Information Technology Reference
In-Depth Information
= AND
= exclusive OR
9
10
I
F
M
20
8
2
11
14
19
7
H
G
1
3
D
13
15
16
18
21
4
6
12
17
N
A
K
L
5
E
C
B
Fig. 1.
Business process model, structured into regions
Fig. 2.
PSTforthemodelshowninFig.1
In order to define a measure for the cognitive load resulting from the effort
to understand the relation between two elements in a BPM, we follow the idea
of Vanhatalo et al. [22] to decompose the BPM into canonical fragments with
a single entry and a single exit. These fragments can be arranged in a process-
structure tree (PST) such that there is exactly one PST for each BPM. For
details we refer to [22], but we introduce the concept of a PST by an example.
Fig. 1 shows a BPM (similar to the ones used in our experiment) and its canonical
fragments that form the PST. Additionally to the fragments that are marked
with dotted boxes, all single activities and the model as a whole are canonical
fragments in the PST. From the example, it can be seen that canonical fragments
can be nested. For example, the fragments D and E are within a larger fragment
C. The depth of the nesting shows how many routing constructs in the BPM
have to be understood in order to reason about the execution of an activity. The
PST of the model is shown in Fig. 2. For a better readability, the control nodes
(called gateways in BPMN) are omitted in this graph.
We argue that the distance between two elements in the PST can serve as a
measure for the interactivity between those elements. Each region in the PST
represents one concept (for example the concept of an exclusive choice or the
concept of parallel branching) that the reader of the model has to understand.
If elements are located in deeply nested control-flow blocks, the reader has to
understand a large number of concepts before being able to answer a question
on the relation between those elements. In this case, the path between the two
