Page 136 -

4 Dynamic Workflow 125
An integral service distributed as part of the YAWL environment that provides
for dynamic flexibility and exception-handling support for YAWL processes is the
Worklet Service. The remainder of this chapter will discuss the unique conceptual
design of the Worklet Service and how it supports dynamic workflow. The way the
Worklet Service handles runtime exceptions is described in the next chapter. The
service implementation (with examples) is detailed in Chap. 11. In addition, Chap. 6
describes Declare, an approach to workflow flexibility using constraints.

4.3 Worklets: Theoretical Basis

Whenever a series of actions is undertaken with a view of achieving a preconceived
result, some plan or set of principles is implemented that guide and shape those
actions towards that goal. To be effective, a plan must be described using constructs
and language that are relevant to both the actions being performed and the desired
result, and be comprehensible by its participants and stakeholders. In workflow
terms, analysts seek to model some aspect of the real world by using a metaphor
that bears some resemblance to the real world, but also represents an understanding
of computational processes. Such metaphors are abstract constructions that form a
common reference model, which assist us in representing the external world through
computers.
The fundamental and widely understood computational metaphor [242] takes
a set of inputs, performs a series of functional steps in a strict sequence, and, on
completion, produces some output that represents the goal of the process. Thus
the computational metaphor describes a single, centralized thread of control, which
very much reflects its mathematical ancestry, and reveals the influence of pioneers
such as von Neumann and his team, and especially Turing, whose abstract machine
proposed “step-at-a-time” processing, and which in turn reflects the influence on
thinking of the contemporaneous development of assembly-line manufacturing.
As the prevailing technological advances influenced the structure of early com-
puters, so too has the computational metaphor become a significant model system
for the conceptualization and interpretation of complex phenomena, from cognition
to economics to ecology. Of particular interest is the way the metaphor has been
applied to the definition of organizational behavior issues and the representation
of organizational work processes. The computational metaphor remains applicable
to well-defined problem domains where goal-directed, sequential, endpoint-driven
planning is required. Such domains were the early beneficiaries of workflow man-
agement systems. Consequently, workflow systems typically provide support for
standardized, repetitive activities that do not vary between execution instances.
Adherence to the metaphor by workflow systems has been an important factor
in their acceptance by organizations with structured work practices. Descriptions
can be found throughout the workflow literature to the “processing,” “manufactur-
ing,” and “assembly-line” modeling metaphors that are employed by commercial
workflow systems. However, while the Workflow Management Coalition claims
that “even office procedures can be processed in an assembly line” [271], there are