Making a Web Site Searchable

Requirements

We all know how to search the Internet. But what if you want the visitors to be able to search your own web site? If you're a big enterprise, you can buy this service from Google. If you are like us, you have to provide the search capabilities yourself. See the little text input field and the Search button to the left of here? Go ahead, try it! Can you have the same functionality on your web site? That depends on your ISP. If you are an ISP, you might consider providing this service to your clients.

The major requirement is that the web server must be running Microsoft software. Sorry, Unix! It turns out that the content index--the software that makes quick searches possible--is built into Windows XP (if you are running XP on your computer, you can index your local disks as well). Reliable Software's web site is hosted on a dedicated server running Windows XP, so we were able to convince our ISP to turn the indexing on. I will describe all the steps in detail.

Creating the Index

This part of the job must be done on a web server by the ISP. The steps are pretty easy.

→ Create a directory for the catalog (content index database). For instance, in our case it was c:\index\relisoft.
Note to ISP: you can have multiple catalogs on the same server, one for each web site.

→ Open Computer Management window from Control Panel > Administrative Tools.

→ Click on Services and Applications and select Indexing Service.

→ Select Action > New > Catalog from the menu. Provide Name (e.g., relisoft) and location (e.g., c:\index\relisoft).

→ Open the Properties window of the newly created catalog. Click on the Tracking tab. Select the WWW Server (in our case it is the default server).
Note: this particular option appears only on the server installation of Windows XP.

→ You can also modify the default Generation options. For instance, you might want the indexing service to index files with unknown extensions (by default most common web file types are indexed). You might also modify the size of the Abstract produced for each file. This abstract may be displayed as part of search results.

→ You can instruct the indexing service as to which directories to index and which to skip. Go back to Computer Management panel and select your catalog (see picture at the top). Use the menu Action > New > Directory. Type in the full path of the directory and select one of the Include in Index radio buttons (see picture below). All subdirectories of the specified directory will inherit this property (unless overriden by another entry in the directory list).

Note to ISP: For multiple web sites, create separate catalogs and specify different root directories to be indexed in each.

→ Restart the index (select Indexing Service and click the Start Indexing button (you might have to first stop indexing, if it's already running). For most web sites, the initial indexing will take only a few seconds. After that, every time you make a modification (edit, add, remove files) to the web site, the index will be updated almost instantaneously.

Next: Creating a Search Form.