The situational method engineering (SME) literature is surveyed and a synoptic evaluation presented in the context of formalizing and regularizing the conceptual framework and underpinning theory. Metamodels proposed for use in SME are evaluated as well as high- level process models for method construction. Method fragments and method chunks are then described formally followed by their identification and creation (from existing methods, from scratch or from past usage). Method creation is then analyzed in terms of various processes for constructing a full methodology from the method fragments/chunks. In particular, we contrast the use of the "map" technique and of the "deontic matrix" technique. The survey is concluded with an evaluation of some ideas on method tailoring and the emerging research on quality evaluation applied to SME.

/pdf/situational-method-engineering-state-of-the-art-review-rsf2nhf319.pdf

Situational method engineering: State-of-the-art review

While previously available methodologies for software like those published in the early days of object technology claimed to be appropriate for every conceivable project, situational method engineering (SME) acknowledges that most projects typically have individual characteristics and situations. Thus, finding the most effective methodology for a particular project needs specific tailoring to that situation. Such a tailored software development methodology needs to take into account all the bits and pieces needed for an organization to develop software, including the software process, the input and output work products, the people involved, the languages used to describe requirements, design, code, and eventually also measures of success or failure. The authors have structured the book into three parts. Part I deals with all the basic concepts, terminology and overall ideas underpinning situational method engineering. As a summary of this part, they present a formal meta-model that enables readers to create their own quality methods and supporting tools. In Part II, they explain how to implement SME in practice, i.e., how to find method components and put them together and how to evaluate the resulting method. For illustration, they also include several industry case studies of customized or constructed processes, highlighting the impact that high-quality engineered methods can have on the success of an industrial software development. Finally, Part III summarizes some of the more recent and forward-looking ideas. This book presents the first summary of the state of the art for SME. For academics, it provides a comprehensive conceptual framework and discusses new research areas. For lecturers, thanks to its step-by-step explanations from basics to the customization and quality assessment of constructed methods, it serves as a solid basis for comprehensive courses on the topic. For industry methodologists, it offers a reference guide on features and technologies to consider when developing in-house software development methods or customising and adopting off-the-shelf ones.

/pdf/situational-method-engineering-330zpuwjqn.pdf

Situational Method Engineering

With the research described in this dissertation, we aim to shed light on the characteristics of process improvement efforts by looking at their evolution (how to change?) rather than their content (what to change?). This research is triggered by three main propositions, derived from earlier work: (i) methods can be dissected into method fragments; (ii) the process of method evolution can be dissected into method increments; and (iii) method increments can be used to guide method construction. These assumptions have not been thoroughly tested and explored yet. Especially the method increment concept is not well understood. The research reported in this dissertation tries to fill this gap by answering the following question: How can method increments be used to support process improvement efforts? The dissertation provides an answer to this question through three parts. The first part describes two studies that investigate the characteristics of process improvement efforts, by looking at such efforts in the domain of agile software development. One study describes a case study, resulting in a method description for agile software product management. While this relates more to the what than the how of process improvement, it provides preliminary insights on the structure of process improvement efforts. We expand upon this insight in the second chapter, in which we analyze and compare the introduction of Scrum at four case companies. The results show a variety of approaches, with heterogeneous effects on effectiveness. In the second part of the dissertation, we develop a better understanding of the notion of method increment. The first chapter expands on the topic of knowledge gathering related to method increments by providing a detailed description and evaluation of a method increment template. In the second chapter of this part, we propose and evaluate a formalization of the process-deliverable diagram, specifically aimed at describing method changes, along with a planning framework and process for generating plans that comply with different types of constraints. In the third part, we describe the design, realization, and evaluation of a knowledge management system that supports incremental method engineering. This system, the Online Method Engine (OME), encompasses the dissemination of method knowledge, linked to process assessment, improvement, and enactment. The OME is based firmly upon the results from the first two parts, and demonstrates a potential solution that helps practitioners during process improvement efforts.

Supporting Process Improvement using Method Increments

Software product line engineering is an approach that develops and maintains families of products while taking advantage of their common aspects and predicted variabilities. Indeed, software product lines (SPL) are an important means for implementing software variability which is the ability of a system to be efficiently extended, changed, customized or configured for use in a particular context. Variability needs in software are constantly increasing because variability moves from mechanics and hardware to software and design decisions are delayed as long as economically feasible. Numerous SPL construction approaches are proposed. Different in nature, these approaches have nevertheless some common disadvantages. We have proceeded to an in-depth analysis of existing approaches for the construction of Software Product Line within a comparison framework in order to identify their drawbacks. We suggest overcoming these drawbacks by an improvement of the tool support for these approaches and for their interactivity with their users. We propose to study a particular software product line which is ERP as experimentation. SPL engineering is considered as unavoidable approaches to support reuse in systems development, as a viable and important software development paradigm. It allows to companies to realize a real improvements in time to market, cost, productivity, quality and flexibility. In fact, SPL techniques are explicitly capitalizing on commonality. They try formally to manage the variations of the products in the product line. To understand the problem of variability in software product lines, we've studied a number of references to provide an overview of the current activities in this area. The reference (28) presents a model for product line variability. They try to model variability by feature model, which describes functional and nonfunctional requirements of the product line. This feature model structures the common and variable features. The architectural variability model is described by architectural variation points. The reference (3) presents the variability as the key to software reuse and he asserts that variability can be associated with different abstraction levels of development (from requirement specification to running code). They propose a

