03.12.2012 Views

xoEPC - Jan Mendling

xoEPC - Jan Mendling

xoEPC - Jan Mendling

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

2 1. Introduction<br />

Dav93, JB96, Sch98a, ACD + 99, Sch00, LR00, ADO00, AH02, MCH03, BKR03, Kha04,<br />

LM04, Hav05, SF06, Sta06, Jes06, SCJ06, WLPS06, KRM06, Smi07]) as well as in-<br />

ternational professional and academic conference series, such as the BPM conference<br />

[AHW03, DPW04, ABCC05, DFS06], confirm the importance of BPM. Despite the<br />

overall recognition of its importance, several fundamental problems remain unsolved by<br />

current approaches.<br />

A particular problem in this context is the lack of research regarding what is to be<br />

considered good design. The few contributions in this area reveal an incomplete under-<br />

standing of quality aspects in this regard. Business process modeling as a sub-discipline<br />

of BPM faces a particular problem. Often, modelers who have little background in formal<br />

methods, design models without understanding the full implications of their specification<br />

(see e.g. [PH07]). As a consequence, process models designed on a business level can<br />

hardly be reused on an execution level since they often suffer from formal errors, such<br />

as deadlocks. 1 Since the costs of errors increase exponentially over the development life<br />

cycle [Moo05], it is of paramount importance to discover errors as early as possible. A<br />

large amount of work has been conducted to try to cure the symptoms of this weak un-<br />

derstanding by providing formal verification techniques, simulation tools, and animation<br />

concepts. Still, several of these approaches cannot be applied since the business process<br />

modeling language in use is not specified appropriately. Furthermore, this stream of re-<br />

search does not get to the root of the problem. As long as we do not understand why<br />

people introduce errors in a process model, we will hardly be able to improve the design<br />

process. There is some evidence on error rates for one particular collection of business<br />

process models from practice [MMN + 06b, MMN + 06c]. 2 We will take this research as a<br />

starting point to contribute to a deeper understanding of errors in business process mod-<br />

els.<br />

1 In the subsequent chapters, we will distinguish between two major types of errors. Firstly, formal<br />

errors can be identified algorithmically with verification techniques. Secondly, inconsistencies between the<br />

real-world business process and the process model can only be detected by talking to stakeholders. The<br />

focus of this thesis will be on formal errors.<br />

2 Classroom experiences are reported, for example, in [MSBS02, Car06].

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!