About the Workflow Patterns

The Workflow Patterns Initiative started at the 4th IFCIS International Conference on Cooperative Information Systems (CoopIS) in Edinburgh in 1999 when Wil van der Aalst and Arthur ter Hofstede started working on workflow patterns. This resulted in a publication at the same conference one year later (CoopIS, Eilat, Israel, 2000) and a technical report both also co-authored by Bartek Kiepuszewski and Alistair Barros.

Initially, the work received little attention outside academia. Although, several workflow products were evaluated using these patterns - the first systematic evaluation of workflow management systems and their capabilities - vendors showed little interest in the research and its results. This changed at the beginning of 2001 when the Workflow Patterns website was launched, which served as a portal for the dissemination of research results (and later vendors responses) relating to the workflow patterns. This event triggered efforts by both researchers and practitioners to apply the patterns in product development and product evaluations. Moreover, several reports on the patterns appeared in Dutch IT journals around the end of 2001 and beginning of 2002. As a result, many Dutch organizations started to use the workflow patterns in selection processes. Ultimately it was this outcome that piqued the interest of workflow vendors.

The website has continued to be extremely successful as indicated by the number of visits it receives on an ongoing basis. As of July 2006, it had served more than 240,000 pages and the site services about 200 page views on a typical working day. This is quite remarkable for such a specialised site.

Further consideration of the workflow patterns and investigations into the various ways in which contemporary products implement these patterns served as the catalyst for research into the fundamentals of workflow technology. This centred on an examination of the expressiveness of the various solutions for control-flow in workflow products and resulted in the PhD work of Bartek Kiepuszewski (available here).

At the end of 2002, the scope of the Workflow Patterns website was expanded further. By this time, concepts and technology originating from the classic workflow management domain as described by the WfMC had been absorbed by other disciplines and systems. Indeed it is now recognised that the majority of today's information systems are "process aware" and therefore incorporate some form of explicit process modeling and enactment. Consequently the results of the patterns research are applicable to a much broader range of fields including ERP-systems, E-business systems, call centre software, case handling tools, etc. In particular in the domain of web services, there are a lot of initiatives to deal with (inter-organizational) processes in a structured way and building on workflow concepts and technology. In 2002-2003 Wil van der Aalst, Arthur ter Hofstede, Marlon Dumas and Petia Wohed evaluated several standards for web services composition (BPEL4WS, BPML, XLANG, WSFL and WSCI).

In 2004, Nick Russell joined the Workflow Patterns Initiative and commenced work on broadening the focus of the patterns to provide a more holistic coverage of the various perspectives relevant to process enactment. The first result of this expansion was the publication of the Workflow Data Patterns which categorised the ways in which data is represented and utilised in processes-aware information systems. This was shortly followed by the Workflow Resource Patterns which provide a detailed exploration of resource management and work distribution in PAIS.

Based on the complete set of patterns (now over 100 patterns in the control-flow, data and resource perspectives), comprehensive evaluations of UML 2.0 Activity Diagrams and BPMN (Business Process Modeling Notation) were undertaken.

In 2005, the area of exception handling was investigated from a patterns perspective leading to the definition of series of Workflow Exception Patterns which characterise the ability of workflow systems to cater for various types of exceptions which may occur in the control-flow, data or resource perspectives. As part of this work, a general graphical language was developed for describing exception handling strategies in a technology independent manner.

In 2006 the original set of 20 control-flow patterns was revisited with a view to providing a formal description of each of the patterns and also to identify any potential gaps in the original set of patterns. This review resulted in the control-flow patterns being completely revised and turned into a set of 43 patterns containing more detailed and formalised descriptions and explicit evaluation criteria. Animations are now provided for all the patterns (except exception handling).

Based on the original collection of control-flow patterns and the evaluations of systems and standards, Wil van der Aalst and Arthur ter Hofstede started the development of YAWL (an acronym for Yet Another Workflow Language) in 2002. It was originally intended to provide a means of definining individual patterns and showing how they could be supported. YAWL can also be used as a starting point for workflow analysis and exchange. Compared to other languages such as the WfMC's XPDL, YAWL is more expressive and has clear and unambiguous semantics. In fact, we believe YAWL to be the world's most powerful language for process modelling.

In 2003, work commenced on an open source reference implementation of the YAWL language. This system provides a complete functional workflow system that is based on the workflow patterns. Over the past three years, a series of releases have been made available to the public. More recently, these releases have involved significant collaboration with industry partners.

The workflow patterns have had a significant impact on the field of workflow technology. Since their initial publication, they have been used widely by academics, practitioners and vendors for a variety of purposes including selection and evaluation of workflow solutions, design enhancement of commercial and open-source offerings, skills transfer and education. The impact section provides a comprehensive overview of the effects that this research initiative has had over the past seven years.