Deliverables
Members
- OntoWeb Members
- Austin Tate, AIAI,
University of Edinburgh, UK (coordinator)
- Nicola Guarino,
LADSEB, Italy (for SIG on Content Standardisation)
- Milena Stefanova,
LADSEB, Italy (SIG monitor at LADSEB)
- Martin Doerr, Forth, Greece
(for links to WG on Cultural Repositories)
- Ellen Schulten, NOOS, France
(for links to WG on Product Classification Standards)
- Stuart Aitken, AIAI,
University of Edinburgh, UK (for links to public Cyc Ontology)
- Corresponding Members
- Martin Bryan, SGML Centre
(for links to eBTWG and ebXML)
- Michael Gruninger,
NIST, USA (for links to NIST PSL)
- Jerry Hobbs, or
David Martin
SRI International, USA (for links to DAML-S)
- Adam Pease,
USA (for links to SUMO)
- Howard Smith,
co-chair BPMI.org, CTO CSC Europe, co-founder Ontology.Org
(for links to BMPL)
- Ben Rode, Cycorp (for links to Cyc
Ontology)
State-of-the-Art
A study of 29 candidate models and representations which might act as
the basis of a process standard to meet a large number of identified
requirements was conducted as part of the NIST PSL project and
published in May 1998.
Knutilla, A., Schlenoff, Ray, C.S., Polyak, S.T., Tate, A., Cheah,
S.C. and Anderson, R.C. (1998) "Process Specification Language: An
Analysis of Existing Representations", NISTIR 6160, National Institute
of Standards and Technology, Gaithersburg, MD, USA.
See http://www.mel.nist.gov/msidlibrary/summary/9807.html
Contempory with the NIST PSL work, and accounting for it, a 1998 look
at possible contributory efforts for work on a shared plan, process
and activity representation was undertaken as part of the DARPA
project on SPAR in 1998. This article contains many of the references
and related URLs for work up to that time that contributed to
standards such as PIF, SPAR and NIST PSL.
Tate, A. (1998), "Roots of SPAR - Shared Planning and Activity
Representation", The Knowledge Engineering Review, Vol 13(1),
pp. 121-128, in Special Issue on "Putting Ontologies to Use"
(eds. Uschold. M. and Tate, A.), Cambridge University Press, March
1998.
See http://www.aiai.ed.ac.uk/project/spar/DOCS/spar-roots.html.
Since that date the main developments have been in the devlopment of a
number of XML encodings and industry "standards" for describing
processes, web services and other kinds of activities, processes and
task-achieving services on the web.
As background, a useful introduction to work on ontologies, and to the
role of processes within ontologies is provided by John Sowa:
A list of potential contributions to a standard process content model
for the web are in the following sections. The classification and
relevance of these is described in the section headed "Approach".
SIG members and others should mail Austin Tate to propose efforts not
mentioned here that may be relevant background for the work of the
SIG.
Approach
It is taken as a given that it is worth seeking to adopt shared
standards for the description and sharing of processes and
activity-related information on the web. It is also taken as a given
that this is a very dynamic area and that even with the best knowledge
available to guide the selection of a proposed process standard, any chosen
representation will only meet some requirements and may prove not to
be a successful and widely adopted standard that will last a long
time. But it is also taken as a given that well described web
information using such an agreed content standard will be understood at
any time in the future, and can be mapped as needed to emerging new
representations.
It may be feasible to adopt a core upper ontology which will have
longer life, but even this should be assumed to date and be replaced
in time. A model of distributing the mechanism for reaching and
communicating agreements on extensions so that not all aspects have to
be centrally agreed and maintained is taken as a requirement -
adopting the process Interchange Format (PIF) "Partial Shared View"
model (see http://ccs.mit.edu/pif/)
is a possible engineering approach.
A proposed classification of these efforts is made below, but this
also may be the subject of comment by SIG members for moving entries
between categories.
- Primary contributions on which a process content standard most
likely will be based.
- Secondary efforts that ought to be considered as requirements that
should be mappable to the chosen process content standard. These
typically cover some limited requirement (e.g. for exchange of simply
linked workflow processes) so do not adopt a very general process
model.
