Evaluation of Systems Analysis Methodologies in a Workflow Context

Fahad Al-Humaidan B. Nick Rossiter

Computing Science Computing and Mathematics

Newcastle University Northumbria University

Newcastle upon Tyne Newcastle upon Tyne

NE1 7RU, UK NE1 8ST, UK

Tele: ++44 191 227 4662

nick.rossiter@unn.ac.uk

Abstract

Complex information systems require a methodology for their development in a structured manner. Many different methodologies exist, each suitable for a particular type of application. The capabilities of a number of methodologies are expressed in tabular form relative to a taxonomy developed for workflow systems and to a more general taxonomy dealing with both hard- and soft-system aspects. The results show that there is no methodology that covers all of the taxonomic aspects identified.

Keywords: methodology, systems analysis, workflow, taxonomy, soft system.

1. Introduction

It is generally accepted that complex information systems require a methodolgy to take their development forward from the initial requirements of users to an implemented documented functioning system, which satisfies the end-users in its functionality and interface (Beynon-Davies, 1998). Many kinds of methodologies have been proposed since the 1970s in the area of systems analysis (Hawryszkiewycz, 1998). A glossary for some of the commonly-used acronyms is given at the end of this report. The large number appears to result from different procedures in the various software development houses and the varying appropriateness of paradigms from application to application. It is probably true to say that every methodology has a target area of application. For example Structured Systems Analysis Design Method (SSADM) is suited to implementation in a transaction-oriented relational database system. Unified Modelling Language (UML) is suited to implementation in an object-oriented environment.

Initially methodologies concentrated on tangible aspects of user problems, that is concepts, which could be readily represented by program code and data structures. However, such aspects form only a part of the whole user problem. In effect any information system has two aspects namely hard system and soft system. The hard system part includes several elements such as data, events, processes and interfaces. The system also needs some resources (people, money and equipment) to achieve the required objectives. So a soft system part is developed including quality issues such as the identification of the problem, the user involvement, the organisational structure, goals and policies, the employee job satisfaction, different points of view, the employee’s values and the system acceptability and usability.

2. Approaches Evaluated

OPM: The Organisation Process Modelling method (Warboys, 1999) deals with aspects of both hard and soft systems.

SSADM: The Structured Systems Analysis and Design Method (SSADM, version 4, 1990) is a detailed method covering almost every element of the information system (Duncan, Rackley and Walker, 1995).

UML: The Unified Modelling Language (UML, version 1.3, 1998) is an expressive modelling language that covers every aspects of the system development process (Booch, 1999). UML can be adapted with Business-Oriented Software Engineering process (BOE Process) to cover more fully the modelling of enterprises.

Unified Process: The Unified Process method of 1999 (Rational Software Corporation, 2000) covers most of the hard system elements. It supports object-oriented techniques as its models are based on object, class and relationship concepts. An activity diagram models business processes.

SSM: While mainly dealing with soft aspects, the Soft Systems Methodology (SSM) also deals with some aspects of hard systems. SSM (Checkland and Scholes, 1990) supports activities and processes through using a conceptual model to represent the activities of the root definition.

Workflow (WFMS): The Workflow Management System (Jablonski and Bussler, 1996) introduces new qualities into the task of combining the people, organisation and processes to form a value chain. Such a chain is a management terminology for a string of companies working together to satisfy market demands. The value chain typically consists of one or a few primary value (product or service) suppliers and many other suppliers that add on to the value that is ultimately presented to the buying public. Workflow systems are used to “document and control the transitions between tasks in a process and bring together the resources (human and information) needed to complete each task” (Stark and Lachal, 1995).

3. Comparison of Methodologies

Table 1 compares the different methodologies in terms of aspects of both the soft and hard system approaches. The taxonomy was developed from an analysis of workflow systems (Al-Humaidan and Rossiter, 2001). Table 1 includes hard system aspects such as data (1), events (2), processes (3) and interfaces (4) (Longworth, 1992a; Longworth, 1992b) and soft system elements such as resource (5), quality (6), business issues (7), problem identification (8), user involvement (9), organisational structure, goals and policies (10), employee job satisfaction (11), different views (12), employee values (13) and system acceptability and usability (14) (Checkland and Scholes, 1990). The quality problems include incorrect requirements handled, neglect of the wider organisation, incorrect analysis and poor reasoning. Poor productivity may result from users changing their requirements, the impact of external events to change requirements, unfeasible implementation plans and poor project control. The soft system approach tries to solve some of these problems by placing emphasis on investigating the problem situation using a variety of techniques to determine the organisational policies and goals. In addition the soft approach focuses on wider issues in the social context which may influence the nature of the problem solution such as the organisational structure, employee job satisfaction, employee’s values and the system usability and acceptability including user involvement (Flynn, 1998).

