I'm Brett Slatkin and this is where I write about programming and related topics. You can contact me here or view my projects.

28 October 2019

I wrote a new edition of my book

This week I announced that Effective Python: Second Edition is now available for preorder. It ships in mid-November this year (mere weeks away!). The new edition is nearly twice the length of the previous one, and substantially revises all of the items of advice in addition to providing 30+ more. I hope you like it!



This second edition was easier to write in some ways and much harder in others. It was easier because I'd written a book before. I know how to put words on the page everyday and keep up the pace of writing. I've been through the whole process of delivering drafts, incorporating technical reviewer feedback, working with various editors, etc. And I was able to reuse the literate programming toolchain that I built in order to write the first edition, which saved a lot of time and allowed me to focus on the words and code instead of the process.

What was most difficult about writing the second edition of the book was how much Python has evolved in the interim. The first edition of the book covered Python 2.7 and Python 3.4. Python 2 will be dead and deprecated in the next 3 months, so I needed to remove all mention of it. For Python 3, there were major changes to the language in versions 3.5 (async/await, unpacking generalizations, type hints), 3.6 (f-strings, __init_subclass__, async generators/comprehensions, variable annotations), 3.7 (asyncio improvements, breakpoint), and this month with 3.8's release (assignment expressions, positional-only arguments).

Each time a new version of Python 3 came out, I had to gain experience with its features and reconsider my previous advice. For example, one of the items I'm most proud of in the first edition ("Consider Coroutines to Run Many Functions Concurrently") is now obsolete. My advice in the second edition is to completely avoid these features of generators (send and throw) in favor of the new async/await syntax from 3.5. However, I'm not sure I would have reached that conclusion had it not been for the improvements to the asyncio module made in 3.7.

The same type of interdependent changes seemed necessary for almost all items of advice in the book, which meant I had to potentially redo everything. I find that the hardest part of writing is editing (e.g., "If I had more time, I would have written a shorter letter."), so this was quite a challenge. I repeatedly delayed the start of my writing sprint so I could learn more about the evolving features and wait for them to mature. Luckily, I felt that the language finally stabilized with version 3.7, and 3.8 presented usability improvements that are worthwhile instead of major structural changes. Thus, the new edition is done and here we are.

Today's Python 3 is a very different language than version 3.0 that was released over 10 years ago. I wonder how the language will continue to grow over the coming years. There are some especially exciting efforts going on related to CPU performance, such as PEP 554 (subinterpreters) and tools like the Mypy to Python C Extension Compiler. I look forward to what comes next (Python 4?), and the potential opportunity to write a third edition of the book (many years from now).
© 2009-2019 Brett Slatkin