[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference nsic00::eis_dw

Title:Executive Information Solutions & Data Warehousing Conference
Notice:Welcome to the Data Warehousing conference
Moderator:26002::HAGGERTY
Created:Thu Sep 01 1994
Last Modified:Thu Jun 05 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:499
Total number of notes:2932

445.0. "AltaVista as access tool?" by IJSAPL::OLTHOF (Spellchecked Henry Although) Thu Feb 06 1997 09:21

Hi All,

Combining data warehouse and internet technologies can be of interest
to many customers. One of my customers wants to integrate data in a
data warehouse and use the simplest tool on earth to access the data:
AltaVista.

The want to be able to use the single line interface, type someting
like products+geographies+1996 and get answers back. Question: can this
be done and if yes, how? What other tools are available? Does AltaVista
operate on databases anyway? Could AltaVista private Extension help?

A lot questions, hoping for some answers, cheers,
Henny
T.RTitleUserPersonal
Name
DateLines
445.1on a related note...EPS::HAGGERTYSBU ASE, Nashua NH USAThu Feb 06 1997 15:00141
[First! for the Web]

Rev Up Your Business -- Search engines are an integral part of the menu for
total Web solutions

---------------------------------------------------------------------------
Date: Thursday, February 6, 1997
Source: VAR Business
---------------------------------------------------------------------------

VAR Business via Individual Inc. : With all the information available today
on the Web, how do you make finding what you're looking for easier? "The
key is in knowledge management," says David Yockelson, vice president and
service director of advanced information management at META Group, a
consulting firm in Stamford, Conn. "There are so many different sources of
information that it's imperative not only to search that information, but
also to find the relevant information."

Incorporating search technology as an integral part of your overall Web
solution decreases production time for your customers by making information
management easier. Search technology alone, however, is not likely to
become your bread and butter. Still, it's a good add-on sale for either new
Web sites or existing ones, and it also has the potential to lead to
additional sales once you get your foot in the door.

Untangling the Web

Here's how a search engine traditionally works: First, the robot or spider
travels the Web, gathering documents found on sites and indexing the
information. The way the information is indexed depends on how the spider
is configured. For example, the spider may gather information about
documents by just indexing the HTML tag, while other spiders may index the
first paragraph or the entire text. Once a search request is made, the
search engine dives into a database and returns the results to the user.

There are some drawbacks to using spiders, however. The indexed information
is usually dumped into the database, often causing database overload.
Another problem is that for many companies, not all of the information that
they want to be searchable exists in HTML files, which is particularly true
of intranets. There's varying textual information, such as legacy data and
documents, as well as nontext data types, such as audio, video and images,
all of which need to be indexed.

Traditional search methods for finding the indexed data are becoming passe
as engines become more automated and expand beyond the usual keyword
search. "Since information is so overwhelming, you need to refine the
structured search syntax as much as possible," says Julia Young, sales
manager for OTEC, a VAR in New York. More advanced and structured search
language types include Boolean, which lets you use search operators such as
AND, OR and NOT; fuzzy, which will still recognize the word even if it's
spelled incorrectly; and proximity, which searches words with close
relationships to each other.

All of these criteria-the search syntax, indexing of the data, scalability,
varying data types, user-friendliness, how well the engine integrates with
your existing applications, performance reliability, and how queries are
returned- must be factored in when deciding which search engine to use. As
Yockelson puts it: "Choosing a search engine isn't always a question of
which is better, but a question of what the customer's needs are and how
the search engine can perform those needs."

The Search Is On

Luckily, the search technology market isn't as saturated with vendors as,
say, the Web authoring tool market. This makes finding the right tool to
fit your clients' needs a bit easier. Astea International Inc., a VAR in
Bedford, Mass., provides custom interaction help desk software and ports it
to the Web. According to CEO Avner Schneu, Astea embeds Fulcrum
Technologies Inc.'s search engine into its help desk applications because
it has a simple, native set of APIs. Also, Fulcrum's retrieval mechanisms
are written in a SQL dialect, so you don't need to learn different tools to
write the queries. Most search engines today also share a common RDBMS and
ODBC interface, which makes writing custom applications for those engines
easier and faster.

