Ontolingua/Kif experiences

Mike Uschold <mfu@aiai.edinburgh.ac.uk>
From: Mike Uschold <mfu@aiai.edinburgh.ac.uk>
Date: Wed, 27 Apr 94 08:59:07 BST
Message-id: <4189.9404270759@subnode.aiai.ed.ac.uk>
To: ontolingua@HPP.Stanford.EDU
Subject: Ontolingua/Kif experiences


We are involved in a large project one of whose main aims is to produce
a software toolkit for enterprise modelling.  There will be strong emphasis on
business process modelling, but we will also support modelling of other
aspects of enterprises. (See summary of project at end of message).

The architecture is to be open, enabling a variety of tools and
languages to be integrated.  One way to achieve this is to have a shared
`enterprise ontology' into which and from which representations from other
tools could be translated/converted as required.

We are looking around for tools and technologies that may assist us in
this, and the only software we have found that is both available and
potentially applicable is Ontolingua/KIF.

Can anyone out there with knowledge and/or experience of Ontolingua and KIF
please comment on the likelihood of our being able to use Ontolingua 
at the heart of our toolkit?

We are especially interested in responses from any persons involved in
applications of Ontolingua large or small and/or writing ontologies in KIF.

Particular issues of concern include:

* Do the services provided by Ontolingua match our needs or will we have
to use something else.  

* If the latter, is there anything else that we might be able to use, or
will be need to develop something from scratch?

* Is Ontolingua sufficiently well documented that one can learn what one
needs to in a reasonable amount of time, or is there a large amount of
scrounging around that one must do to find/get what one needs?

* What are the main strengths of Ontolingua relevant to our project?

* What are main weaknesses of Ontolingua relevant to our project?

* Suitability of KIF (+ Frame Ontology) as the language for expressing
	- expressive power
	- speed of learning KIF
	- ease of use once familiar with it
	- etc

* The enterprise domain is very large, and we cannot start from scratch
building ontologies for all the sub-areas includeing resources, business
processes, time, etc.  We hope to select existing ontologies and
merge/integrate them into a whole.  

Does anyone have any experience with doing this sort of thing on small,
medium, or large scale?

We are also considering using the KADS Domain modelling language for
representing the ontology.  We welcome any comments/experiences anyone
may have in using this language and associated support software ad how
it may compare to Ontolingua/Kif.

Thank you
Mike Uschold,   

AI Applications Institute,     (AIAI)               
The University of Edinburgh,
80 South Bridge, Edinburgh EH1 1HN 

                  The `Enterprise  Project'

             Technical project manager: John Fraser

Project Partners

    The project partners are AIAI (lead partner), IBM, Lloyds'
Register, Logica and Unilever.

Project Duration

    The project is due to be completed in autumn 1996.

Project Aims

    The broad aim of Enterprise is to develop a software toolset for
modelling various aspects of enterprises.  The toolset will support a
generic core of practical knowledge based modelling and simulation tools
and methods for business application.
    Significant input will be provided by Unilever, Lloyd's Register and
IBM, who will set user requirements and evalu- ate the toolset in the
context of their selected applications.


    The toolset will be particularly beneficial to middle man- agers at
a tactical level as part of an on-line operational mod- elling, planning
and simulation system. Support will be avail- able for the specification
of information systems solutions for parts of the business
re-engineering process: it is hoped that such specification can be made,
when appropriate, in a form amenable to existing workflow management
systems. Aids to tracking, evaluating and modifying the implementation
of the chosen solution will also be provided.


    An important focal point for the project will be the assess- ment,
specification and implementation of interfaces between the Enterprise
toolset and existing decision support and mod- elling tools (e.g.
mathematical simulation systems). The tools and methods that we aim to
integrate and enhance will be based on a number of key standard methods
and recently de- veloped technologies, including object-oriented
requirements capture, design and development, knowledge-based plan rep-
resentation, multi-agent problem-solving and open architec- tures.


    The project will also pay particular attention to the in-
tegrability of its deliverables with established business mod- elling
approaches (two targets being IBM's BSDM and the internationally used
IDEF methods) and current management support systems (e.g. RDBMS,
mathematical modelling tools).

Public Awareness

    An important aspect of the project is specific requirement
to increase public awareness. There are three key components:

1. a broad survey of the state of the art in enterprise mod-
      elling techniques, tools and methods;

2. the development of two demonstrators using two third         
          party software tools;				        

3. information dissemination activities, aimed at building
      up interest in the user community.

Technology Areas

    These  are  the  main  technology  areas  which  will  be  ad-
dressed during the project:

    o knowledge representation and domain ontologies
    o knowledge interchange
    o process  modelling  and  application  of  plan  representa-
    o requirements capture
    o organisational modelling
    o strategic planning
    o distributed AI systems