Site Search Engine: High Level Functional Specification

Search Engine: Functional Specs
www.USNA63.org

This page provides a high level Functional Specification for a Search Engine to be deployed against the pages of the Class of 1963 web site: http://www.USNA63.org.

The overall model for the design of the Search Engine can be found at the Web Site of

the Class of 1987, and
the United States Naval Academy site.

These are two different models, but both are lighting fast, thus they are models for what we want to do. The Class of 1987's version provides the hits without weighting the results; USNA's search engine provides an extensive 'relevance' weighting for the search text. For our purposes, the pages hit (as in Class of 1987) is sufficient, and these Functional Specs follow closer to that model.

It is the sincere hope of the Spec Writer that one or both of these referenced search engines can be acquired for the Class of 1963 either from the referenced Webmaster, or commercially, thus saving Hostmaster '63 the development and maintenance of creating a new such engine for the Class of 1963.

Earlier Hostmaster review of the Class of 1987 Search Engine found that cgi script to be large and unwieldy. However, if indeed the Class of 1963 is blessed with sufficient disk space, it is recommended that a mechanism be found to bring that search engine program in its entirety to USNA63.org. Here are some of Hostmaster's comments on the Class of 1987's apparent commercial product:

"The search engine I think you have in mind doesn't exist on the shelf in an acceptable fashion. I can do better than the META web crawler spoor, and I think I can do better than '87. As you point out, they have very little material to search and I'll speculate (without evidence) they are using a tool from their hosting vendor. "

A Glossary of Terms for search engines in general is available as an Appendix to these Specs.

Functional Specifications

Location/Availability: One of the reference models, the Search Engine for the Class of 1987 pages, is available (along with a link to the Site Map) at the top of every page on the site. Although laudable, this is not a requirement for the Class of 1963. Availability on the primary pages (USNA63 Home, Wardroom, Quarterdeck, perhaps Plan of the Day) is required, in that links to these primary pages are available in the navigation bar at the bottom of each page. Consideraton should be given to displaying the Search Engine box on certain secondary pages, such as Classmates, Companies, Foundation, etc.

Entry Box: A cgi box similar to that employed for the search entries on the Dynamic Class Roster should be made available to the user. The amount of text entered as <search string> should be at least 20 characters, but does not need to be greater than 255. Pressing on a 'Submit' button, or a carriage return entry should signify to the program that the entry is complete.

Search Capability: A basic word search should be initiated when the entry is submitted. All text pages (both .html and .html, in both Production and Development areas) should be searched. The engine should collect all 'hits' on the entered text string. No weighting need be given regarding how many hits occur on a particular page.

Display of Results: A page should be generated to display the Search Engine results. A title line should indicate "Search results for <search string>." A successful search implies at least one hit on one page on the site. Individual results should display as the URL of the 'hit' page, and, if feasible, the title of the 'hit' page. A maximum number of results displayed should be determined, such as 20, with a line somewhere on the Results Page that reads, 'Your entry is found on more than 20 pages on www.USNA63.org. The first 20 are displayed.' If the search is unsuccessful, an additional message to that effect should be displayed, along with either the search entry box itself, or an easy link to return to the entry box.

Error Checking: Error messages to the user should be kept to a minimum, much as they are with the Dynamic Search Roster. It should be difficult to enter a 'wrong' text string into the entry box; an unsuccessful search message should be the results if indeed such is the case.

Response Time: The user should not feel that the search time is excessive. The model engines are very quick. A response in less than five seconds is the strong goal, if not the requirement.

Nice To Have: On Results page, display total number of pages on the site that were actually searched.

Some Search Engine terminology is provided as a Glossary to these Functional Specs.

This page is
Functional Specs: Search Engine -

updated:
29 July 2000

Navigate to:

Shipmate Columns