Extending Search

We have discussed implementing search and scaling the search engine to support a large number of users, but what about extending the functionality of search? SharePoint provides APIs, web parts , and other components that make it possible to add new functionality or modify existing search functionality.

You can customize the user interface to search by modifying the search web parts. For instance, some users need more than the simple search provides but are confused by all the options in advanced search. By customizing the search web part, you could create a streamlined version of advanced search that displays or hides fields as you desire . Similarly, you could reorganize the search results web part if you wanted to highlight different elements of the search results.

SharePoint Portal Server provides an Advanced Search Administration mode. This provides additional settings that are not available by default to the SharePoint administrator. For instance, you can create as many search scopes as you like, and group together crawled sites for simpler administration.

Editing a Noise Word File

Some common words are ignored during the indexing process in order to reduce the size of the index and improve meaningful search results. For instance, a search on "the" or "an" would not likely return meaningful hits. SharePoint stores a list of noise words in a text file that you can modify. SharePoint comes with 20 noise files for 18 languages. By default SharePoint uses the neutral noise file noiseneu.txt in the server directory\Program Files\SharePoint Portal Server\DATA\Config.

The noise word file can produce some interesting results. For instance, users who abbreviate information technology to "IT" will be disappointed that no hits are returned, because "it" is included in the standard noise word file (Figure 13.25). All single character words are similarly excluded.

Figure 13.25. Default SharePoint Noise Word File

To include all words in indexes, delete all entries from the noise word list. Do not delete the file, as this will result in all single characters being treated as noise words.

You can edit the noise word file in any Unicode editor, such as Notepad. If you change the contents of the noise word file, be sure to run a full update of the index. Otherwise, the search engine may accept search terms as valid but produce no results because the term was ignored during the indexing process or reject terms that should be valid because you deleted them from the noise word file.

Microsoft SQL Server uses a noise word list of its own that is maintained independently of SharePoint. You may want to consider whether changes in the noise word list of one system should be reflected in the other.

Search Feature Comparison

The portal has many needs for search capability, and you face several options for how to implement them. You may find that an integrated approach that uses more than one product or technology works best for you. Table 13.2 summarizes the search capabilities of several Microsoft products.

Table 13.2. Search Technology Feature Comparison (Source: Adapted and updated from white paper "Microsoft Full Text Search Technologies," June 2001.)

	SharePoint Portal Server 2003	Indexing Service	SQL Server 2000	Exchange 2000 Server	Office XP on Windows XP
Crawls:
File system
					Local only
Web sites
		Local only, through file system
Lotus Notes
Exchange 5.5
	Public folders
Exchange 2000
	Public folders			Public folders and private mail boxes
SQL tables
	Through ASP
SharePoint Portal Server workspaces
Third-party protocols
Best Bets
Categories
End user UI	Dashboard site	Windows Explorer on Windows 2000 and custom	Custom	Outlook through Advanced Find, custom	Office search task pane

While SharePoint is clearly the most advanced search engine offered by Microsoft, you may want to consider how it fits with the other search products in the family. Most importantly, give your users what they need to bring them to the rich content you offer in the portal.