Klaus Schmid

PuLSE: a methodology to develop software product lines

Software product lines have recently been introduced as one of the most promising advances for efficient software development. Yet upon close examination, there are few guidelines or methodologies available to develop and deploy product lines beyond existing domain engineering approaches. The latter have had mixed success within commercial enterprises because of their deployment complexity, lack of customizability, and especially their misplaced focus, that is on domains as opposed to products. To tackle these problems we developed the PuLSETM (Product Line Software Engineering) methodology for the purpose of enabling the conception and deployment of software product lines within a large variety of enterprise contexts. This is achieved via product-centric focus throughout the phases of PuLSETM, customizability of its components, incremental introduction capability, maturity scale for structured evolution, and adaptations to a few main product development situations. PuLSETM is the result of a bottom-up effort: the methodology captures and leverages the results (the lessons learned) from our technology transfer activities with our industrial customers. We present in this paper the main ideas behind PuLSETM and illustrate the methodology with a running example taken from our transfer experience.

Cool features and tough decisions: a comparison of variability modeling approaches

Variability modeling is essential for defining and managing the commonalities and variabilities in software product lines. Numerous variability modeling approaches exist today to support domain and application engineering activities. Most are based on feature modeling (FM) or decision modeling (DM), but so far no systematic comparison exists between these two classes of approaches. Over the last two decades many new features have been added to both FM and DM and it is tough to decide which approach to use for what purpose. This paper clarifies the relation between FM and DM. We aim to systematize the research field of variability modeling and to explore potential synergies. We compare multiple aspects of FM and DM ranging from historical origins and rationale, through syntactic and semantic richness, to tool support, identifying commonalities and differences. We hope that this effort will improve the understanding of the range of approaches to variability modeling by discussing the possible variations. This will provide insights to users considering adopting variability modeling in practice and to designers of new languages, such as the new OMG Common Variability Language.

A customizable approach to full lifecycle variability management

In order to enable a smooth transition to product line development for an organization that so far only performed single system development, it is necessary to keep as many of the existing notations and approaches in place as possible.This requires adaptability of the basic variability management approach to the specific situation at hand. In this paper we describe an approach that enables homogenous variability management across the different lifecycle stages, independent of the specific notation. The approach is accompanied by a meta-model and a process for introducing the variability management approach by developing a notation-independent representation. This approach has so far been applied in several cases where our Product Line engineering method PuLSETM has been introduced into a software development organization.

The economic impact of product line adoption and evolution

An organization faces many challenging decisions when transitioning to product line development: What is the best way to adopt a product line approach? How can we avoid disrupting regular product development? Once adopted, how should we evolve the product line? The article discusses how to optimize a product lines economic benefits by considering the adoption context and using product line scoping techniques.

A systematic approach to derive the scope of software product lines

Product line scoping is a critical activity because it elicits the common realms upon which the different products of a product line can be optimally engineered with respect to economies of scope. This, in turn, upper bounds the overall economic benefits that can be accrued from product line based development. Inherently, product line scoping is difficult because of the complexity of the factors that must be taken into account. Many are not known a priori. Traditional scoping approaches (from domain engineering) have focused on the notion of application domains. However, domains proved difficult to optimally scope and engineer from an enterprise standpoint because a domain captures extraneous elements that are of no interest to an enterprise which must focus on particular products, whether existing, under development, or anticipated. Hence, the domain view provides a flawed economic basis for making a scoping decision. We introduce PuLSE-Eco, a technique especially developed to address the aforementioned issues. Its main characteristics are: a complete product-centric orientation done via product maps, the separation of concerns achieved through the definition and operationalization of strategic business objectives, and last, diverse types of analyses performed upon product maps allowing scoping decisions based on these objectives. We illustrate the technique with a running example.

A comparison of decision modeling approaches in product lines

It has been shown that product line engineering can significantly improve the productivity, quality and time-to-market of software development by leveraging extensive reuse. Variability models are currently the most advanced approach to define, document and manage the commonalities and variabilities of reusable artifacts such as software components, requirements, test cases, etc. These models provide the basis for automating the derivation of new products and are thus the key artifact to leverage the flexibility and adaptability of systems in a product line. Among the existing approaches to variability modeling feature modeling and decision modeling have gained most importance. A significant amount of research exists on comparing and analyzing different feature modeling approaches. However, despite their significant role in product line research and practical applications, only little effort has been devoted to compare and analyze decision modeling approaches. In order to address this shortcoming and to provide a basis for more structured research on decision modeling in the future, we present a comparative analysis of representative approaches. We identify their major modeling concepts and present an analysis of their commonalities and variabilities.

Building Dynamic Software Product Lines

Dynamic software product lines extend existing product line engineering approaches by moving their capabilities to runtime, helping to ensure that system adaptations lead to desirable properties.

Applying Product Line Concepts in Small and Medium-Sized Companies

Small and medium-sized enterprises work under heavy constraints: They need to be very flexible and fast in their reaction to customer requests, thus limiting their possibility for long-term planning. In a partially publicly funded project, the authors have started to apply their Product Line Software Engineering method developed at Fraunhofer IESE, in six small and medium-sized companies addressing six different domains. The article presents first experience and lessons learned from 24 months of project work; including first results within the companies.

A Cost Model for Software Product Lines

In this paper we present a first-order cost model that describes the costs associated with developing products in a product line organization. The model addresses a number of issues that we present as a set of scenarios. The goal of this work is to develop models of varying granularity that support a managers decision-making needs at a variety of levels. The basis of these models is the relationships among the artifacts of the product line.

Variability modeling for distributed development - a comparison with established practice

The variability model is a central artifact in product line engineering. Existing approaches typically treat this as a single centralized artifact which describes the configuration of other artifacts. This approach is very problematic in distributed development as a monolithic variability model requires significant coordination among the involved development teams. This holds in particular if multiple independent organizations are involved. At this point very little work exists that explicitly supports variability modeling in a distributed setting. In this paper we address the question how existing, real-world, large-scale projects deal with this problem as a source of inspiration on how to deal with this in variability management.