Initial List of Primary Contributing Efforts
- NIST Process Specification Language
(PSL - http://www.mel.nist.gov/psl/) -
contact: Michael Grunninger (NIST)
- The Process Specification Language (PSL) has been designed to
facilitate correct and complete exchange of process information among
manufacturing and business software systems. Included in these
applications are scheduling, process modeling, process planning,
production planning, simulation, project management, workflow, and
business process reengineering. PSL consists of an ontology for
process-related terminology together with a syntax for specifying
process descriptions.
- The PSL Ontology is organized into PSL-Core (which incorporates
the earlier work from the Process Interchange Format - PIF - project) and a
partially ordered set of extensions. All axioms and definitions for
the ontology are written in KIF (the Knowledge Interchange Format).
All theories within the PSL Ontology that are currently being
standardized have been proven to be sound and complete with respect to
the intended semantics of their terminology.
- NIST PSL was developed within a rigorous programme involving
identification of a wide range of requirements for many uses of
process representations, studying 29 potential candidate
representations, developing validation scenarios, and having extensive
reviews by a wide range of participants.
- PSL is a New Work Item (ISO 18629) within Joint Working Group 8 of
Sub-committee 4 Industrial data and Sub-committee 5 Manufacturing
integration of Technical committee ISO TC 184,Industrial automation
systems and integration . Part 1 of the standard has been accepted as
a Committee Draft.
- Business Process Modelling Language
(BPML - http://www.bpmi.org) -
contact: Howard Smith (CSC/BPMI) or Ismael Ghalimi (Intalio/BPMI)
- BPMI.org defines open specifications such as the Business Process
Modelling Language (BPML) and the Business Process Query Language
(BPQL) that will enable the standards-based management of business
processes with forthcoming Business Process Management Systems (BPMS),
in much the same way SQL enabled the standards-based management of
business data with off-the-shelf Database Management Systems
(DBMS). Whereas the enterprise wide management of data using DBMS is
based on the relational data model, the enterprise wide management of
processes using BPMS is based on process calculus.
- e-Business XML and the
e-Business Transition ad-hoc Working Group
(eBTWG - http://www.ebtwg.org) - contact: Martin Bryan (SGML Centre)
- More
details at http://www.ebtwg.org/projects/documentation/bioreference/.
- ebXML's goal is the standarisation of XML business specifications
and development of a technical framework which enables
interoperability, security and consistent use of business
information. This has been initiatied by UN/CEFACT and OASIS
(Organisation for the Advancement of Structured Information
Standards). The ebXML framework will represent a generic semantics of
concepts and business processes from different market sectors, while
allowing applications to use XML or EDI (http://www.edi.org) syntax.
- The eBTWG Business Proces Modelling Group are developing a Business
Process Information Exchange Schema (http://www.ebtwg.org/projects/bpimeschema.html)
and Business Process Specification Schema (http://www.ebtwg.org/projects/bpschema.html).
- These efforts are based on the Resources, economic Events and
economic Activity model (REA -
http://www.reamodel.org or http://www.msu.edu/user/mccarth4/rea-ontology/).
This ontology was developed in order to support Generalised Accounting
Systems and Shared Data Environments. It is also aimed at Supply Chain
Integration. Recently REA has been extended with the concept of
economic agreement which is of importance for electronic transactions
(in the context of ebXML).
- Shared
Planning and Activity Representation (SPAR -
http://www.aiai.ed.ac.uk/project/spar/)
- contact: Austin Tate (AIAI, Edinburgh)
- SPAR was developed by a group concerned with military plan representation
who also had experience of the development of initial standards
for plan, process and activity description. A large review team
was also involved. This work accounted for the experience gained on
the PIF, CPR and NIST PSL standards efforts and involved many of
the same people involved in these groups.
- SPAR deepened and extended the core upper ontologies of PIF and
NIST PSL and built upon the experience of applying CPR to a range of
military needs. It can be viewed as providing a second level of shared
model or ontology beyond what is offered by NIST PSL and is
compatible with NIST PSL.
