第2章計算論的思考と設計の再利用性

By the end of this section, you will be able to:

• Describe how business solutions design heuristics and how patterns are used

• Discuss the role of enterprise architecture (EA) and related architecture domains

• Differentiate between enterprise and solution architecture

While computational thinking commonly employs a bottom-up strategy for crafting well-structured components, adaptive design reuse adopts a top-down methodology, emphasizing the creation and assembly of business solutions by combining existing design components. A design component is a reusable element within a larger system that serves a specific purpose. These components, akin to low-level solution building blocks, necessitate minimal modifications. This approach, often termed computing in the large, diverges from individual algorithmic instructions. Instead of composing instructions for the execution of specific actions through computational thinking’s decomposition, adaptive design reuse identifies fitting components based on the articulated requirements of the business solution’s potential users, commonly referred to as stakeholders. The method(s) used to identify these needs will be discussed in detail in Chapter 9 Software Engineering. While adaptive design reuse typically considers algorithmic designs as building blocks, it uses similar techniques (i.e., decomposition, logical thinking and pattern recognition, abstraction/generalization, componentization, testing, and debugging) to identify and reuse building blocks.

The structural designs that stem from the top-down adaptive design reuse approach to business solution design are referred to as business solutions architecture models. A business solution architecture is a structural design that is meant to address the needs of prospective solution users. When a structural design meets these needs, it is considered “architecturally sound” for the problem at hand. Furthermore, to accelerate the putting together of complete solutions, the adaptive design reuse approach relies on architectural models or component assemblies that were developed from previous designs. When the granularity of these architectural models is at the level of subsystems, they are referred to as system family architectures or architectural patterns. An architectural pattern is a reusable solution to a recurring problem in software architecture design. Chapter 10 Enterprise and Solution Architectures Management provides more detail about the various levels of patterns and how to organize them and bookkeep them into pattern catalogs to facilitate reuse.

Business solutions are strategies/systems created to solve specific challenges in a business. They aim to make operations more efficient, improve decision-making, and contribute to overall success. Creating business solutions is a multifaceted and intricate process, demanding knowledge in different technological domains and the relevant business sector. A crucial step in this procedure is outlining the solution’s architecture, serving as a master blueprint, and guiding the entire design and implementation process. A blueprint is a detailed plan or design that outlines the structure, components, and specifications of a building, product, system, or process.

Computational thinking and adaptive design reuse are simply methods used to bootstrap and/or accelerate the design of business solutions by providing a comprehensive set of methods for developing software solutions to business problems—for example, gathering solutions needs, building, and deploying turnkey solutions. These solutions will be discussed in detail in Chapter 9 Software Engineering.

As explained earlier, the computational thinking and adaptive design reuse approach only provide high-level techniques to help create or reuse components. As a result, one of the typical criticisms is that these approaches are too vague, as it is sometimes not clear how they differ from other forms of thought. This is why experts at applying these methods are usually seasoned architects who can instinctively recognize when various types of components may be reused. When you ask Thoughtworks’ enterprise architecture expert Martin Fowler why a particular business solution architecture is sound in a given context, he will often reply that it simply “smells good” to him.

Rather than looking at computational thinking and adaptive design reuse as a best practice set of techniques, it is best to consider them as process patterns that may be applied and adapted to help create a business solution architecture model, which represents the content of the application architecture. Process patterns may, in turn, leverage other process patterns that are more focused. When process patterns that are an inherent part of the process of creating a work product become rules of thumb, they are referred to as heuristics.