4 Discussion and Conclusions

The different methodologies that are used for developing an information system deal with the hard and soft systems aspects as follows.

Organisational Process Modelling (OPM) is a simple method, which handles principally the interactions between agents as they achieve their goals for modelling the organisational process. It deals with some aspects of the hard system approach and most of the soft system issues. For the latter it uses some of the Soft Systems Methodology (SSM) techniques to deal with the problem. The most serious omission from this method is its lack of facilities for representing data structures.

Structured Systems Analysis and Design Method (SSADM) is a detailed method which covers almost every element of the information system. It deals with every aspect of the hard system issues but only some of the soft system issues. So, there is a trend in the later versions of SSADM to use SSM in the early phases.

Method System elements	OPM	SSADM	UML	Unified Process	SSM	WFMS
1. Data	Not supported	Logical Data Model (LDM)	The class diagram	A. The class diagram B. Databases	Not supported	A. Form-based workflow: fields connected to database. B. Engine-based workflow: data stored in database
2. Events	Goal model (Conceptual Model CM)	Entity Life History (ELH)	The Behaviour (interaction) diagrams	The Behaviour (interaction) diagrams	Not supported	Two types of events: internal and external. They trigger the starting and execution of process instances.
3. Processes	Method model (Role Activity Diagram)	Data Flow Diagram (DFD)	The activity diagram	The activity diagram	Conceptual models	A. Form-based workflow: logic of process. B. Engine-based workflow: process information held.
4. Interfaces	Not supported	Dialogue Design	Modelled in class and component diagrams.	A, User screen sketches/prototypes. B. Internal interfaces: classes and components.	Not supported	Five types of interfaces can be used with engine-based systems. In form-based systems form is interface.
5. Resources	Not supported	Requirements Catalogue (RC).	Modelled by using the stereotype feature.	Project manager plans and schedules process resources.	In root definition, activities related to resources in conceptual models.	Information and human.
6. Quality	Not supported	Requirements Catalogue (RC).	A. In analysis explorative prototypes. B. In design experimental prototypes	A. Inception and elaboration phases: explorative prototypes. B. Tests: Integration, configuration, negative and stress.	Measures for activities in conceptual models. Some activities monitor these measures taking control action to improve matters in proposed system.	Identifying rules followed to perform specific process. Improve supported process by identifying weaknesses and reducing time to perform tasks.
7. Business issues	Method model (Role Activity Diagram)	A. Data Flow Diagrams (DFD). B. Entity Life History (ELH).	Activity diagrams describe and model business process.	Developing business model that defines business processes.	Combination of different perceptions in conceptual models that specify business system options.	Support for business processes.
8. Identify the problem or problem objectives	The system model is used to define the problem scope.	The strategic planning defines the problem that needs to be solved.	The strategic planning defines the problem that needs to be solved.	The strategic planning defines the problem that needs to be solved.	Rich picture presents problem situation including different people perceptions.	Identified in enterprise planning and business area analysis.
9. User involve-ment	A. Gathering information about system. B. Validating models and final system.	A. Gathering information about system. B. Reviewing products of each stage.	A. Gathering info- rmation about system in use case models, CRC and tech. dictionary. B. Review/check prototypes.	A. Gathering info-rmation about system in use cases, business or domain models, suppl. requirements. B. Check/validate arte-facts of iteration//phases.	A. Gathering information about problem situation. B. Choosing activities to construct consensus primary task model. C. Debating to define required changes.	Encourage involvement of users in implementing workflow system.
10. Organi-sational structure, goals and policies	OPM analyses process to define organisation values.	Strategic planning looks at organi-sational structure giving Project Initial Document.	Activity diagram models organisational structure and integration.	Documented in the business model and supplementary requirements.	Presented in A) Rich picture model. B) Primary task model.	Present organisational structure and population. Goals specified in enterprise planning and business area analysis.
11. Employee job satisfaction	Not supported	User may choose Business System Option (BSO) that defining impact on users and training.	Allowing employees to choose suitable way to perform assigned job.	Project feasibility, risk management, team structure, project schedule, project under-standability and sense of accomplishment	User involvement in stages of SSM.	A. System offers tasks to employees who are free to accept them or not. B. System-delivered model enabling users to reject or delegate responsibilities.
12. Different point of views	A. Consider process owner’s view or change process. B. Use dialectic concept C. Rich picture to represent views.	Different views of the system are documented in Requirements Catalogue.	Analyst considers different views of system and resolves contradictions.	Different views are integrated to reach best answer.	Different views are modelled in conceptual models and combined in ways to accommodate them and reconcile conflicts.	Define several process paths to support different views of process.
13. Employee values	Recommends use of SSM to define emp-loyee goals/views	Not Supported.	Not Supported.	Not supported	Documented in Analysis Two that specifies roles, norms and values.	Stored in organisational population.
14. System accepta-bility and usability	OPM attempts to match users’ task and structure of the software system.	A. User involve-ment in developing system. B. Use prototype. C. Study of system impact on staff.	Involvement of users in experi-mental prototypes to verify usability/accepta-bility of system.	A. Involving user in developing system. B. Performing acceptance test. C. Providing users with doc/help line.	Achievement of user requirements and user involvement promotes acceptance/ usability of delivered system.	Acceptance of workflow systems increases if workers’ and business problems are solved. Services relating to user requests must be efficient to satisfy their users.