- SPAR 0.1a is a very detailed model that also addresses engineering
issues of changing ontologies in future. SPAR 0.2 is a much simpler
"sentence" level description of the core model (essentially the same
as NIST PSL Core) along with a number of terminology and model
extensions grouped into packages based on their function. Note that
SPAR 0.2 is not a direct replacement for 0.1a and this earlier
document could still offer ideas for future ontologies underlying
process standards on the web.
- Core Plan
Representation (CPR - (http://projects.teknowledge.com/CPR2/)
- contact: Adam Pease
- CPR is a model that expresses
information common to many plan, process, and activity models. The
goal of this effort is to leverage common functionality and facilitate
the reuse and sharing of information between a variety of planning and
control systems. The CPR embodies a standard that is general enough to
cover a spectrum of domains from planning and process management to
workflow and activity models. The representation is powerful enough to
support complex, hierarchical plan structures. The initial application
of the CPR is in addressing plan interchange requirements of several
military planning systems, but the model goes beyond military planning
and presents a more general plan representation.
- IEEE Standard Upper Ontology
(SUO - http://suo.ieee.org) - contact: Adam Pease
- The IEEE SUO will specify an upper ontology that will enable computers
to utilize it for applications such as data interoperability,
information search and retrieval, automated inferencing, and natural
language processing. The Suggested Upper Merged Ontology available at
http://www.ontologyportal.org
is a "starter document" for this
effort which includes a formally specified process and activity model
along with hundreds of other supporting terms and thousands of axioms.
The process model merges PSL and other activity models and supporting
material.
- http://reliant.teknowledge.com/DAML/SUO.daml
- DAML encoding of proposed SUO
Initial List of Other Contributing Efforts
- Cyc (http://www.cyc.com)
- contact: Stuart Aitken (AIAI, Edinburgh) or Ben Rode (Cycorp)
- The Cyc Upper Ontology developed by Cycorp
is intended to be common ontology which can be
shared by more specific theories, or used in knowledge-based applications.
The Upper Ontology includes a Time Vocabulary which provides the
central concept of Event, plus simple and complex temporal relations.
Additional vocabulary for composite events (processes) is being developed under
the DARPA Rapid Knowledge Formation (RKF) program and may be included in
OpenCyc (http://www.opencyc.org).
- Planning Domain
Description Language (PDDL - http://cs-www.cs.yale.edu/homes/dvm/)
- PDDL was developed by the AI Planning Systems (AIPS) Competition Committee
for use in defining planning problem domains. It provides a widely-used
syntax for expressing STRIPS-like planning operators and
other elements of the domain, plus the start and goal states of
planning problems.
- <I-N-OVA> -
Constraint Representation of Activity
(http://www.aiai.ed.ac.uk/project/ix/inca.html)
- contact Austin Tate (AIAI, Edinburgh)
- <I-N-OVA> is a simple model for activity specification which
is based entirely on the representation of constraints on the whole
space of legitimate behaviour in an application domain. It is
therefore very extendable and can be used as an underlying and
unifying representation to underpin more elaborate or specialised
representations. It was designed to be able to deal with
representations needs from the very simple to the most complex.
- <I-N-OVA> had the broadest coverage of the requirements in a
study of 29 possible contributing representations for the NIST PSL
work. The areas not covered were beyond the design remit for
<I-N-OVA> and dealt with activity-relatable-object models or
agent-to-agent relationship models - things meant to be orthogonal
to <I-N-OVA>.
- <I-N-OVA> is itself based on the <I-N-CA> constraint
ontology for any synthesised artifact or product. It is an ontology
which can underpin desciptions of designs, configurations, plans, etc.
A process description or plan is treated as a specialised type of design.
- The aim of the work is that PIF, NIST PSL, CPR and SPAR all
converge on a single core model of activity based on <I-N-CA>
and that this form a strong extensible underlying model for web
standards for activity, process, plan, task, capabillity and other
related descriptions.
- DAML for Services (DAML-S -
http://www.daml.org/services/)
- contact: Jerry Hobbs or David Martin (SRI International)
- DAML-S is an ontology for describing Web services,
currently under development by a
collection of researchers funded by the DAML (Darpa Agent Markup
Language) program. The ontology is layered atop the DAML+OIL
(http://www.daml.org/language/)
ontology language.
