS instance. We would then configure the instance and begin the installation of ORR, which is a well-defined set of core and custom services.

We will instantiate the ESIP-relevant code, forking it from the MMI ORR GitHub repository. Local customizations will be made as needed to the ESIP fork. Modifications may be offered by anyone to either the MMI or ESIP fork; during the evaluation period, MMI will manage modifications to both. At the evaluation period close, the ESIP fork will be moved to the ESIP GitHub repository, and full control will be given to ESIP, unless an alternate agreement is reached. We emphasize our interest in collaboratively enhancing this software and its related utilities, whether for ESIP-specific needs, or broader community value; and the option for ESIP to fully maintain and enhance the software after the demonstration period.

While the software is being tested and finalized, we will coordinate with ESIP system managers to create the appropriate domain forwarding, so that ESIP URLs will point to the new instance (included in the testing). We will set up automated operations monitoring of the system. During this period, we will work with ESIP to begin defining evaluation goals and procedures for a side-by-side comparison with the existing ESIP ontology portal.

When the system is ready, we’ll notify ESIP’s management and Semantic Web Cluster, and coordinate with them on wider announcements. The system will also be advertised to the MMI community, drawing further attention to ESIP’s semantic activities and capabilities.

Timeline

(assumed) June 1: Project approval
June 30: AWS instance acquired, software migrated to it
ESIP Summer Meeting: Progress Report given at Semantic Web session
July 31: AWS instance and software configured, initial operations achieved
August 31: ESIP domain forwarding setup and working, initial testing complete
September 15: Notification of COR readiness to ESIP management and SW Cluster
September 30: General announcement of COR availability; completion report issued

Budget

We request $4,750, allocated as follows: AWS instance costs: $2000 ($500/month * 4 months) Travel/registration for ESIP workshops and related semantic events: $1,750 John Graybeal will attend ESIP Summer Meeting to present status update John Graybeal will attend OGC Geospatial Ontology Summit This Summit overlaps with EarthCube All Hands Meeting; Semantic Web Cluster Chair will be attending All Hands Meeting with John attending the Summit to maximize the exposure of this project Software/hardware acquisition as needed for development support: $1,000

We will manage budget variances (for example, in AWS usage/costs) as appropriate.

c. COR Team Membership

John Graybeal, MMI Project Lead: John has overseen the MMI ORR since its creation 7 years ago, and curated much of its current content. He will assume primary responsibility for the installation, configuration, test, and operation of the COR ORR instance on AWS.

Carlos Rueda, MMI Technical Lead: Carlos wrote most of ORR’s custom software and its associated utilities, and performs system development and operational maintenance. He will provide technical advice and support for the AWS installation, and consult regarding new technical requirements and questions that may arise during the evaluation period.

Felimon Gayanilo, ORR Operations Lead: Felimon leads operational IT support for the existing ORR instance, and will consult on the setup and configuration of this instance.

Tom Narock, ESIP Semantic Liaison: Tom is the ESIP Semantic Web Cluster Chair, and will facilitate communications and collaboration between this project and ESIP’s semantic work.

d. Anticipated Outcomes

We anticipate the following products (see next section for other outcomes):

  • A working ORR instance (COR) available at the ESIP domain;
  • All original and customized source code made available on ESIP’s GitHub site; and
  • An installation report, describing
  • Resources used (time, cost),
  • Issues found and resolved (using ORR’s GitHub issues page), and
  • Lessons Learned.

In FY2016 we anticipate additional products resulting from the same installation, should a continuation proposal be funded.

e. Impact of Project

We expect this project to provide perspective about the features, cost, availability, and usability of semantic services, as well as comparative information about these two repository solutions. This development proposal will particularly yield information about the challenges of deploying the technology on the cloud, and of providing (or changing) simple ESIP branding.

Increased visibility and attention to the systems (due to increased activity and social outreach) will increase ESIP’s semantic credibility, and better prepare ESIP to consider its long-term opportunities and options. These benefits and others will multiply should a continuation proposal be funded.

Finally, as a result of direct and indirect engagements, and the unique alignment of this activity with EarthCube, OGC, and other initiatives, the entire earth science semantic community will engage strongly with semantic technologies, and may start to coalesce around the need for a respected international provider (like ESIP) to provide these services on a long-term, cost-reimbursed basis. Consistent with ESIP’s vision and mission, we anticipate the COR software based on ORR may form the core of a sharing platform, where developers can discuss and contribute software, and form a core community to sustain operations and development of the ESIP COR deployment.