Table 1: Comparison of Methodologies in Terms of both Hard (1-4) and Soft (5-14) System Aspects.

Unified Modelling Language (UML) is an expressive modelling language that covers all hard aspects of the system development process. UML can be used with any object-oriented development method, such as Unified Process which covers most of the hard system aspects and also some soft system aspects. Unified Process does not support soft aspects such as employee values.

Soft Systems Methodology (SSM) deals with some hard system aspects and all of the soft system aspects. SSM does not support hard system aspects such as data structures, events and the design of interfaces.

WorkFlow Management Systems (WFMS) ostensibly meet every feature of the taxonomy but this is not surprising as the taxonomy is based on WFMS. Certainly WFMS is comprehensive in soft aspects. However, more comprehensive abstractions are found in UML for handling hard aspects such as data and processes.

It can be concluded that there is no methodology that covers all aspects fully. Organisational Process Modelling (OPM) and Soft Systems Methodology (SSM) are relatively strong on soft aspects and weak on hard aspects. Unified Modelling Language (UML) and Unified Process are relatively strong on hard aspects and weak on soft aspects. Structured Systems Analysis and Design Method (SSADM) is perhaps the most comprehensive but some soft aspects are omitted. Workflow systems are promising. They provide a method to deal with both the human (soft) and information (hard) issues but are less advanced in information abstractions than object-oriented systems such as UML. The combination of techniques such as UML and Workflow appears to be the way forward.

References

Al-Humaidan, F, and Rossiter, B N, (2001), A Taxonomy and Evaluation for Systems Analysis Methodologies in a Workflow Context: Structured Systems Analysis Design Method (SSADM), Unified Modelling Language (UML), Unified Process, Soft Systems Methodology (SSM) and Organisation Process Modelling (OPM), Computing Science Technical Report no.751, 32pp. University of Newcastle upon Tyne (2001).

Booch, G., and et al., (1999), The Unified Modelling Language User Guide, Addison Wesley Longman Ltd.

Checkland, P., and Scholes, J., (1990). Soft Systems Methodology in Action, John Wiley and Sons Ltd..

Duncan, J., Rackley, L., and Walker, A., (1995), SSADM in Practice, A version 4 text, Macmillan Press Ltd.

Flynn, Donal J. (1998), Information Systems Requirements: Determination and Analysis, McGraw-Hill.

Jablonski, S., and Bussler, C., (1996), Workflow Management: Modelling Concepts, Architecture and Implementation, International Thomson Computer Press, first edition.

Longworth, G., (1992a), Introducing SSADM, Version 4, NCC Blackwell Limited, First edition.

Longworth, G., (1992b), A users Guide to SSADM, Version 4, NCC Blackwell Limited, First edition.

Rational Software Corporation, (1999, June), OMG Unified Modelling Language Specification, version 1.3, www.rational.com.

Rational Software Corporation, (2000, October), The Unified Process, www.rational.com.

Stark, H. and Lachal, L., (1995), Ovum Evaluates: Workflow, Ovum, London, UK.

Warboys, Brian, Kawalek, Peter, Robertson, Ian, and Greenwood, Mark, (1999), Business Information Systems-A Process Approach, McGraw-Hill Publishing Company.