Factors you need to consider in the technical design of any search integration include:

• The API's exposed by the engine
• Your technical requirements in terms of the consuming application
• How simple you want it to be to consume the search service from your code
• How much of an engine's advanced search syntax or capabilities need to be exposed
• Management of search indexing and spidering activities
• Performance and scalability

I have experience implementing and integrating with a variety of search engines, including: Autonomy, Verity, Thunderstone, Oracle Text Indexes, MS SQL Server Full Text catalogs, and Lucene .

The integration techniques have varied widely based, in part, on the factors described above.  For example, Oracle Text searches can be performed via simple SQL statements executed through your environments usual database access support. In those cases, simple DB wrapper classes are often sufficient. For Thunderstone, an engine with a more proprietary interface, I actually developed a Thunderstone JDBC driver, allowing that engine to be accessed very simply by java developers comfortable with the JDBC API.

Contact Practical Apps for help implementing and integrating with these or other search engines.