Has the question of SharePoint Search vs. FAST Search come up in your planning sessions? Do your users have requirements that can “only” be met with FAST Search? Is Search Server Express an appropriate decision in your environment? Microsoft has provided a range of enterprise search products so teams have options when deciding how to provide search functionality in their project. This means that teams need to make an additional decision when choosing to provide search, which product is right for me? Every situation needs to be evaluated on the merits, but I’ll provide you some quick questions to point you towards an appropriate product.

1. Installation on a single server and no licensing costs? Go with Search Server Express. This edition provides for up to 10 million items(which requires installation on an external database. No database, you’re limited to approx 500,000 items.)

2. Provide Search using out-of-the box features for up to 50 million items? Go with SharePoint Server 2010. SharePoint Server 2010 provides search capability for up to 100 million items, so 50 million items still provides plenty of upside room to grow. However, if the corpus of information to be searched is under 50 million items, SharePoint Server 2010 should be capable of meeting your needs.

3. Searching between 50 million and 100 million items? This range is not so cut and dry, both SharePoint Server 2010 and FAST Search are reasonable choices at this range. The hardware requirements for both search technology will be similar, so more research is necessary at this level? Are you sure that you won’t go over 100 million items to index? Are you planning on augmenting SharePoint Search with third party functionality that may be available in FAST Search? What are the needs defined by your search administrator (at this level, search rankings, relevancy, and sources needs to be actively managed, and is likely becoming a full time role).

4. Searching over 100 million items? Go with FAST Search for SharePoint. Microsoft has tested FAST Search with over a billion items and provides prescriptive guidance for sizing and operating farms well over 500 million items.

5. Building a Search specific application? In addition to customer/user facing applications, these technologies are appropriate for component use by developers. Your architect and development team needs to define their needs and look at the extension points available in these products to determine the proper product. The FAST Search architecture provides many more extension points to allow developers to customize how the engine crawls, indexes, and organizes search data. Episode 64 of the SharePoint Pod Show goes into detail about these types of decisions, and I encourage you to give it a listen.

I remember the first time I saw Google Suggest, it would run the search in real-time and provide a dropdown with possible queries. One of my teammates always joked about making software with one button that would automatically know what you wanted to do. I showed it to him and suggested that Google was getting closer.

I've been working on my presentation for CapArea.NET about jQuery and SharePoint and I knew I would have to demo this search as you type code. The proof of concept was created by Jan Tielens, and the solution was fleshed out by the team at muhimbi. It overrides the default search box and starts returning results as you type. It's super cool and it's all HTML and JavaScript, it's not a web part, and there are no assemblies. Just add this to your master page, and you get a search upgrade.

While I was testing it I realized one thing. This search is using the keyword search. That's not too helpful since no results get returned until you spell out the entire keyword. I ended up using the LIKE operator against the Name and Description to get a much more helpful search that brings up results while you are still spelling the keywords. It required two changes to the javascript.

1. Changing the query to use the FullText rather than the keyword query.
2. Changing the jQuery selectors as the return values are different for the different query types.

In addition to this I changed the path to jQuery to load from the 12 hive, and I added some protection against SQL injection in the query.

Update 9/26/2011: David Lozzi has extended this code (and code contributed by several others) to create the Search as You Type Project on CodePlex. I would encourage all of you to take advantage of his hard work with this great project.

Don't let this title fool you, this isn't about searching Office documents in SharePoint. This is about searching your SharePoint server inside your Office application (Word, Outlook...). This functionality has existed since Office 2003, but whenever I show it to people, they love it.

1. Launch any Office component (Word, Excel, etc...).
2. Open the Research Pane
   1. Office 2003: On the Tools menu, click Research.
   2. Office 2007: On the Review tab, click Research
3. In the Research task pane, click Research options.
4. While you are here, I usually disable research services that I don't use.
5. Add research services, click Add Services.

   
   Figure 51 - Share Point Portal Server 2003

6. Add the path to your MOSS search web service, type http://your root directory/_vti_bin/search.asmx.(Example:http://Share/_vti_bin/search.asmx.)
7. Click Add. The service is automatically enabled for searching, and it will appear in the Search for list the next time you open the Research task pane.
8. To use the integrated Search you can
   1. Select the portal name in the Research task pane and enter a search term, then press the Search icon
   2. (In Word and Outlook), select a word or phrase, right-click and choose Look Up. This is my preferred approach because you don't waste the real estate of having the Research pane open when you don't need it, and it is fewer clicks to open when you do need it.

This can also be configured in your Internet Explorer Research settings.

This approach provides the default search options. If you want to create your own specialized search, there is plenty of documentation.