Browse Prior Art Database

Virtual URLs for Browsing and Searching Large Information Spaces

IP.com Disclosure Number: IPCOM000123297D
Original Publication Date: 1998-Sep-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 4 page(s) / 177K

Publishing Venue

IBM

Related People

Elo, S: AUTHOR [+2]

Abstract

Motivations As an information space grows, it becomes more difficult to find the information you need. For example, the IBM Intranet is a complex information space containing over a million web pages hosted on more than three thousand servers across the world. Semantically similar pages, or ones relating to the same subject, are often stored on different servers and not linked to one another. For example, even if an employee finds a good starting point to download software, the employee probably misses many other sites that host downloadable software of interest.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 35% of the total text.

Virtual URLs for Browsing and Searching Large Information Spaces

   Motivations

   As an information space grows, it becomes more difficult to
find the information you need.  For example, the IBM Intranet is a
complex information space containing over a million web pages hosted
on more than three thousand servers across the world.  Semantically
similar pages, or ones relating to the same subject, are often stored
on different servers and not linked to one another.  For example,
even if an employee finds a good starting point to download software,
the employee probably misses many other sites that host downloadable
software of interest.

   Our solution is to create a virtual service.  In a virtual
service, a heterogeneous collection is structured into a single,
logical information space.  The service presents an integrated
interface that minimizes the physical and cognitive load of
switching between browsing and searching.  A virtual service begins
with webmasters at different sites classifying pages within a single
hierarchical organization, or taxonomy.  While the physical location
of pages is unchanged, pages on the same subject get grouped together
under the taxonomy.  The new organization masks the existing
inconsistent names of hosts and directories.  Each page now has a
logical address, or VRL, that reflects its position within the
taxonomy.  A more stable information space is created because a
taxonomy does not change as rapidly as individual pages classified
within it.

   As users of the virtual service navigate the new information
space, the VRL Mapper handles their requests.  The VRL Mapper never
lets a user-entered VRL fail.  Instead of giving the user an error
message for an invalid URL, the Mapper tries hard to find nodes or
pages based on the VRL requested.  In practice, regardless of the
requested VRL, the system always returns at least some options on
what else a user can try.  The system also explains how it came up
with the options.  The inference rules and explanations of the VRL
Mapper improve upon standard web server error handling.

   Virtual Resource Locators

   A Virtual Resource Locator is the logical address of a
piece of information.  A VRL has the same syntax as a URL.  While a
URL describes a physical location on a server on the net, a VRL
describes a logical location within an information taxonomy.  A VRL
address is more meaningful and easier to understand.  Because items
can be classified under more than one taxonomy category, a page may
have several VRLs.  This allows users to access the same information
from a number of different perspectives.

   The process of classifying a page into a category in a
taxonomy creates metadata, or information about the subject matter
of the page.  Usually, metadata is stored either in a metadata
server or as special tags in the page, like the XML syntax.  Because
the VRL Mapper needs fast access to the taxonomy and the metadata
associated wit...