Browse Prior Art Database

Parci Architecture

IP.com Disclosure Number: IPCOM000115012D
Original Publication Date: 1995-Mar-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 6 page(s) / 361K

Publishing Venue

IBM

Related People

Bejar, A: AUTHOR

Abstract

Proposed is an architecture which supports flexible reuse of varied code and data in a single system. It combines a program to parse and search requests; keeps a list of keywords associated with information sources and employs a program to route the search request to the relevant information source; solves the problem of requesting information from several sources without knowledge of their contents or source usages; utilizes the best of existing code.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 17% of the total text.

Parci Architecture

      Proposed is an architecture which supports flexible reuse of
varied code and data in a single system.  It combines a program to
parse and search requests; keeps a list of keywords associated with
information sources and employs a program to route the search request
to the relevant information source; solves the problem of requesting
information from several sources without knowledge of their contents
or source usages; utilizes the best of existing code.

      As systems evolve the amount of information available to the
user grows rapidly different forms and the ways of accessing the
information vary as much as information itself.  Most of these
different information sources are self contained, and so if something
is needed from more than one, even on a common topic one either has
to go into each one and access it, or build a tool with direct access
capabilities between them generating a dependency.  Since information
sources have different ways of accessing the information contained,
tools for going through it can be difficult and long to try and take
the information out for display somewhere else, losing already
developed tools that might be better for looking and navigating that
information source.  Say it is wanted to make a front end program
that brings together the information in electronic books, a knowledge
base and a Host database; the problem looks like Fig. 1.

      The computer is required to take control of where the
information lies, and in this case the front end that is needed has
to know which information source to access, how to do it, how to
extract or display the results.  That takes long to develop.  It can
be done directly, writing device drivers for each program that
depends on the front end, and letting the front end know what lies in
which information source.  That dependency makes for a front end that
is not flexible and were the information sources to change either
form or content it would be difficult to maintain and expand without
rewriting.  It is not flexible and only a short term solution.  Such
problems originate in the dependence between the front end and where
the information lies which is important because as computers evolve
the information is constantly updated, and new sources added so any
solution that tries to cope needs to be able to change and adapt as
the information sources change, and evolve with the user's needs.

Consider the following propositions:
  o  Information can be indexed, by keywords.  That index can be
      reduced to the keyword and where it is to be known as the
      information source.
  o  The important distinction between information sources lies in
      their contents, which can be differentiated by their list of
      keywords.  If this lies in more than one place then it should
      be possible to look in them all.
  o  A program that would generate a list of keywords and where they
     ...