Another approach, typically used for fulfilling large system needs, is
full- text document management search capabilities. Full-text indexes every
word of the document automatically, acting as a structure for larger
archives. Excalibur Technologies Inc. supports full-text and provides a CGI
interface to a Web browser, which connects to a document imaging server on
the back end.

"Full-text is the most cost-effective way to index large amounts of data,"
says Brad Osborne, director of the imaging division at $151 million BTG, a
total solutions VAR in Vienna, Va. "The downside of using full-text is that
it doesn't have any agreed-upon interface, so you have to do some rewriting
for each additional application you create." BTG does document searching
and archiving for large litigation firms and the government; the search
engine is either used on internal networks or ported to the Web. Since
full-text indexes every word, you don't have to define any database fields.

Database Connectivity

More often than not, even when using full-text, a database on the back end
is also needed. For example, sometimes you may have fields that need to be
predefined, such as a company name, location, contact or date, and that
requires database knowledge. "Writing full-text isn't the answer," says
Osborne. "We make money providing the full application package."

When a database is involved, you'll have the opportunity to provide the
full application package because you'll need to write a custom interface
from the database to the search engine to connect the two elements. Most
vendors offer VARs toolkits and SDKs (software development kits), making it
easier to beef up database connectivity and customize search systems. For
example, using Excalibur's RetrievalWare SDK, VARs can ease database
connectivity through VisualBasic custom controls and RDBMS interface DLLs.

Adding database connectivity to internal or external Web sites also
increases security concerns, which could lead to another added sale. To
secure your database, you'll need to separate the sensitive data and public
data; one way is to keep the two database types on separate servers or to
install firewalls. If you're letting outside users gain access to your
database, then store the database internally, but push the pieces of the
database you want your users to see outside the firewall. This way, users
have read-only rights and can't make any changes to the existing database.

Once you've defined your customer's needs and installed the right search
system, add-on sales are bound to follow. Other possibilities for a total
Web solution include incorporating search functionality to groupware and
e-mail systems. But it all begins with the search.

-Quick Scan

AltaVista Internet Software/Digital Equipment Corp. Littleton, Mass. (800)
336- 7890, www.altavista.software.digital.com

Excalibur Technologies Inc. Vienna, Va. (703) 761-3700, www.excalib.com

Fulcrum Technologies Inc. Ottawa, Ontario (613) 238-1761, www.fulcrum.com

Innotech Multimedia Corp. North York, Ontario (416) 492-3838,
www.netresults- search.com

Copyright 1997 CMP Media Inc.

<<VAR Business -- 02-01-97, p. 63>>
---------------------------------------------------------------------------
[Feedback] [Index] [Search] [Customize] [Individual, Inc.]
[Main] [Up] [< Prev Story] [Next Story >] [Calendar]
---------------------------------------------------------------------------
[Individual, Inc.]
445.2AltaVista as front end...NETRIX::&quot;[email protected]&quot;Carlos BerlangaThu Feb 20 1997 20:3122
I have heard that there was a group internally trying to accomplish this and
they ran into some fairly 
basic problems. The major one was that AltaVista uses a primarily free form
text source from which it 
creates its own internal index scheme and database. The reply alluded to this.
The problem was how to 
integrate the raw data format that the database holds its information in and
create it in a way that 
AltaVista can understand. I do not believe that this is currently in the plans
for any additions to 
AltaVista, nor do I see it comming in the near future. As far as AltaVista is
concerned, the world is 
just a bunch of files that it opens up, looks into and retrieves information
from. That information is 
sorted, indexed and stored so that the more common items can be accessed
quickly and easily. 

So, in terms of using AltaVista as a traditional front end to a DBMS, I don't
think its currently doable. 

-Carlos-
[Posted by WWW Notes gateway]