Hands-On Design Patterns and Best Practices with Julia

Proven solutions to common problems in software design for Julia 1.x

Tom Kwong

• Éditeur: Packt Publishing
• Année: 2020
• ISBN:
  9781838646615
• Édition: 1
• Sujet: Informatik
• Sous-sujet: Programmieralgorithmus

Design and develop high-performance, reusable, and maintainable applications using traditional and modern Julia patterns with this comprehensive guide

Key Features

• Explore useful design patterns along with object-oriented programming in Julia 1.0
• Implement macros and metaprogramming techniques to make your code faster, concise, and efficient
• Develop the skills necessary to implement design patterns for creating robust and maintainable applications

Book Description

Design patterns are fundamental techniques for developing reusable and maintainable code. They provide a set of proven solutions that allow developers to solve problems in software development quickly. This book will demonstrate how to leverage design patterns with real-world applications.

Starting with an overview of design patterns and best practices in application design, you'll learn about some of the most fundamental Julia features such as modules, data types, functions/interfaces, and metaprogramming. You'll then get to grips with the modern Julia design patterns for building large-scale applications with a focus on performance, reusability, robustness, and maintainability. The book also covers anti-patterns and how to avoid common mistakes and pitfalls in development. You'll see how traditional object-oriented patterns can be implemented differently and more effectively in Julia. Finally, you'll explore various use cases and examples, such as how expert Julia developers use design patterns in their open source packages.

By the end of this Julia programming book, you'll have learned methods to improve software design, extensibility, and reusability, and be able to use design patterns efficiently to overcome common challenges in software development.

What you will learn

• Master the Julia language features that are key to developing large-scale software applications
• Discover design patterns to improve overall application architecture and design
• Develop reusable programs that are modular, extendable, performant, and easy to maintain
• Weigh up the pros and cons of using different design patterns for use cases
• Explore methods for transitioning from object-oriented programming to using equivalent or more advanced Julia techniques

Who this book is for

This book is for beginner to intermediate-level Julia programmers who want to enhance their skills in designing and developing large-scale applications.

Section 1: Getting Started with Design Patterns

The aim of this section is to introduce you to how design patterns are used in general and how Julia is different from the object-oriented programming paradigm.

This section contains the following chapter:

• Chapter 1, Design Patterns and Related Principles
  

Design Patterns and Related Principles

Nowadays, learning and applying design patterns is an important aspect of software engineering. Design patterns are like water – you can't live without them. Don't believe me? Just ask hiring managers, and you will find that many of them have design patterns in their job postings as well as related questions in job interviews. It is a common belief that design patterns are important ingredients for software development and everyone should know them.

In this chapter, we will provide some context about why design patterns are useful and how they have served us well in the past few decades. By understanding the motivation behind design patterns, we will be able to set forth a set of guiding principles for developing software. The following topics will be discussed in this chapter:

• The history of design patterns
• Software design principles
• Software quality objectives

Let's get started!

The history of design patterns

Design patterns is not a new concept to computer programmers. Since personal computers became more affordable and popular in the 1980s, the programming profession flourished and a lot of code was written for a variety of applications.

I remember that, when I was 14 years old, learning the GOTO statement for a BASIC program was one of the coolest things. It literally allowed me to take a control flow to a different part of the code at any time. Perhaps not too surprisingly, when I learned about structured programming and the Pascal language in college, I started to realize how GOTO statements produce messy spaghetti code. Using GOTO for branching purposes is a pattern. It's just a bad one because it makes code difficult to understand, follow, and debug. In today's lingua franca, we call them anti-patterns. When it comes to structured programming techniques, organizing code in small functions is a pattern as well, one that has been taught as a mainstream subject in programming courses.

When I graduated from college, I started my programming career and spent plenty of time hacking away. I had the opportunity to do various kinds of research and find out how systems are designed. For example, I realized that the Unix operating system has a beautiful design. That is because it consists of many small programs, which individually do not have a ton of functionality, but you can compose them in any number of ways to solve more complex problems. I was also fond of the Scheme programming language, which came out of MIT's AI Lab. The simplicity and versatility of the language still amazes me today. Scheme's heritage can be traced to Lisp, which had some influence on how the Julia language was designed.

