Название: Regular Expression Puzzles and AI Coding Assistants: 24 puzzles solved by the author, with and without assistance
Автор: David Q. Mertz
Издательство: Manning Publications
Год: 2023
Страниц: 147
Язык: английский
Формат: pdf (true)
Размер: 10.2 MB

Learn how AI-assisted coding using ChatGPT and GitHub Copilot can dramatically increase your productivity (and fun) writing regular expressions and other programs.

Regular Expression Puzzles and AI Coding Assistants is the story of two competitors. On one side is David Mertz, an expert programmer and the author of the Web's most popular Regex tutorial. On the other are the AI powerhouse coding assistants, GitHub Copilot and OpenAI ChatGPT.

Here's how the contest works: David invents 24 Regex problems he calls puzzles and shows you how to tackle each one. When he's done, he has Copilot and ChatGPT work the same puzzles. What they produce intrigues him. Which side is likelier to get it right? Which will write simple and elegant code? Which makes smarter use of lesser known Regex library features? Read the book to find out.

David also offers AI best practices, showing how smart prompts return better results. By the end, you'll be a master at solving your own Regex puzzles, whether you use AI or not.

This is a book for programmers with a little bit of experience in some programming language. As every reviewer of the proposal antecedent to this book commented rather vociferously, and as every programmer of nearly every programming language recognizes at once, of course every computer programmer and software developer uses regular expression already in the course of their daily work. I do not expect to introduce a wholly new concept to many of my readers. Or at least I do not expect to introduce that concept to them.

Within this book, where code other than regular expressions themselves are shown, it is generally within the Python programming language. This is especially true of those puzzles where I present “AI thoughts” that I have solicited from GitHub’s Copilot or OpenAI’s ChatGPT. It turns out that these tools currently often balk at a request to “write a regular expression to do such-and-such” but are generally happy to comply with a request to “write a Python program using regular expressions” to do the same task.

About the technology
Groundbreaking large language model research from OpenAI, Google, Amazon, and others have transformed expectations of machine-generated software. But how do these AI assistants, like ChatGPT and GitHub Copilot, measure up against regular expressions—a workhorse technology for developers used to describe, find, and manipulate patterns in text. Regular expressions are compact, complex, and subtle. Will AI assistants handle the challenge?

About the book
Regular Expression Puzzles and AI Coding Assistants is the perfect starting point for programmers of any experience level who want to understand the capabilities—and the limitations—of these exciting new tools. Author David Mertz presents 24 challenging regex puzzles, their traditional human-made solutions, and the fascinating answers given by popular AI assistants. Alongside these eye-opening puzzles you’ll learn how to write prompts, integrate AI-generated coding suggestions, and interact with the assistant to get the results you want. By the end of the book, you'll have a clear understanding of where AI assistants can reliably write code for you and where you’ll still need a human touch. Plus, you’ll learn a lot about regular expressions!

Who should read this book
This is a book for playful programmers, and those who wish to expand their understanding and rethink their assumptions (albeit about a few relatively small matters). While the tutorial in the appendix suffices to bring you from a complete naivete about regular expressions to a reasonable understanding, your motivation to understand and work with this tool should precede your enjoyment of this text.

But this is also a book for the millions of software developers who have read intriguing and exuberant discussions of the promise of AI coding assistants, and who perhaps have already begun working with them. These tools have a lot of promise, and will have a growing future utility; they also have limitations that will not abate completely as the technologies improve. Using tools to aid in coding is wonderful, and understanding the scope and limitations of your tools is even better.