Desktop search is the holy grail of personal computing. We need to be able to find the exact file we want, when we want it, without waiting all day for the search to complete and without needing to continually refine the search to drill down for something better. Amazing desktop search was the promise of the eternally delayed WinFS component of Longhorn, dating back further to the ancient Egyptian file system formerly known as Cairo. Spotlight in OS X Tiger comes close to this vision but is still lacking in some areas. Longhorn will improve the search function in Windows through the elimination of Rover and the addition of some metatag indexing. These improvements raise a few privacy concerns that need addressing in terms of sharing files in the open, but they solve other privacy issues with the Secure Startup infrastructure.
In the meantime, the big three Web search companies--Google, MSN, and Yahoo--are competing with Copernic to provide the best desktop search experience. For this article, I'm focusing specifically on Google Desktop Search, MSN Search Toolbar with Windows Desktop Search (formerly called MSN Desktop Search), and Copernic Desktop Search, because I haven't successfully gotten Yahoo Desktop Search to function without crashing or slowing my system to a crawl in a Windows XP SP2 environment.
The choice among the three took on urgency recently, because Microsoft released the final version of Windows Desktop Search.
The common theme between the three apps covered here is indexing data on the hard drive more effectively than with the indexing built into Windows. Each company achieves this goal with varying degrees of success. If you use any of these services full-time, by all means tell XP not to index your hard disk, because when XP indexes, you take a big performance hit. To do it, right-click on your hard drive in Windows Explorer, choose Properties, and uncheck the box next to "Allow Indexing Service to index this disk for faster file searching," as shown in Figure 1.
Figure 1. Uncheck Allow Indexing Service to eliminate the performance hit
The third-party apps index a different subset of files on your hard drive, consuming system resources in the process. Ideally, you need to run the search tool all the time so that it catalogs changes to your information in the background, providing an accurate picture of your files at all times.
Copernic Desktop Search, shown in Figure 2, indexes file metadata and information for the following applications and file types stored locally: Word, PowerPoint, Excel, WordPerfect, PDF, HTML, .txt, .rtf, .asp, .c, .cpp, .cs, .csv, .h, .inc, .ini, Java, .pas, .xml, and .xsl. In a more limited capacity, it also stores details for .gz, .hlp, .rar, .swf, .tar, and .vsd. This makes it a convenient resource for users and developers, because the attention to code files makes finding a particular line of code on your system just as easy as locating the document where you referenced a quote from Blade Runner. Files may be located on any drive connected to your system or any mapped network drive that reconnects automatically.
Figure 2. Copernic Desktop Search presents targeted search options when you type in a query
Email indexing supports Outlook 2000+, OE 5+, Eudora 5, and Thunderbird. If you use more than one of these applications on your system, Copernic will index as many of them as you like. Indexing of contacts is supported for Outlook, Outlook Express, Windows Address Book, and Thunderbird.
Browser favorites and history are indexed for Internet Explorer, Firefox, Mozilla, and Netscape 6+. Copernic requires you to pick one browser to index; it doesn't watch them all. Favorites are indexed by default. Browser history indexing is turned off by default.
Copernic currently does not search IM chat logs.
Music, image, and video metadata is indexed for the following file types: iTunes, MP3, OGG, WMA, WAV, EXIF, JPEG, GIF, MPEG, MOV, WMV, and ASF. Copernic's advanced configuration is able to label other known audio and video file types as media formats, but metadata scanning is currently limited to these types.
Results of searches are displayed inside the application interface, organized by file type.
Google Desktop Search, shown in Figure 3, supports full text search of Word, Excel, PowerPoint, PDF, HTML, .txt, and .rdf files by default and can be extended to support WordPerfect and a variety of programming language file types with third-party plugins. Files located on other machines may be searched using a third-party plugin, assuming all PCs are running Google Desktop Search.
Figure 3. Google Desktop Search displays results as a full page of results in your browser--click for full-size image
Email support for Outlook, Outlook Express, Netscape Mail, and Thunderbird is included in the base install. Gmail indexing is supported with a plugin. There is currently no support for searching contact information or non-email components of personal information manager products. A third-party plugin adds a Google search bar to work in Outlook directly.
Supported browsers include Internet Explorer, Netscape, Firefox, and Mozilla. Favorites and Web history are both indexed by default, with the option to turn off the indexing of secure Web pages during the install process (although it's checked by default).
AOL Instant Messenger chat histories are indexed by default, although you are prompted to uncheck this feature during initial configuration.
Music, photo, and video metadata is indexed for common file types including MP3, OGG, WMA, WAV, JPEG, GIF, MOV, and WMV. Other file types, such as MPEG, are indexed but don't return any metadata. Cataloging of media files is limited to whatever Google predetermines is a media file, until someone creates a plugin for the toolbar.
Search results are displayed in a pane that looks identical to a Google Web page, organized by Google's predetermined relevance. I find this confusing in determining where I'm searching for information, because it looks the same as going to the Google home page (which adds a Desktop link to the home page when you visit Google).
Windows Desktop Search, shown in Figure 4, primarily supports searching files associated with the Microsoft universe. Word, Excel, PowerPoint, HTML, .txt, and .rdf files are supported. PDF support is available via a third-party plugin. Windows Desktop Search supports the searching of My Documents only, the entire computer, or configurable locations anywhere accessible on the network.
Figure 4. Windows Desktop Search displays results as you type
Email support is limited to Outlook and Outlook Express. The indexing of attachments is optional, with indexing occurring only when the email application is open. The upside to this limitation is a solid depth of search, supporting the search of calendar items, notes, and tasks in addition to email.
Browser history is not indexed by MSN's search. Favorites are indexed automatically, although there is no option for indexing alternative browsers like Firefox or Netscape.
Music, photo, and video files return as much metadata as is available, focusing primarily on the file's name, author, and embedded comments.
Search results are displayed on the fly in a box that pops up from the search box in the toolbar. Pressing Enter displays more details in a Web page that is distinctively different from the MSN Search home page. The search results are broken down by file type, but may be filtered to display results only from email, for instance.
Which is the best of the three? I'm biased toward Copernic because it is far more flexible than either Windows Desktop Search or Google Desktop Search. Not only does Copernic allow you to configure which files on which drives are indexed, but it also lets you configure the indexing frequency.
That Google makes my desktop search virtually indistinguishable from a Web search is a major turnoff, especially since it commingles my Web history with local files if I have history indexing turned on. MSN's pop-up box that filters as I type is one of the more elegant features in this category of products. All apps return fairly similar results when using the same search terms.
Based on my usage of the three products, MSN and Copernic do a better job of peering inside files to find information than does Google. If you use many alternatives to Microsoft apps, Windows Desktop Search won't make sense because it limits you to searching primarily for Microsoft file types.
One concern that all the apps ignore is the issue of personal information security. Indexed data is not encrypted, which means it's accessible without authentication. Proceed with caution when indexing private information on your hard drive.
Despite the inherent risk, though, all three apps are a better search solution than the built-in Windows Explorer search function. Until someone comes up with a better product, or Longhorn proves it can find my data more effectively, Copernic Desktop Search will be my personal data finder.
Jake Ludington is the author of the best-selling guide Converting VHS to DVD. He publishes audio and video tips at MediaBlab.com.
Return to the Windows DevCenter.
Copyright © 2009 O'Reilly Media, Inc.