dtSearch Desktop Version 7
By Mike Riley
What a difference a few years makes in the software tools market. When I reviewed dtSearch for Informant s Microsoft OfficePro (the precursor to asp.netPRO magazine), I was delighted with the efficient indexing and search capabilities dtSearch Version 6 brought to the Windows platform. Since that time, however, Google has become the predominant search portal having bundled their algorithms into a desktop package called Google Desktop (available for free at http://desktop.google.com/?promo=mp-gds-v1-1) and a preconfigured network appliance designed to spider corporate intranet and file servers. Microsoft has also entered the search market and is directly competing with Google for search dominance. Like Google Desktop, Microsoft s MSN Desktop Search, available at http://desktop.msn.com, is also free. Because dtSearch Desktop promotes its primary use as a desktop search utility, its root value proposition has been commoditized and what once seemed well worth the money has been readjusted by its competitors to what should be a free utility.
That said, dtSearch Desktop does carry several advantages that advanced users will appreciate. Search queries can be extremely granular with multiple features such as stemming, phonic, fuzzy, and synonym search filters applied to the request. Indexes can be targeted to specific directories or over 40 different supported file types, ranging from Microsoft Access databases to Microsoft Outlook e-mail stores to Word, Excel, and PowerPoint and Adobe Acrobat documents. In addition to local files, dtSearch can also be used to spider and index entire Web sites, providing extremely fast search results on unwieldy corporate intranet sites, as well as your favorite Internet sites. dtSearch also beats Microsoft s WinFS technology to the punch by allowing its indexes to be shared across a network, thereby providing licensed clients the ability to aggregate desktop contents for distributed client libraries.
Figure 1: Indexing time depends on the volume of data to be analyzed. This example took about an hour to index 4.2 gigabytes of data.
However, compared to the free desktop-centric alternatives, there are some annoying disadvantages. Unlike the constantly updating search indexes of Google and MSN clients, dtSearch indexes have to either be manually updated or scheduled via the Windows Scheduler, preventing real-time additions to the index. There is also no easy interface for non-technical users. Google became the search champion not only because of its relevant search results but also because of its ridiculously simple, single textbox entry that masks enormous complexity on the back-end. Although one could be programmed for dtSearch, it s nevertheless a feature that should have been included to bring it to parity with its competition. In addition to this problem, the relevance of the search results compared to Google or Microsoft s desktop utility is also a challenge.
The days of highly refined search queries with special keywords and symbolic syntax are fading. End-users expectations on relevant results using simple phrases, keywords, or, in the case of commercial Web sites like AskJeeves and Brainboost that use natural language English queries, have permanently altered the perceived effort required to construct a meaningful query. Given the years of experience dtSearch Corp. has in the Windows (and more recently, Linux) desktop space, the company could have leapfrogged the alternative free offerings with truly amazing innovations. Lamentably, such innovations will have to wait for future versions.
Figure 2: Results are displayed as a list of relevant files meeting the search criteria and a representation of the selected file in the viewer pane.
Although dtSearch offers a superior API compared to that of Google or Microsoft, once again these alternatives also offer hooks into their technology to leverage and extend for custom purposes. dtSearch does provide several examples in languages and platforms ranging from Java to VB6 to Delphi to .NET that allow developers to create rich-client custom front-ends, but the practical use of this need may be limited to kiosk and specialized document management applications. The clincher for Microsoft users is that many developers are waiting to see what Microsoft s WinFS search APIs will offer especially in light of Microsoft s intent to compete vigorously with Google by potentially bundling such capabilities into future versions of Windows.
Figure 3: Search queries can range from rudimentary to complex. It takes some time to understand how each of these parameters can affect return results.
I really wanted to replace my reliance on Google Desktop with dtSearch, and I was excited by the programmatic, targeted search indexes and broad file type support dtSearch offers in their latest release. However, after using both products for several weeks with the intent of uninstalling the losing search utility from my desktop, the program that lost its place was dtSearch.
Figure 4: The dtSearch API is now available for the .NET platform, with good documentation on the exposed interfaces and assemblies.
Perhaps the next version will regain my confidence and infuse enough excitement to justify the $200 purchase price. But for now, the free alternatives offer a majority of features for most desktop customers, save the more advanced power search users and document management programmers.
Mike Riley is an advanced computing professional specializing in emerging technologies and new development trends. He also is a contributing editor for asp.netPRO. Readers may contact Mike at mailto:[email protected].
Web Site: http://www.dtsearch.com