Microsoft SharePoint 2007 Technologies

120 6.2 Configuring Search

A familiar search experience across all applications is also a goal for

Microsoft (and the end user!), and this is being realized through the use of

the common search engine. As mentioned previously, Windows Desktop

Search, which is used by Windows XP and Vista, is also used by Outlook

2007. Exchange 2007 also harnesses the common search engine so that

applications like Outlook Web Access and Exchange ActiveSync also provide

a similar search experience to the end user.

Comprehensive indexing is also a goal, and the protocol handler and iFil-

ter architecture help here. The BDC is also crucial here, as items held in

back-end business applications can also be indexed and returned in search

results.

Lastly a flexible, scalable, manageable, and secure search system is a key

goal. SPS 2003 suffered here, particularly in larger enterprises where, for

example, it was difficult to keep large indexes up to date on query servers due

to the way index propagation from index servers functioned.

6.2 Configuring Search

Let’s take a look at the major components that you need to think about when

configuring search.

6.2.1 Managing Search

Search is a MOSS shared service and, as such, is managed via the SSP

Administration option in Central Administration. With SPS 2003, the entry

point into the management of the various components of search was a bit

untidy and resulted in navigation frustration. In SharePoint Server 2007, the

overall administration has been much improved, and a bird’s-eye view of all

administration tasks is available from a single page. This page, shown in Fig-

ure 6.2, groups the management tasks into three sections: crawl settings,

scopes and authoritative pages. Note that for WSS there is no graphical user

interface to search. It is enabled by default, and you have to use stsadm to

make any changes to it. Figure 6.3 illustrates the use of stsadm to list the

search configuration on a WSS farm.

The crawl settings section gives you access to the options that let you

manage the content of the index: content sources, crawl rules and log files,

file types, property mappings, and so on, as well as giving you an overall view

of the status of your indexing operations. The scopes section allow you to

manage what users see in the search scope pull-down on a standard search

page, but the main end-user customizations are performed in the search

results Web Part pages, as we will discuss shortly. Authoritative pages are used

to define Web sites that are tagged as being authoritative or nonauthoritative;

they ultimately affect the relevance ranking of search results.

6.2 Configuring Search 121

Chapter 6

As with SPS 2003, the physical full text index files are stored on NTFS,

but there has been a change in the location of other parts of the search sys-

tem. Details of much of the configuration, indexed properties, and other per-

tinent information such as anchor text from links are now stored in a SQL

database associated with the search service. Note that the definition of con-

tent sources is still also stored in the registry on index servers (at

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office Server\12.0\

Search\Applications\<appid>\Gather). Also, note that you have one index

catalog per shared service provider. This again is a change from SPS 2003,

but it simplifies the deployment and management of the index. Microsoft

states that a single index can support up to 50 million documents, so a single

index should suffice for most organizations.

A few areas of search have not changed that dramatically (or at all) since

SPS 2003. This includes the thesaurus, which is used to augment or replace

search query terms. The crawl log is easier to search in MOSS than it was in

SPS 2003, and you still have control over the file types that the engine

should attempt to index.

The ability to perform server name mappings—that is, the replacement

of URLs in the result set that is passed back to calling clients still exists. This

is typically used in situations where the content was crawled using an internal

URL, but the clients need to access the results via an external URL.

Figure 6.2

MOSS search

administration.