FAST Technical Resources

How can you help ensure that your search project is a commercial and technical success? You have a better chance at success when you understand the stages, stakeholders, and requirements for each phase in the search project. This way, you can align your business objectives continually with the implementation, deployment, and maintenance of your search project and thus be able to achieve your strategic goals for search.

Content is the external data - from many different structured and unstructured sources - that is fed into a search engine. Before being stored, the content is refined for optimal retrieval.

Linguistic tools, such as spell-checking of queries or grammatical normalizing of content or queries, can simplify the user's search experience greatly. However, as with most other aspects of search, it is essential to know the users for optimal application of linguistics.

A search engine may be viewed as a set of servers that will run with little or no intervention. However, to guarantee user satisfaction, search performance and quality must be continually improved. The prerequisite is to monitor the performance and behavior of the system.

Relevance models determine the ordering of search results, but such models need not be based only on algorithmic information retrieval methods - they also can incorporate the business logic at hand to organize the answers in a way that is optimal, commercially or organizationally.

Good search engines do most of the hard work behind the scenes and thus simplify the user's search experience. For example, vague or misspelled queries can be refined in query processing, and results can be filtered, merged, and post-processed for intuitive navigation in result processing.

Relevancy is the measure of how well a set of results answers or addresses the intent of a given query.

Often, search applications are integrated into large software applications or complex information systems. These environments can be complex, and there are multiple integration points. Careful planning and consideration of the users' needs are a good investment to help ensure long-term success of search integration.

Enterprise search is mission-critical for many enterprises. For example, e-Commerce sellers or criminal investigators cannot afford for their search boxes to be out of order. In the face of inevitable hardware failures, search systems use redundancy that is architected carefully for constant availability. Search providers should differentiate between availability on the content ingestion side and on the user query side.

Search engines are built to provide easy access to available information. But in an enterprise setting, it is essential that this information is delivered in a more secure way. Good solutions manage to combine and balance security requirements with search.

The old adage, "If the user can’t find it, it ain't there," is true for search applications - for both the query entry page and the results page. Designers of search applications sometimes forget that their users tend to be less technical.