September 24, 1998
The Intranet Search: A Matthew Bender Case Study
By Tanyalynette Rosmarino
"If a WebMaster waits until the HTML/content is perfect, a corporate intranet would never blossom and grow." says Matthew Bender's T.L. Rosmarino.
|
|
Learn to search before you crawl
When you work at a
multi-million dollar law
publishing company, nothing is
more important than content
(except maybe spelling).
Content, content, content.
The notable exception to this
rule is, of course, the
ability to find the content. As
the information on my
company's intranet became
more abundant, locating specific content became more
complicated. Unless you had
been using the intranet from the
start, you were in for some clunky navigation. "I
want to give the users the ability to search the intranet"
became my mantra. As the Intranet/Extranet Web Administrator, it
is my job to not only build the better search feature,
but to build it to last until the next release is
proof-of-concept.
My first task was to sit down and figure out what it was
that I wanted to search. Because most of the documents
available currently were text-based, any simple
search would do the job. But what are the hidden costs in
a simple Band-Aid? Answer: The two-Ms: manpower
and money. Would the search component stand-up to the
PDF, RTF documents, and screen shots I knew where
being developed? Would the users be able to search on
more than just a few keywords? Did I want thumbnail
shots of the documents or a document summary? What
searching needs did my users expect, and which
features would I like to make available to them? Version 1
could be a simple text-based index search and get the job done, but it
would also need to be able to adapt to
daily file changes.
Which brings me to the crux of intranet development: The
intranet administrator is the name and face
associated with the content, regardless of any other
development teams. To release "perfect" HTML is to
dream. Every seasoned Webmaster knows that the first
release is the test release. End-user feedback is required
before the first rewrite, along with a reinvestigation
of the end users' needs. If a Webmaster waits until the
HTML is perfect, a corporate intranet would
never blossom and grow. For all the sunshine in the world,
a little rain must fall. Don't be afraid of posting
files too
soon. The best intranet applications were developed by
gathering user input of available intranet
functionality. In short, build it, let them come and criticize, then make it
better.
I do not have a lot of time for R&D or for new problems; my time is my most
precious resource. I wanted a search engine that could be quickly
downloaded, decompressed, and installed with no problems or hidden
"features." I wanted the search butler to be easily
configured and quick to start. Most importantly, I wanted
the same interface to be available for both my intranet
server and my extranet server. Lastly, I wanted a
pre-written HTML front page.
All right--straight to the Net. (How did anyone do research
before the Net?) One stop at Metacrawler and I had
a list of more than 100 intranet search engines.
Finding a search engine that ran on an NT-based
Intel-equipped machine was much easier than trying to find the URL for
Delta Airlines. My other qualifying factors included Boolean
Search and keyword phrases.
|
