Introducing Faceted Search

This is an example of how Faceted Search can be implemented in a site:

The main features of Faceted Search are described below.

Facets

Facets are a central concept within Faceted Search. They are used to group the search result by category, for example the results over a certain period, results by author or by keyword. Facets can be used to refine the search query, they help you to 'zoom in' to the result you are looking for.

If you select an entry in a facet, it is added to the search filter.

Suggestions

When you start typing a search term, the search suggestions dropdown pops up. It will list suggestions based on previous searches that were successful. Since Faceted Search records all search queries, the suggestions system is self-learning.

Did you mean

The recorded search queries are also used to implement the 'did you mean' feature. Previous searches are used to display search terms with a similar syntax. By default, 'did you mean' is only displayed if the current search has no results.

When using SQL Server 2008, Faceted Search will also display the removed noise words and the thesaurus enrichment.

Ranking

Search result are sorted by relevance. The ranking system used many parameters to rank the search result. Most of these parameters can be configured and fine tuned to influence to result order. The result can also by sorted on other criteria, such as publication date.

Load and save queries

Search queries (including all parameters such as filters), can be saved for later use. This is especially useful for dynamic sites were content changes frequently. Site implementations can also use this information as a source for a personalized newsletter or RSS feed. 

Search expressions

Advanced search queries can be constructed by using boolean AND and OR, searching for literal string using quotes or excluding results using NOT.

Search within results

The search can be refined using the option 'search within results'. This is an easy way to limit the search results especially suited for those who are unfamiliar with search expressions.

Result summary

The search result will be displayed in a paged list. For each result, relevant information can be displayed, such as the publication date and the author. The result can also include an abstract of the searched content. The search terms are used to generate an abstract that is relevant to the search query.

Result highlighting

The search terms are highlighted within the results pane. Also, the links to the target items are enriched with the search terms. By adding a behavior to the site configuration, result highlighting can even be activated for the link targets.

Only search within titles

If you are looking for a specific content page, it is sometimes useful to limit the search to one or more specific content fields. By default, the search can be limited to the title field, but other fields can be added in a similar fashion.

Word and PDF documents

By using SQL Server full-text indexing, Faceted Search can also search within Microsoft Office documents, PDF documents and any other (binary) file format for which an iFilter is installed. Documents stored in de file system and accessible by Smartsite can also be included in the search using FileLink items in conjunction with SyncFileLinks.

Rich user experience

The Faceted Search user interface makes full use of progressive enhancement. It can be used in a non-scripting environment, but, whenever possible, it makes full use of modern browsers. The results pane is updated using an ajax callback and the facets are updated asynchronously. Facet boxes can be collapsed and expanded without the need for a screen refresh.

Customizable

Because Faceted Search is built as a set of smartlet, it is an open implementation which is fully customizable. Customization can reach from setting specific smartlet properties and using overriding css classes to creating a complete custom version of the smartlet.

Advanced Filters

Faceted Search uses filters to limit the search result. Filters are normally added by selecting a facet entry. In advanced mode, these facets filters can be inverted (NOT) or combined using boolean AND or OR operators. Filters can also be removed individually and are persisted when saving a query.

Stemmers and word breakers

By building on the SQL Server full-text indexing technology, word breakers and stemmers are used. This involves finding word boundaries (word-breaking) and conjugating verbs (stemming). Word breakers and stemmers are language specific, and the rules for linguistic analysis differ for different languages.

Thesaurus

When using SQL Server 2008, a special faceted Search thesaurus can be exported to the database server. This thesaurus can be used to replace or enrich search terms.