Regular Expressions Cookbook [Anglais] [Broché]

Although I run the risk of fawning all over this book here, Jan Goyvaerts and Steven Levithan's "Regular Expressions Cookbook (Second Edition)" (O'Reilly, 2012) is a technical text that I will gladly describe using words like "essential" and "indispensable" and "invaluable". It should be on every working programmer's bookshelf, if not on her desk. It is exhaustive and rigorous, covering the major regex flavors across eight popular/widespread general purpose languages. [1] If your work brings you in regular contact with regular expressions, then you need easy access to this book.

To begin with, Goyvaerts and Levithan present an in-depth discussion of each regex feature, starting with the very basics (e.g., making matches against literal expressions) and working up into some pretty sophisticated topics (e.g., writing parsers). True to the title, their approach is a "cookbook" style: a general problem is stated, a solution is presented (or multiple solutions, if that's what it takes), and then they go into an almost painful (but neatly sectioned) level of detail about the solution, describing it token-for-token in some cases. Now, by "neatly sectioned" I mean that their discussion of each solution is broken down by language [2] wherein they are careful to point out flavor- and/or language-specific nuances, quirks, bugs, and/or unique features. They are very careful about this part--if a particular feature does not work in a language (e.g., how JavaScript lacks named capturing groups) then they show you how to work around that deficiency; but perhaps more importantly, if a feature is unique to a language, they point it out as such and caution you against using them (i.e., to keep your regexes general and portable). [3]

Later chapters (i.e., 4 through 9) look at more specific problems--e.g., performing validation on email addresses, [4] dealing with Roman numerals, combing for text in the Apache Common Log Format, or parsing URLs. The recipes are all cross-referenced with each other, so if a particular solution really only solves about 75% of your problem, they're prepared to point you in the right direction. They get right to the point, and then tell you where to go for more. What else can be said about these chapters except that they're like the magnificent arsenal you'll be wishing for when the text zombies swarm at your gate.

All of this makes the "Regular Expressions Cookbook" very skimmable. It is easy to pick it up, find the particular recipe that is going to help you out of a jam, and power through with that solution in hand. Do you "just" need a quick JavaScript solution? Done. Curious how it might compare to the solution in Java or Ruby? No problem. You skim the surface, or you can go as deep as you need [5] on some very narrow and specific sub-sub-subject within the corpus of regular expressions knowledge. (That being said, take their advice and be sure to read the first three chapters so that you *are* properly equipped for those deep dives later on.)

As I said before, if your work regularly brings you in contact with regular expressions, you'll want to arm yourself with this. Highly recommended.

---

[1] Goyvaerts and Levithan define the regex flavors as: .NET, Java, JavaScript, PCRE, Perl, Python, and Ruby; the specific languages covered include: C#, Java, JavaScript (and Levithan's XRegExp library), PHP, Perl, Python, Ruby, and VB.NET. They also have a list in chapter 3 of 11 other languages which--while not specifically covered--are applicable because they adhere to one of the flavors.

[2] I should add "where appropriate" here, and note that the per-language sections in each discussion are much more common in the early chapters (2 and 3, with a pretty sharp drop-off starting in 4). This is because they're covering the fundamentals, and there's a lot more in the way of quirks and nuances to tread lightly around at this point.

[3] In other words: they remind you not to get too clever. "Sure you could do that as a one-liner... but no one's going to know what that means next week. Not even you."

[4] Which, validating an email address is not as easy as it sounds.

[5] Or as deep as you want, if you're in to that sort of thing.

The authors say it best when they say that the book is intended for those who:

"regularly work with text on a computer, whether that's searching through a pile of documents, manipulating text in a text editor, or developing software that needs to search through or manipulate text."

The first three chapters of this book cover useful tools, basic regular expression skills, and programming with regular expressions. Chapters 4 through 9 contain the practical regular expressions recipes. The programming languages that are covered in this book are C#, Java, JavaScript, PHP, Perl, Python, Ruby, and VB.NET. Every recipe that is in the book has solutions and explanations for all eight languages.

The recipes are organized and easy to look up. For example, finding regular expressions that deal with validation, words, lines, URLs, and etc can be easily located in the table of contents. The author's tone is straightforward, direct, and informative. This is not the kind of book where you read from start to finish, but rather a book where you just skim through to find the recipe that you need in whatever given situation.

With all the languages that the book covers, there are just too many languages to put into a single book. I feel as though instead of having a book with all the languages, it would be better to have a book with a single language. However, I must give praise on how well the authors were able to consolidate detailed solutions in 8 programming languages.

For everything that the book is and everything that it covers, the book offers a lot of information at a bargain. If you work with regular expressions and need a reference book, this book is definitely the way to go.

O'Reilly cookbooks are awesome. But just like I don't read the recipes cover-to-cover in regular cookbooks, I don't read all the recipes in the O'Reilly cookbooks either. Also just like regular cookbooks, the day before Thanksgiving is not a good time to open the cookbook for the first time, I at least glance thru all the recipes to know what is there, pick out a few that I can use right away, and dog ear the ones I think I will come back to. So here are the criteria that I review this book with:

1. Easy Navigation: Yep, this book is easy to navigate. If I need to do, say form validation, I know I should start at Chapter 4 "Validation and Formatting".
2. Clear and precise explanation: Yes, I think the explanation are short and precise to the topic of discussion.
3. Pointer for more information: This is hard to do, but the book has a section on "See Also" for correlation between recipes and a general pointer toward 'Master Regular Expression' in the introduction chapter.
4. Easy Reading: Hum.. here is more of a wish list of mine, I wish the book is broken down into different books by language. The book covers these languages, VB.NET, C#, Java, JavaScript, XRegExp, PHP, Perl, Python, and Ruby. I typically skip down to Python and occasionally stop at C# and PHP. The book is over 600 pages and listed at $49.99. I would have been happy to pay 1/5 of the price to get one that just focus on Python, and another 1/5 of the price to get one on PHP.

All in all, it is a good value and a keeper on the bookshelf. But I really think it should be broken down into language-specific cookbook as most reader probably use only one or two languages on a daily basis. With today's print-on-demand, e-book format, I think it would be very minimal work for the author and a whole lot of less skipping for the readers. Just my 2 cents.