28-09-2012, 10:51 AM
Comparison of Tools and Languages for Business Process Reengineering
1Comparison of Tools.pdf (Size: 127.07 KB / Downloads: 85)
Abstract
The paper presents the comparison of main Business Process Reengineering (BPR) tools from the point of view of modeling languages supported by them. One of the tools considered is the GRADE tool developed by IMCS LU. The proposed comparison criteria are language support for the selected basic modeling activities common to most BPR methodologies. The main emphasis is on the semantic properties of modeling languages. The results of comparison are summarised in a table.
Introduction
The term Business Process Reengineering (BPR) has been actively used for at least 4 years, since the publishing of the book by Hammer and Champi [1]. Nearly every company has declared to do some BPR activities recently. This has created a booming market for various BPR support tools - according to the estimates by Gartner Group [2] - about 100 million USD per year, with the annual rate of growth at 30%. There are about 40 BPR tools on the market.
Some industry-level evaluations of BPR tools are performed regularly [2-8], the most respected of which are the regular annual evaluations, performed by Gartner Group [2,3,4,5]. The main goal of Gartner reports is to serve as a tool buyers’ guide of high quality to the US market, therefore the main emphasis there is on the tool-technical criteria. Only the tools highly used in USA are included in Gartner reports. The goal of the vast BPR tool survey by Bach, Brecht et al [6] is also an industrial evaluation, however a deeper analysis of modeling facilities, including metamodels, is provided.
The goal of this paper is to compare the underlying principles of modeling incorporated in tools. These modeling principles of a tool determine its modeling language and so our main emphasis is to compare modeling languages for BPR. The authors of this paper are within the team developing the BPR tool GRADE [9]. Therefore one of the goals of the paper is to compare GRADE to other BPR tools from the modeling language point of view.
Comparison principles
The considered area of BPR
The business process or more precisely the business system is understood in this paper in the broadest possible way. BPR is considered as an activity corresponding to its classical definition given by Hammer and Champi[1], namely “fundamental rethinking and redesign of business processes to achieve dramatic improvements in critical contemporary measures of performance such as cost, quality, service and speed.” There may be a lot of ways to achieve this goal or, in other words, different BPR methodologies (see e.g. [10-14]). Except for some very specific ones [15],
Tools and languages selected for comparison.
As it was already mentioned, this comparison is made by some of the authors of the GRADE tool [9], being developed by joint efforts of IMCS LU and Infologistik GmbH (with earlier participation of the company “Dati”). Therefore the comparison inevitably is “GRADE to other tools”. The other tools are taken from the top-ranking ones according to the Gartner reports [2,3]. Since we analyse tools from the modeling language point of view, tools supporting the same language are grouped together. The following tools/languages have been selected.
Short description of modeling languages used by tools
We have to give at least a very short description of the selected modeling languages to be able during comparison refer to the main concepts of each language. Since the main area of comparison is the dynamic behaviour representation most attention is devoted to this aspect. Fig.1- fig.4 show the brief examples of all languages describing the same fragment of a business system - an initial fragment of order processing in a company.
GRADE and GRAPES-BM language
A sufficiently comprehensive description of GRAPES-BM business modeling language supported by the GRADE tool (namely, its version 3.0) has been given in the previous DB&IS conference [9]. From this version the main application area of GRADE has been meant namely the BPR. Now the GRADE version 4.0 has been released. The main business modeling language features in GRADE 4.0 are the same as in GRADE 3.0, though the tool quality, especially its user interfaces, have improved significantly. The main new language feature in this version is the support of Object Modeling Techniques with Class diagrams either in the classical OMT form [16] or in the new UML 1.0 [17] form. Class diagrams, according to almost all modeling methodologies, should be used for capturing the first overall view of a business system. GRADE 4.0, to a certain degree, has integrated Class diagrams with other modeling aspects. Class diagrams can be “reverse-engineered” from ER models, relevant parts of Class diagrams may be used for generating organisational structure (ORG) or Data Type Definition (DD) diagrams.
ARIS modeling language
The ARIS tool by IDS Prof. Scheer has its own modeling language, which has no specific name. The current evaluation is based on ARIS version 3.1a and fragments of the upcoming version (ARIS Easy Design 1.0).
The modeling language contains about 50 diagram types. Each of the diagram types has a list of permitted objects, many of which are common to several diagram types.. The current number of object types is about 110 and is expected to grow, but they are separated into groups with common properties. Each diagram type defines also permitted links between object types. A link may have types defining its intended semantics and attributes. Any object has a fixed list of attributes (e.g. function has about 140 attributes), part of which are common to all objects and part - object-specific. Among the common ones there is a fixed number of user definable attributes.
Conclusions
The comparison of modeling languages for the main BPR has revealed many similarities, but also several different approaches to BPR language design. Yet the main conclusion may be a couple of issues having not found a good solution in any of languages:
• though OMT (now UML) techniques is supported by nearly all tools, no language has offered natural links between classic Business modeling and Object modeling
• higher level business goals have not got a real semantic link with other parts of modeling