dtSearch is a text searching and retrieval application that's been around just about
as long as I can remember. All this experience has given them time to refine their
algorithms, and they deliver very fast indexing and searching as well as the ability
to manage collections of documents covering terabytes of text. At the same time,
they haven't stopped innovating - their current beta includes 64-bit and Office
2007 support, and their shipping version covers C++, COM, Java and .NET APIs (as
well as searching straight from their own user interface). This weekend I took at
look at setting up their Web product, which is designed to allow adding search to
your own Web site. I was pleasantly surprised at how simple it was to set up.
There are three basic steps to hooking up dtSearch Web. First, you need to decide
what to include in the index. This can be a collection of documents (not just text
- it can also parse documents such as Microsoft Office or Adobe Acrobat files) or
a Web site; the program comes with its own spider, which respects robots.txt, for
the latter purpose. I ran the spider over the whole Larkware.com site, which only
took a few minutes. One nice thing about the latter is that it can cache everything
it sees, so that if your site is dynamically generated, the search interface doesn't
force everything to be regenerated when displaying hits. The next step is to install
the dtSearch Web software itself, which is simply a matter of running a wizard,
selecting a Web site, and clicking a button; the software takes care of little details
like properly registering its ISAPI extension with IIS on Windows 2003 (IIS 4 or
later is required). Finally, you build a search form. Or, more accurately, you let
dtSearch build you a search form. You can customize it if you want, but I'm lazy.
The results? Well, judge for yourself: click this link to give the new
dtSearch form on Larkware a spin.
dtSearch offers more options for complex searching than I've exposed on that particular
form. You can do phrase searching, Boolean searching, proximity searching, phonic
searching, wildcard searching, and more. There's even a semantic concept searching
option that uses a thesaurus to search on similar terms. Depending on your needs,
you can get dtSearch for anything from your desktop to an entire enterprise. Visit
their Web site for more information on the pricing options as well as a 30-day trial
download.