Today's software enterprise systems continue to su er from symptoms like aging and the lack of evolution capabilities. One of the major reasons for this problem is that the future maintenance and evolutions are undervalued during traditional software systems implementation. The best way to overcome the problem of software aging and the misalignment with business is by placing the change in the center of software system lifecycle. This thesis proposal presents the main activities required to create a method allowing the integration of change in the lifecycle of SMEs information systems (IS). By following a design science approach in IS, this study will explore the main challenges faced by small and medium enterprises (SMEs) regarding the evolutions of their IS, and the bene ts of an agile approach to integrate change throughout the IS lifecycle.
Small and medium enterprises (SMEs) represent a sector with one of the largest growth in the economy. Following a long term and well known trend of large companies, SMEs become increasingly dependent on information systems to support their operations (Premkumar, 2003). While there has been extensive research concerning the lifecycle and the evolution of information systems (IS) in large companies, only few research has explored the topic of lifecycle and evolution of IS in SMEs.
Nowadays, business environments progress and change rapidly to move with evolving markets. Those evolutions lead to a considerable amount of changes in business processes. Some of those are supported by IS and since the business processes change or new market opportunities appear, the IS needs to be adapted in order to continue to support the business. Regarding this context, the need of IS evolution is intrinsic.
Currently, if SMEs want to compete e ectively within their market, the
adoption of a sustainable enterprise IS is necessary. In terms of information system
needs, SMEs tend to adopt a kind of incremental approach to problem
resolution. Namely, when they need an application to manage accounting, they buy
(or develop) it; when they need an application to manage stocks, they buy (or
develop) it, etc.
IS undergo several distinguish stages during its life. According to
To integrate change in the software development, a typical approach is to
follow an iterative and incremental process. These development methodologies
are also known as agile software development methodologies
According to
Based on these aspects, during preliminary research work, we addressed the following research question : \How to apply agile principle to the deployment of an open source ERP in SMEs". The output of this previous part was an agile and incremental methodology for ERP deployment. We published a rst paper in the proceedings of the AIM 2009 Conference (Metrailler and Estier, 2009), and an extended version of our research is currently under review (Metrailler and Estier, 2011).
By comparison with traditional (waterfall) methodologies, this approach allows, among others, an improvement of exibility and of project priorities in planning, a reduction of the ERP complexity through several rollouts and a better distribution of the extra workload among stakeholders. This methodology has been developed in collaboration with business partners who are currently using it and obtain good results.
Our agile deployment methodology progress in an incremental manner, which means that between two rollouts, the system is used in production. Since each rollout provides new features, there is an evolution of the system in production. Consequently, this approach helps SMEs to integrate a continuous evolutional perspective in their ERP lifecycle. Therefore, the continuation of this research focuses on identifying challenges relative to IS evolution in SMEs and on the opportunities concerning the use of an agile approach for evolution.
This paper is structured as follows. In next section, we expose the research questions. In section 3, we present an overview of previous research in systems evolution. In Section 4, the research methodology used to conduct this study is described, namely design science in IS. Section 5 presents the current state of our work and the required stages to carry out this research. In addition, Section 5 presents the research activities for each stage of the research methodology. Finally, the expected contributions and the limitations of this research are presented in Sect. 6. 2
Regarding the fact that to support business activities the IS must evolve, and the fact that an agile, iterative and incremental process allows the integration of change, the general research question is as follows:
To answer this question, we aim to design an agile method to conduct IS evolution in SMEs and to evaluate its impact on IS evolution. In order to accomplish this, additional research questions also need answers: Concerning SMEs IS in general: { How do SMEs deal with the changes of their IS? Are there di erences between in-house developed software systems and commercial o -the-shelf (COTS) systems? { What are the advantages and disadvantages of both in-house developed systems and COTS systems regarding system evolution? { Do known laws of software evolution apply to SMEs information systems evolution?
{ In the rst stage of IS lifecycle, is an iterative and agile implementation methodology more e cient regarding the incoming evolutions of IS? { How can we enhance such iterative and agile implementation methodology to integrate an evolutional perspective? 3
In this section, we present a brief overview of previous research on which we refer and on which we will rely to design our artifact to conduct evolution in SMEs IS. 3.1
The term evolution is elusive to de ne. Dictionaries and common sense refer to the following de nition: \Evolution is a process of gradual development in a particular situation or thing over a period of time1".
When it comes to software systems, the term evolution has many di erent
interpretations depending on the role of stakeholder. To de ne evolution in a
way that is independent of subjective interpretations and that captures
characteristics of evolution in software systems, Lehman (cited in
One of the principal researchers in the eld of software evolution is Professor
M.M. Lehman. He is the principal developer of the laws of software evolution.
These laws have been developed for more than 30 years with the rst version
published in 1974. They are the basis for the elaboration of a theory of software
evolution
Lehman made a fundamental distinction between three types of programs.