/pdf/framework-for-evolving-software-product-line-2uuqq630qx.pdf

Framework for Evolving Software Product Line

Context: Method engineering approaches are often based on the assumption that method users are able to explicitly express their situational method requirements. Similar to systems requirements, method requirements are often vague and hard to explicate. In this paper we address the issue of involving method users early in method configuration. This is done through borrowing ideas from user-centered design and prototyping, and implementing them on the method engineering layer. Objective: We design a computerized tool, MC Sandbox, to capture method requirements through the use of method-user-centered method configuration, hence bridging the gap between systems developers' and method engineers' understanding of and expectations on a situational method. Method: The research method adopted can be characterized as multi-grounded action research. Our implementation of multi-grounded action research follows the traditional 'canonical' action research method, which has cycles of diagnosing, action planning, action taking, evaluating, and specifying learning. The research project comprised three such action research cycles where 10 action cases were performed. Results: MC Sandbox has proven useful in eliciting and negotiating method requirements in a continuously ongoing dialog between the method users and the method engineers during configuration workshops. The results also show that the method engineer role rotated among the systems developers and that they were indeed committed to the negotiated methods during the systems development projects. Conclusion: It is possible for method users to actively participate and construct suitable situational methods if they are provided with appropriate high-level modelling concepts, such as method components, configuration packages and configuration templates. This way, the project members' understanding of the current development practice develops incrementally, both in terms of understanding the needs and available method support. In addition, both method requirements and commitments are made explicit, which are important aspects when working with method configuration from a collaboration point of view.

MC Sandbox: Devising a tool for method-user-centered method configuration

In Method Engineering (ME) science, the key issue is the consideration of information system development methods as fragments. Numerous ME approaches have produced several definitions of method parts. Different in nature, these fragments have nevertheless some common disadvantages: lack of implementation tools, insufficient standardization effort, and so on. On the whole, the observed drawbacks are related to the shortage of usage orientation. We have proceeded to an in-depth analysis of existing method fragments within a comparison framework in order to identify their drawbacks. We suggest overcoming them by an improvement of the “method service” concept. In this paper, the method service is defined through the service paradigm applied to a specific method fragment – chunk. A discussion on the possibility to develop a unique representation of method fragment completes our contribution.

/pdf/from-method-fragments-to-method-services-1ekswob3ef.pdf

From Method Fragments to Method Services

The Method Engineering (ME) community is a prolific research domain where competing Situational Method Engineering (SME) approaches have been defined and used for composing, adapting or/and configuring a method into modular constructs according to their own modularization vision. This diversity shows the richness of the ME domain but implies some drawback like unnecessary confusion for non ME expert, lack of standard & interoperability, lack of implementation tool. However, researchers are agreed that a common ground in SME is a hot matter of discussion. Assuming that the differences between SME approaches are purposeful, we propose to reach a semantic common ground on what types of core concepts constitute a method descriptor. To achieve it, an ontology-based approach is applied in SME to design an ontology of method descriptors as a domain ontology. The semantics of the six most popular SME approaches modular constructs are defined according to this ontology in order to show its usage and its relevance. Finally, usage scenarios have been sketched to show that the ontology can be the start up phase for reducing the ME drawbacks mentioned above.

/pdf/towards-common-ground-in-sme-an-ontology-of-method-3jc7evccup.pdf

Towards Common Ground in SME: An Ontology of Method Descriptors

In Method Engineering (ME) science, the key issue is the consideration of information system development methods as fragments. Numerous ME approaches have produced several definitions of method parts. Different in nature, these fragments have nevertheless some common disadvantages: lack of implementation tools, insufficient standardization effort, and so on. On the whole, the observed drawbacks are related to the shortage of usage orientation. We have proceeded to an in-depth analysis of existing method fragments within a comparison framework in order to identify their drawbacks. We suggest overcoming them by an improvement of the ?method service? concept. In this paper, the method service is defined through the service paradigm applied to a specific method fragment ? chunk. A discussion on the possibility to develop a unique representation of method fragment completes our contribution.

Situational Method Engineering aims at constructing methods adapted to a given situation, either by a construction from a set of predefined method components or by a customization of an existing method using different techniques: configuration, extension, reduction, and so on. However, these techniques are still limited in practice, as considered complicated and heavy to implement. In this paper, we describe a practitioner experience of a gradual integration of different method components (issued from agile methods of software development). In a real case of a development company, we have practiced and observed the progressive introduction of agile method components instead of the construction or customization of methods in one go. We discuss the lessons learned from this experience and define different research perspectives.

/pdf/progressive-integration-of-method-components-a-case-of-agile-qgyw20nz9p.pdf

Adrian Iacovelli

Papers

From Method Fragments to Method Services

Towards Common Ground in SME: An Ontology of Method Descriptors

From Method Fragments to Method Services

Progressive integration of method components: A case of agile is development methods