- The aim is to supply Web service providers with a core set of markup
language constructs for describing the properties, capabilities, and
behavior of their Web services in unambiguous, computer-intepretable
form.
- DAML-S is designed to facilitate the automation of Web service tasks
including discovery, execution, composition and interoperation.
- In addition to the specification of processes, DAML-S provides for the
declaration of "profiles" (advertisements) and groundings" (bindings to
specific communication protocols) of Web services.
- Web Services Description
Language (WSDL - http://www.w3.org/TR/wsdl) and
Web
Services Flow Language (WSFL - http://www-106.ibm.com/developerworks/library/ws-ref4/)
- Metadata Coalition Business Engineering Model (MDC-BEM)
- Workflow Management Coalition
(http://www.wfmc.org) Workflow Reference Model and Workflow
Process Description Language (WPDL)
- OMG has several working groups related to manufacturing and
e-commerce. In particular, see the Electronic Commerce Reference Model,
the
UML specification
(http://www.omg.org/cgi-bin/doc?ad/99-06-08) and the
List
of Working Groups (Domain Technology Committees)
(http://www.omg.org/news/about/omg_technology_plenary.htm)
- OAG and RosettaNet
(http://www.openapplications.org/)
specify protocols for B2B applications and agents, and these have
similarities to service ontologies.
- CIDOC CRM (http://cidoc.ics.forth.gr) - contact Martin Doerr (Forth, Greece)
- The CIDOC CRM is a high level ontology for describing cultural
heritage. Central to it is the event-based description of history in
a physical sense. As such it contains several rather general temporal
concepts, concepts of objects and actors. Process descriptions about
past events should be comparable to those. It contains further basic
concepts of how places and objects relate, in particular places on
mobile objects. Further the notion of time is analyzed as indeteminate
intervals, determinate intervals and duration. The CIDOC CRM version
3.2 is going to be proposed as a Committee Draft by ISO/TC46/SC4 in
October 2001.
SIG Workplan
For the OntoWeb SIG activities, there were three goals (achieved by
10-Dec-2001):
- Deliver a "state-of-the-art report" with concise descriptions of
the most relevant standards.
- Decide the standards cluster on which we will focus our ontology-based
harmonisation effort. This means we should select a FEW standards that:
- Are suitable for ontology-based harmonisation
- Are of high interest for the European Community
- Show a concrete interest (and allocation of resources) from the
standardisation bodies
- Involve at least 2-3 OntoWeb members willing to invest resources on
their own funds.
- To isolate the main foundational issues we want to discuss with the
other OntoWeb WGs.
Recommendation on Focus for Future Study within OntoWeb
NIST PSL, BPML and eBTWG's business process schema are suggested as
the bases to concentrate on as they all have standards activity behind
them. These represent a range of interests and alternative
requirements to generalise the approach. The work on CPR 2 and SPAR
are also relevant, and indeed some of this provided background used in
NIST PSL, but since these activities are not part of a standardisation
process they are not considered as relevant for use directly within
OntoWeb. All these 5 bases have OntoWeb membership and other
corresponding member interests. IEEE SUO is of interest as it is
being actively developed and incorporates the results of the NIST PSL
and CPR efforts.
Credentials
The working group members represent some of the most experienced
people worldwide who have been involved in creating shared plan,
process and activity representations or standards for some
years. Several of them are actively leading efforts to develop or
refine standards at international level.
Austin Tate, the coordinator, has the following background:
- A researcher with 30 years of experience in AI planning techniques
and plan representations.
- Fellow of the Workflow Management Coalition and a former member of
it's WG1 which developed WPDL.
- Active in the development of the US Core Plan Representation (CPR).
- A member of the Process Interchange Format (PIF) team.
- An active member of the team which developed the NIST PSL Core.
- Leader of the group which developed DARPA SPAR.
- A Principal Investigator on the EPSRC-funded
Advanced Knowledge Technologies (AKT)
project - which partially supported the work in this WG.
Page maintained by Austin Tate
(a.tate@ed.ac.uk)
Updated: Tue Sep 7 18:27:47 2004