Matrix, Search Engines: GNA Meta-Index

Globewide Network Academy's Meta-Index

GNA is a worldwide non-profit consortium of educational and research organizations whose mission is to improve education by creating a marketplace for online courses. It also offers administrative and technical services to organizations and individuals who are presenting such courses.

Key Links

• Course Catalog: http://www.gnacademy.org:8001/uu-gna/documents/catalog/topic.html
• Topic tree: http://uu-gna.mit.edu:8001/uu-gna/schools/vslis/topics/

• GNA Miranda Search: http://www.gnacademy.org:8001/cgi-bin/miranda
• GNA (full text) Catalog Search: http://www.gnacademy.org:8001/gna-catalog/catsearch

Organization

• In general the collection has no strong organization, but rather various areas are grouped under index pages. Key areas of interest are listed above, or can be found from the GNA front page.
• There are 2 different ways to search the GNA collection. Most of the documentation is devoted to a techical discussion of the search engines, but no single document to describe why you should choose one or the other. Ideally, the functionality would be unified into a single interface, but there is no indication that this will happen.
  • The Miranda search engine is based on common UNIX software for scanning text files, such as awk and sort, and performed with a simple shell script.
  • The Catalog search gives no indication as to its underlying engine, but clearly states that it is a prototype. It is used to search the document tree that describes the courses offered by GNA.
  • The server allows multiple keywords, which are treated as if linked by the Boolean AND. No other control is given over substring searches, except to eliminate a few common conjunctions and letters.
  • The server forces a limit of 100 or 200 documents to each search, with no user control over this number.
  • Search results display document name, topic, and the date it was added to the database, with other miscellaneous (and often empty data fields).
• None of these search pages require forms-support to use.
• The catalogs and the search tools only span the contents of the GNA Web site and user-suggested documents, which means that this server has little value for searching the Internet in general.

Administration

• The document placed on this archive are automatically searchable by the mechanism involved, but there is no effort to standardize the organization of the catalog as a whole. In fact, locating a particular piece of information proves to be rather difficult due to the poor layout of the documentation.
• Each of the search engines currently return results within about 20 or 30 seconds, but there are serious questions regarding the scalability of these tools. I feel that the search scripts used on this server are of the simplest design and lowest quality, and will need to be replaced as the traffic on this server increases.
• Documents are added regularly, and I could find no dead links. The documents describe the collection as a "compendium of several thousand Internet references."
• Additional Services
  • The help pages and other information seem to consist of strictly technical descriptions and policy statements.
  • A What's New page logs the addition of services and changes to the GNA archive.