Complementary to our work on GRID data inventory and query systems, Terradue offers a specialist service for the management and distribution of very large spatial data sets carried out collaboratively with best-known open source projects,

With duetopia we work to exploit and strengthen best practice in distributed data processing, archiving and discovery. Our emphasis is on the delivery of robust operational systems while keeping a concrete roadmap connecting what is built now to next generation data processing and storage systems.

This platform was designed to easy:

• The publication of metadata into well-known registries and indexes
• The syndication of data between peers automatically
• The discovery of data relevant to an application
• The establishment of clear rights to reuse data, minimising negotiation

As such, duetopia is a description of a set of services for indexing, searching, sharing and storing very large spatial data sets. The parts taken together constitute the core of what is needed to participate in a "Spatial Data Infrastructure"

Background

As a follow-up to her initial articles on Minimal metadata abstract model, Planning duetopia and the More inputs for duetopia, Jo Walsh wrote three essays that lead the ground for the duetopia development.

The first essay, called On metadata about geographic information, provides an overview of concerns around providing metadata about geographic information to users of a spatial data distribution service. It identifies best practices and was intended to provide supporting context for the use of a minimal abstract model and to offer an overview of current and future concerns in geographic metadata production and use.

The second essay, called On the Development of Spatial Ontologies, provides a critical overview of the concepts of “geosemantics”, data abstraction and packaging in the light of recent developments in collaborative information modeling over the Internet. What does it mean to have “semantic interoperability”, and how do we get at it?

The third and last essay of this initial collection, is titled On Spatial Data Search and takes a closer look at the apparent lack of cost effectiveness of the Spatial Data Infrastructures effort in developing "catalog services" and community specification for spatial search. The GIS industry coins specifications for search interfaces, where it might do better to reach out to other information retrieval and Internet communities. A bigger picture of data search surrounds the spatial problem; and conventional text search engines move to assume the same functions.



Also of interest for duetopia are the following Jo's entries in our blog:

• Impact of the INSPIRE Discovery Services Implementing Rules - 2008-11-21
  
  
• Stepwise Geospatial OpenSearch - 2008-06-12
  
  
• OpenSearch - A Design for Compliance - 2008-02-06
  
  
• The Joy of Internationalisation - 2008-01-15
  
  
• INSPIRE - A Design for Compliance - 2007-12-03
  
  
• Updates from the spatial metadata world - 2007-10-08
  
  
• More inputs for duetopia ... - 2007-08-23
  
  
• Reflections on the 13th EC-GIS workshop in Porto - 2007-07-16
  
  
• Planning duetopia - 2007-06-27
  
  
• Minimal metadata abstract model - 2007-03-29
  

Source Code

The first public version of duetopia is released as open source and hosted in google code pages. This version presents the initial implementation of the libraries to provide import and export from common standard formats into a shared model (ISO 19115, ISO 19139, FGDC, RDF/XML).

A GeoDjango-based graphical user interface is also included to allow collaborative editing of metadata records in the repository, and a search interface that is designed to comply with the INSPIRE Implementing Rules for Discovery Services in the EU in a minimal way.

The current implementation of the machine-machine interface is using OpenSearch (with Geo extensions) to provide provides an Atom feed of geographic metadata in RDF/XML.

The goal of this first version is to lay the ground to integrating systems rather than requiring certain standards. A common subset of well-known standards for geographic metadata is identified, building on common work in the open source geospatial community. This subset corresponds closely to the draft abstract model described by the Implementing Rules for Metadata in the INSPIRE European SDI. However, more emphasis is placed on the machine-readable and repeatable properties of data: unique identifiers which can be used to annotate the provenance and processing history of data sets; URLs to represent license constraints; contact details for people responsible for data sets. The minimal model can be rendered as ISO 19115/19139, FGDC, RDF, a Dublin Core Application Profile, GeoRSS/Atom, in KML Metadata: presenting a flexible face to third party search services as much as possible.

Future Work

A "lego model" approach is taken allowing one to glue open source components to the edges of existing systems. Taxonomy and ontology are left very open where the use of URI schemes for keywords is recommended. Data annotation is distributed with the emphasis not on how the data is described, but on how it is used. For this reason user registration will in the future be required by GeoPortal component. In the near future duetopia promises closer connection and more feedback between data registry/repository services, and GIS application clients. For example, an analyst, in linking data sets, creates valuable knowledge which, if recorded for use in a search service, is a 'tacit recommendation' for others.

In future versions the access to data can be limited by time, space or person. The common GRID world concept of a Certificate Authority, a third party vouching for the integrity of both the client and the service in an exchange, will be used.