Those written to satisfy a xed speci cation (S-type, S for speci cation), those
developed to satisfy a need in the real world (E-type, E for Evolving), and a third
type of programs the P-type (P for Problem). This third type always satis es
the de nition of either S- or E-type, thus, in his subsequent research Lehman
ignored the type P. This distinction is known as the SPE taxonomy
Several researchers described the evolution of E-type softwares. To compare
and categorize evolutions of such systems, three decades ago, Lientz and Swanson
(1980) proposed a software maintenance typology that distinguishes among
perfective, adaptive and corrective maintenance activities. More recently,
A signi cant amount of research has been completed on what distinguish SMEs
from larger rms, it points out that the principal distinctions are the limited
resources and the limited knowledge base of SMEs
SMEs information systems can be grouped in two main categories: enterprise systems developed by the company (in-house developed software systems) and enterprise systems purchased by the company (COTS) like ERP or SaaS (Software as a Service) and BoB (Best of Breed) approaches. Based on the fact that all these solutions will eventually evolve according to the business context of the company, our interest is to determine the advantages and inconvenients of each category and whether a solution is more able to evolve in SMEs business context.
A recent study of
We believe that the existing research has missed to explore the evolution of SMEs information system.
The research concerning the laws of software evolution is based on the evolution of large E-type software. Lehman's Evolution laws are originally based on observations regarding the evolution of IBM's OS/360 and OS/370.
Empirical research on the evolution of open source software point out that
the evolution of such software breaks some of Lehman's laws. One way among
others to explain these results is the decentralized manner by which such software
are developed and the loosely-coupled community of developers (Scacchi, 2003).
Another study concerning the evolution of Eclipse plug-ins
When it comes to the taxonomy of software evolution or the typology of software change, there is no application of them to SMEs context. The fact of carrying out such a research could point out interesting elements concerning namely the temporal properties, the object, the system properties or the challenges relative to software changes in SME.
On another side, there is no research concerning the evolution challenges relative to in house developed or COTS systems in SME, is there an approach which is more able to evolve than the other? 4
To carry out this research, we use a design science approach in IS. According to
Pe ers et al. (2007) propose a Design Science Research Methodology (DSRM) process model which is composed by six steps (1. Identify problem and motivate, 2. De ne objectives of a solution, 3. Design and development, 4. Demonstration, 5. Evaluation, 6. Communication). This process is structured in a nominally sequential order. However, it is not expected that researcher should always proceed in order from activity one to six. Pe ers et al. (2007) describe four \entry points" through which the researcher can start (from 1 to 4).
During the preliminary research work, we followed a problem-centered approach by starting with step 1. We identi ed some issues that emanate from SMEs, namely the problem of exibility in the traditional (waterfall) ERP deployment approach, di culties related to the workload, the limited resources of SMEs, and the complexity of appropriating the tool. Using an iterative approach we performed each step to create our agile deployment methodology.
The sequel of our research will follow a Design- and Development-Centered approach as described by Pe ers et al. (2007) and illustrated in Fig. 1. Indeed, based on the agile deployment methodology we developed previously, we want to build a model of evolution management integrated throughout the lifecycle of SMEs IS. Obviously, new questions arise concerning the evolution of SMEs IS: their nature, the causes which trigger them, the applicability of our methodology to other type of IS, etc.
Our research activities for each stage of the DSRM process model (see Fig.1) are described in the next section.
Nominal process sequence
Identify Problem & Motivate Conduct SMEs IS evolution with an agile approach
Define
Objectives of a fIrceeenn eEvonSIlilmcfuoeotplciumoryotncpiovlaieennssIS sustainabiliy y r o e h T
Design & Development
Method inspired by common evolutions and agile principles
Process Iteration e Demonstration g d e l w tonooKwH Caartsriefeaficsunttseuaedtnhiedesittso
Evaluation
Communication s i lys e a g i,trscenAM lonedKw Iomafptphlieelomaterptnirfotaajcetitcoitnn rya eg iliicspnD loednKw McoInSfaejornuedrnncaelss Design & DevelopmentCentered
Approach
Possible Research Entry Points
This section gives a brief description of the current state of our research and introduces the ideas we are currently working on. The stages of the research follow the DSRM process model described in Sect. 4.
As mentionned before, the sequel of this research will follow a Design- and Development- Centered approach (Pe ers et al., 2007). In our case, the entry point of the research is the reuse of our agile methodology concepts to build the artifact of this research, namely a method to conduct IS evolution in SMEs. 5.1
After having identi ed the entry point of the research, the next activity in Pe ers
DSRM (Fig. 1) is to de ne the research problem and to justify the value of
the solution. The global research problem has already been de ned in Sect. 1.
