Adding keyword search to a Web site

The approaches to adding search range from minimal in-house Web programming to programming a system from scratch.

Using services from a search company (e.g. Google)

These approaches take advantage of services that already exist.

• Directly access search engine

  • Web page provides form that directly queries search engine
  • Inexpensive (use of Google is free)
  • Requires no Web programming
  • Allows for little customization
  • Site pages must be accessible externally
  • Example (when my tomcat server is running)
  
  

• Purchase software and hardware dedicated to your site

  • Configurable but requires little in-house programming
  • Customized to your site's needs
  • Can search intranet content
  • Consulting and product can be costly
  • Google's search appliance is one example

Using an API that links to a search engine

• Server page accesses search results through an API
• Query can be tweaked to limit search
• Display of results can be easily customized
• Some programming is required for issuing query and displaying results
• Site pages must be accessible externally
• Example (when my tomcat server is running)

Here is a link to the Google Search API.

Develop your own search application from scratch

• Requires the most programming
• Allows for full control of search
• Display of results is fully customizable
• Example (when my tomcat server is running)

Factors that affect search results

• Scope of search (just index pages, just links, or all site content?)
• Removing "stop" words (e.g. and, or, the, about)
• Case sensitivity
• Word stemming (e.g. "computing" or "compute" might become "comput")
• Matching synonyms and related words

The source code for the examples on this page is available in this zip file.