Integrating geodata discovery services with OpenSearch

Abstract

When publishing information in the web, the aim is to make it reach all the

potential interested users, thus content discovery mechanisms are a key web

component. Web content discovery is mainly achieved with general purpose

search engines, which use text indexation to select content by keywords.

In the case of Geographic Information (GI), exposing content to text search

engines is not effective. GI needs specific spatial indexing, and use of special- ized format parsers. Google holds a geoindexing engine, where content can be

exposed to. But this needs specific data treatment, and indexing process is out

of the control of data publishers.

In the case of thematic databases, it can be convenient to provide a cus- tom web search service. Some well known examples are on-line dictionar- ies (Wikipedia.org, Wordreference.com), or stores (Amazon.com, Ebay.com).

A9.com proposed a widely used specification called OpenSearch, that provides

means to describe a search engine, its accepted parameters, response formats

and autodiscovery mechanisms.

Geographic Information repositories can also benefit from the use of dedicated

search services. Several approaches to GI search and discovery mechanisms can

be found: The mentioned Google’s Geoindex, the classical OGC standards ap- proach via Catalogue Services for the Web (CSW) and filter description syntax

(used for instance in WFS feature selection), or the modern Web 2.0 services

approach, consisting of specific web access interfaces (the so-called APIs). This

variety of strategies results in a fragmentation of geodata discovery techniques,

and increases the user’s needed effort to reach geographic contents.

There is a need for a simple, common interface that can be applied to a wide

range of geographic information services. OpenSearch is extensible, and its

“geo” draft extension can be used in a variety of scenarios to fill this gap,

providing a standard, lightweight and flexible discovery mechanism for GI. This

works demonstrates how OpenSearch-geo can be applied successfully to a wide

range of scenarios.