Today's software enterprise systems continue to su er from symptoms like aging
and the lack of evolution capabilities. One of major reasons of this problem is
that the future maintenance and evolutions are undervalued during traditional
software systems implementation. The best way to overcome the problem of
software aging and misalignment is to center the software system lifecycle on
change management
The objective of our artifact is to conduct SMEs IS evolution by applying principles of agile software development methodologies to the management of IS evolutions. Our previous work points out that an agile approach for ERP deployment is more aligned with SMEs context than traditional waterfall approaches. Moreover, it is well known that agile development methodologies enable the integration of change throughout software development. We believe that such approach enables the simpli cation of IS evolutions, an improvement of IS reactivity towards business adjustments and an expansion of the IS lifecycle by maintaining an evolution stage instead of going through servicing and phase-out stages. The IS should be designed for evolution from the earliest stage of its lifecycle, by having the evolution integrated in its conception and deployment. 5.3
Despite these facts, the design and the development of such an artifact require
investigations regarding the evolutions of SMEs system. As mentioned in the
literature review (see Sect. 3), it is not clear whether general E-type software
evolution laws are applicable to medium sized systems or to systems that
incorporate software applications from di erent vendors. In order to address IS
evolution issues, it is essential to identify these issues. In the literature, there is
no research identifying software evolutions in SMEs and their context.
Nevertheless, the literature provides tools to undertake such research, that are typology
and taxonomy of software evolution
After having shed light on these interrogations, we would have the basis to start the development of our artifact. The design and development activities will be an iterative process around small case studies to re ne and demonstrate the use of the artifact. When our artifact seems to be mature enough, we would apply it to conduct real SME's IS evolutions. 5.4
After the development of the artifact, it is necessary to evaluate the artifact
to determinate how well it works
In this section we discuss about limitations and contributions of our study.
Our study has limitations. The rst limitation we distinguish is relating to the iterative and agile approach uses by our artifact to manage IS evolution. Our previous work is based on an open source system, consequently, we have a complete access to its architecture and to its data model. This transparency allows us to master the technical dependencies among components which permits us to uncouple them if necessary. We do not apply yet our previous work to a proprietary system, but we guess the IS's exibility will be reduced to the editor's choices.
The second limitation to the adoption of our artifact results from the iterative approach and the integration of the evolution perspective in the IS lifecycle. In fact, the IS will be in continuous evolution and those who have to manage it may need to provide an extra e ort.
Moreover, we assume that the evolution perspective should be integrated since the beginning of the system implementation. If elements of the systems are strongly interconnected since its implementation, it is possible that our artifact could not be integrally used.
Although this study has limitations, it o ers several contributions to practice as well as research. In terms of research, this work will extend the knowledge on evolution of SMEs software systems. The exploratory, qualitative research has a primary contribution of identifying the strong and weak points of both in-house developed and COTS systems regarding their evolution capabilities. Its secondary contribution lies in the identi cation of the main reasons why software evolves in SMEs and of what types are these evolutions. Moreover, these results will allow us to evaluate how the evolution of small-medium software systems conforms to the laws of software evolution.
By describing and understanding the evolutions of SMEs software systems and the main activities generated by these evolutions, one then has the potential to suggest improvements. The contribution of the design and development phase of a design science research is to determine how to more sustainably conduct the lifecycle of SMEs software systems. To these ends, this study delivers a method to integrate an evolutional perspective into the lifecycle of such software systems. Lientz, B. P. and Swanson, E. B. (1980). Software maintenance management: a study of the maintenance of computer application software in 487 data processing organizations, volume 4. Addison-Wesley Reading MA.
Loh, T. C. and Koh, S. C. L. (2004). Critical elements for a successful enterprise resource planning implementation in small-and medium-sized enterprises. International Journal of Production Research, 42(17):3433{3455.
Madhavji, N. H., Lehman, M., Perry, D., and Ramil, J. F. (2006). Software evolution and feedback. Wiley Online Library.
Mens, T., Wermelinger, M., Ducasse, S., Demeyer, S., Hirschfeld, R., and Jazayeri, M. (2005). Challenges in Software Evolution. Eighth International Workshop on Principles of Software Evolution (IWPSE'05), pages 13{22. Metrailler, A. and Estier, T. (2009). Deploiement agile d'ERP open source en
PME. In Actes du Colloque AIM 2009.
Metrailler, A. and Estier, T. (2011). Agile Deployment Methodology for Open Source ERP in SME. Submited for publication to "Revue Systemes d'Information et Management", under review.
Pe ers, K., Tuunanen, T., Rothenberger, M. a., and Chatterjee, S. (2007). A Design Science Research Methodology for Information Systems Research. Journal of Management Information Systems, 24(3):45{77.
Premkumar, G. (2003). A meta-analysis of research on information technology implementation in small business. Journal of Organizational Computing and Electronic Commerce, 13(2):91{121.
Scacchi, W. (2003). Understanding Open Source Software Evolution : Applying , Breaking , and Rethinking the Laws of Software Evolution. Arti cial Life. Stender, M. (2002). Outline of an Agile Incremental Implementation Methodology for Enterprise Systems. In Proceedings of the Eighth Americas Conference on Information Systems, pages 907{917.