The rise of design patterns

In 1994, while I was diving deep into C++ and distributed computing for a financial application, four software professionals, also known as the Gang of Four or GoF, came together and published a book about design patterns, and it took the object-oriented programming community by storm. The group collected and classified 23 design patterns that were commonly utilized when developing large-scale systems. They also chose to explain the concepts using Unified Modeling Language (UML) and C++ and Smalltalk.

For the first time, a set of design patterns had been collected, organized, explained, and widely distributed to software developers. Perhaps one of the most significant decisions by the group was to organize these patterns in a highly structured and easily consumable format. Since then, programmers have been able to communicate with each other easily about how they design their software. In addition, they can visually present software design with a universal notation. When one person talks about the Singleton pattern, another person can immediately understand and even visualize in his/her mind how that component works. Isn't that convenient?

Even more surprisingly, design patterns suddenly became the gospel when it come to building good software. In some ways, using them was even perceived as the only way to write good software. GoF patterns were so widely preached across the development community that many people abused them and used them everywhere without good reason. The problem is – When all you have is a hammer, everything looks like a nail! Not everything can be solved or should be solved by the same patterns. When design patterns are overused or misused, the code becomes more abstract, more complicated, and more difficult to manage.

So, what have we learned from the past? We recognize that every abstraction comes with a cost. Every design pattern comes with its own pros and cons. One of the main objectives of this book is to discuss not just the how but also the why and why not, and under what circumstances a pattern should be used or not used. We, as software professionals, will then be equipped with the information we need to make good judgment calls about when to apply these patterns.

More thoughts about GoF patterns

GoF design patterns are classified into three main categories:

• Creational patterns: These cover how to construct objects in various ways. Since object-oriented programming brings together data and behavior, and a class may inherit the structure and behavior of an ancestor class, there are some complexities involved when building a large application. Creational patterns help standardize object creation methods in various situations.
• Structural patterns: These cover how objects can be extended or composed to make bigger thing. The purpose of these patterns is to allow software components to be reused or replaced more easily.
• Behavioral patterns: This cover how objects can be designed to perform separate tasks and communicate with each other. A large application can be decomposed into independent components and the code becomes easier to maintain. The object-oriented programming paradigm requires solid interaction between objects. The purpose of these patterns is to make software components more flexible and more convenient for collaboration with each other.

One school of thought is that design patterns are created to address limitations in their respective programming language. Two years after the GoF book was published, Peter Norvig published research showing that 16 of the 23 design patterns are either unnecessary or can be simplified in a dynamic programming language such as Lisp.

This is not an unimportant observation. In the context of object-oriented programming, additional abstraction from a class hierarchy requires the software designer to think about how objects are instantiated and interact with each other. In a strong, statically typed language such as Java, it is even more necessary to reason about the behavior and interaction of objects. In Chapter 11, Traditional Object-Oriented Patterns, we will circle back to this topic and discuss how Julia works differently compared to object-oriented programming.

For now, we will start with the basics and review some software design principles. These principles are like the North star, guiding us as we build applications.

How do we describe patterns in this book?

If you are new to Julia programming, this book will help you understand how to write more idiomatic Julia code. We will also focus on describing some of the most useful patterns that are already used in the existing open source Julia ecosystem. That includes Julia's own Base and stdlib packages as the Julia runtime is largely written in Julia itself. We will also reference other packages that are used for numerical computing and web programming.

For ease of reference, we will organize our patterns by name. For example, the Holy Traits pattern refers to a specific method for implementing traits. The Domain-Specific Language pattern talks about how to build new syntax to represent specific domain concepts. The sole purpose of having a name is just ease of reference.

When we discuss these design patterns in this book, we will try to understand the motivation behind them. What specific problem are we trying to solve? What would be a real-world situation where such a pattern would be useful? Then, we will ...