Chapter 6 Geographic Search Engines (p.143-144)

Andrew Daviel,
Felix Kaegi
Geotags, Inc

1. INTRODUCTION

Most people now are familiar with the concept of a "Search Engine" on the Internet, such as Google or the original WebCrawler. These search engines perform what is termed "open text" search, where the user enters some keywords or phrases, and the search engine returns a list of pages that contain one or more of them.

Many people are also familiar with the concept of a Metadata search, though they may not realize it. Most public libraries have an index, usually now on computer, and the user can search it for particular items. It is not possible to search for phrases within books, instead one must search for information about a book, such as Author, Subject or Title. These entries are termed Metadata (data about data) and in libraries have historically been compiled manually.

HTML documents have historically contained only one explicit item of metadata - the title - but may optionally contain arbitrary metadata using the HTML META element, most commonly Description and Keywords. Opentext search engines such as Google index some of this metadata, but treat it as part of the text (although some elements such as Title may be preferentially weighted compared to words in the body of the document)

A text search is not the only possible form of search, however, although it is now the most ubiquitous. For instance, a researcher might wish to find places which have less than ten centimetres of rainfall a year, or a motorist may wish to find the nearest service station. This kind of search is usually done with a database where it is known that certain fields have specific properties, such as being numeric.

The Geotags search engine described here is an example of a locationbased search engine - that is, one that can answer the question "Where is the nearest restaurant ?".

2. HTML MARKUP

While a normal search engine ranks pages according to their relevancy and popularity, the Geotags search engine ranks them according to their distance from a particular point. For this to be possible, each page must be identified with a geographic position. Currently, it is required that each page submitted to the search engine to be indexed must include geographic metadata in the form of HTML META elements. The META element is familiar to a majority of document authors since keywords and description identifiers may be used to influence many search engine results.

The format of position data used for the search engine was chosen to be simple, compact and unambiguous, since it is believed that in many cases this data will be transcribed from navigation units, gazetteers or other documents and manually added to web pages. Two elements are used; a (Latitude, Longitude) coordinate pair given in decimal degrees of arc, and a region code taken from either ISO 3166-2 or ISO 3166-1.

These region codes form a restricted vocabulary that is unambiguous and suitable for machine interpretation. These two metadata elements may be easily added to existing HTML or emerging XHTML documents, and are in a standard form understood by many HTML authoring software packages and human authors. The following are examples of the position and region elements: