Information Technology Reference
In-Depth Information
elements in the PST contains many arcs. On the other hand, if both elements are
located into the same control block without additional nesting, they will also be
in the same region of the PST, i.e. there are exactly two arcs in the PST between
the elements. The assumption that the PST-distance can be an indicator of the
diculty to reason about a relation between two model elements is in line with
the conceptual model of cognitive complexity by Cant et al. [23] that has been
developed with respect to understanding software. Cant et al. discuss nesting
within a a piece of software and argue that “the number of 'steps' [groups of
control-flow statements; note from the authors] involved indicates the number
of chunks which need to be considered” [23].
Formally, we define the
PST-distance
between two elements
A
and
B
of a BPM
as the number of arcs between
A
and
B
in the PST minus one. This means that
elements in a sequence or in the same control block have a PST-distance of 1.
For example, in Fig. 1 the activities 17 and 18 which are executed in parallel
inside the same control block have a PST-distance of 1 while the activities 16
and 17 (the latter is inside the fragments M and N) have a PST-distance of 3.
Element Separateness: Cut-Vertices.
A second aspect we take into account
when discussing the interactivity between elements
A
and
B
in a BPM is the
special case where a single arc in the BPM separates the BPM into two disjoint
parts
P
1
and
P
2
such that
A
P
2
.
In terms of graph theory this means that the connected graph
G
that forms
the BPM has a so-called cut-vertex on a path from
A
to
B
, i.e. a vertex that
when removed causes that the remaining graph is not connected anymore. If
such a cut-vertex between
A
and
B
exists, the mental model of the relationships
between
A
and
B
becomes much easier, because
A
is located “before” and
B
is located “after” an easy-to-spot reference point (the cut-vertex). For example,
in Fig. 1 it is easy to see that activity 7 cannot be executed after activity 17.
Because of the cut-vertices before and after activity 16, this can be concluded
without analyzing the control structures in which the activities 7 and 17 are
embedded. The assumption that the presence of a cut-vertex makes it easier to
understand a model is backed by results by Mendling and Strembeck [2] who
found that a large number of cut-vertices in a model has a positive effect on its
understandability.
∈
P
1
and
B
∈
3 Research Model
Having laid out the relevant theoretical factors related to local understandability
of process models, we will now draw several propositions to suggest how these
factors will influence cognitive diculty in comprehension tasks. Prior research
on process model comprehension has almost exclusively focused on global model
understanding, a focus of study that we extend in this paper by looking at the
understandability of relations between elements in a process model.
Fig. 3 shows our research model. The model proposes that the cognitive dif-
ficulty of understanding the relation between model elements is influenced by
