Finder Concepts

This section outlines the major concepts that make Finder a sensible choice with content-rich websites.

Keyword Searches

While simple, Joomla's native search engine is limited in both scope an performance.  It searches for parts of words using database wildcard characters which becomes more and more processor intensive as the amount of content in your site increases.  It is also unaware of the relevance of finding words in the title versus finding them in the body text or meta information.

Finder is different.  When content items are saved, it will be passed a number of fields to index just as the body text, meta keywords, meta description, the title and possibly the Url if in Search Engine Friendly format.  All HTML formatting is strips from the content and it is sliced up into individual words.  Finder allows you to assign a weighting to words depending on where they are found.  For example, you might add a weight of 5 to words that are found in the Title of the content.  This means that when searching for a word, content which has that word in the title will be considered more relevant than if it was just found in the body of the content.  Likewise, words in the meta description or meta keywords may also be given additional weights.

Because Finder indexes keywords, it is able to perform simple suggestions when the desired words are not found using what is called the Soundex algorithm.  Please note, Soundex is only suited to English words.

Taxonomy

Joomla's native search engine does provide some taxonomical support but it is very limited.  You can limit to search areas, such as categories or section, but you cannot break it down to individual (or multiple) categories or sections.

When Finder indexes content, it also has the ability to build a taxonomy system to which the content can be mapped.  Take regular Joomla Articles.  An article is usually found in a Joomla section and category.  When Finder indexes the article, it looks at what section and category the article is in and adds a map between the article and both it's section and it's category.  You can see this in the Content Maps are of Finder's administrator.

Finder's taxonomy is fixed at two levels deep.  The top level are called Branches.  A branch is a major container such as a Category, a Section, an Event or Venue (such as for the Event List extension) or Publications and Issues (such as in the Jxtended Magazine extension).  Under each branch are any number of children.  For example, under the Category branch would be the names of all the categories actually in use.  Finder's branches do not care where the categories come from - it considers that a category is just a category whether it comes from Joomla, DocMan, Magazine or some other extension.  This means that the user does not have to know about how your content is arranged when searching your site (just find me anything in the "News" category from anywhere please!).

Each taxonomy branch and child can be individually toggled for display in the advanced search.  For various reasons you may wish to hide various categories from your visitors, or whole branches - this could be to reduce clutter, allow the visitor to home in on the most relevant areas, and so on.  Taxonomy states are not linked to their sources.  This means that that a child of the Category Branch called "News" could be published independantly of the "real" Joomla category called "News".  Also, when the real category is deleted, Finder still keeps track of the category because it's possible that another extension might also be using a category called "News".

Search Filters

By default Finder gives you the ability to perform an advanced search which allows you to filter the result list by any combination of the taxonomy mappings.  For example, the advanced search can allow you to just find content in the News category (which could come from Joomla, or DocMan, or another extension using categories), or at a particular venue (in the case of EventList).  This is pretty neat, but sometimes you want to deliberately narrow the search which is available to the visitor.  This is where Search Filters come in handy.

An excellent example of using a search filter is when you are using Catalog as a contacts database.  You may have a generic search page to do the exhaustive site search.  But you could also create a Search Filter that only looks in the Contact Class that you created in Catalog.  You might assign this to a module called "Search Contacts".  Using that search allows you to restrict the results only to matching Contacts in Catalog.

Indexing

After the initial index that you would perform when you first install Finder, indexing will usually happen as save content is saved.

Extensions the we produce (such as new releases of Magazine and Catalog) will also syncronise indexed content when you unpublished it or delete it.  This is generally not possible in other Joomla 1.5 extensions due to the lack of triggers to intercepts the appropriate events.