Clemson CPSC 950 - Modeling Variation in Production Planning Artifacts

Unformatted text preview:

Modeling Variation in Production Planning Artifacts Gary J. Chastek Software Engineering Institute [email protected] John D. McGregor Clemson University [email protected] Abstract Production planning and variation modeling are interdependent parallel activities critical to the success of a software product line. Software product line organizations design a production capability to satisfy their business goals. That production capability is dependent on and must support the full range of product variation. Current techniques for variation modeling identify and handle variations among products but fail to recognize variations that result from business goals such as rapid time to market which are satisfied by how products are built. In this paper we present a view of our production planning technique and describe our preliminary research into its relation to variation modeling. 1. Introduction Variation distinguishes a software product line from single product development. A variation is any concern that will vary from one product to another or that may vary in time. Product feature models alone are inadequate for modeling the full range of variations the software product line organization must manage. They capture product but not production variations. Goals like mass customization are production rather than product issues, and can yield strategically-significant variations. The variations in how products are built are a direct result of the goals of the organization. For example, a business goal to compete in the global market can lead to variations in testing processes used depending upon the market for which each product is intended. These variations are the result of strategic decisions made during early product line planning activities of the “What to Build” (WTB) pattern [1], shown in Figure 1. Scoping uses the marketing analysis to identify the members of the product line. During this activity variations in features among the products are identified. The market analysis identifies market segments and analyzes the differences among them. Building the business case requires the creation of a justification for the use of the organization’s assets to create and operate the product line organization. It also provides the opportunity to identify additional variations that are not directly related to product content but that are related to product production. We will call the variations found during the WTB activities strategic variations. A change in a strategic variation results in a change to at least one of the WTB artifacts, and vice versa. For example, expanding the scope of the product line may result in additional test requirements if the additional product is intended for a new market. UnderstandingRelevantDomainsMarket AnalysisTechnologyForecastingBuilding a Business CaseScopingDomainModelsProductSetMarketClimateProduct LineScopeJustificationProduct SetBusinessCaseTechnologyPredictionsTechnologyPredictionsMarketClimateKeyInformsActivity OutputArtifa ctUnderstandingRelevantDomainsMarket AnalysisTechnologyForecastingBuilding a Business CaseScopingDomainModelsProductSetMarketClimateProduct LineScopeJustificationProduct SetBusinessCaseTechnologyPredictionsTechnologyPredictionsMarketClimateKeyInformsActivity OutputArtifa ctKeyInformsActivity OutputArtifa ctArtifa ct Figure 1 What to build pattern In addition to strategic we also define tactical variations as those variations that arise from the resolution of strategic variations. For example, the development process will have a step in test planning where the levels of test coverage are determined by the intended market. Tactical variations typically do not necessitate a change in a strategic variation but could if, for example, the choice an architectural mechanism provided the possibility of supporting a wider range of products. The progression from strategic to tactical variation and ultimately to variation point corresponds to positions along the Variation axis in the space defined in [2], as illustrated in Figure 2. The strategic variationsrepresent “what concerns” are important and are placed further out on the Variation axis than the tactical variations that represent “how” those concerns are addressed. Similarly, tactical variations are placed further out than variation points since tactical variations represent the broad concern while a specific variation point addresses only a portion of that concern. Finally, a bound variation point represents a fixed point in the two-dimensional single-product plane. This progression corresponds to the hierarchical separation of concerns described by [3]. Strategic VariationTactical VariationVariation PointVariationStrategic VariationTactical VariationVariation PointVariation Figure 2 Variations in context The goal of this paper is to describe our preliminary research into the relation between variation modeling and production planning. Its contribution is to identify key software product line variations not currently considered (i.e., variations related to how products are produced1) and to extend the work of Berg et al [2] by explicitly identifying strategic and tactical variation points. The subsequent sections describe our work in production planning, variations in the production system and their relation to variation modeling in general, and finally the current directions of our ongoing research. 2. Production Planning Artifacts How a software product line organization builds its products is a system, the production system [5], that has both functionality (e.g., the development tools employed) and quality attributes (e.g., how quickly a specified product can be delivered). Production planning devises a production system that • Satisfies the organization’s goals and constraints for its product line • Coordinates the design of the core assets with the production system • Communicates the effective use of the production system to the product developers 1 Schmid and Eichelberger have addressed binding time as a meta-variability


View Full Document

Clemson CPSC 950 - Modeling Variation in Production Planning Artifacts

Download Modeling Variation in Production Planning Artifacts
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Modeling Variation in Production Planning Artifacts and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Modeling Variation in Production Planning Artifacts 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?