Effective STL: 50 Specific Ways to Improve Your Use of the Standard Template Library 1st Edition

"Effective STL" is the final installment of Scott Meyers' "Effective" trilogy. While the first two volumes ("Effective C++" and "More Effective C++") discussed general tips on C++, "Effective STL" focuses exclusively on the STL, which is taken to mean the parts of the C++ standard library that work with iterators. It goes without saying that this book should be read after "Effective C++, Third Edition". One could probably get away with skipping over "More Effective C++", thus going directly from "Effective C++" to "Effective STL".The Good: in typical Meyers style, the writing is generally relaxed and even pleasant, e.g. "This works, but only if you're not terribly picky about what you mean by 'works'." (Item 7). Like in "Effective C++" (but in contradistinction to "More Effective C++") most Items are quite short and therefore easily digestible. Similarly, in this volume, too, Meyers often reiterates the main point of each Item before concluding: this sometimes feels needlessly repetitive, but is mostly great for driving home the essential issues. As in the earlier volumes, things are for the most part organized intuitively. Meyers has a knack for pedagogy and thus for different ways of getting the message across, cf. the few tables that are included in the book, most notably the one on search options in Item 45. The author's organizational prowess is also evidenced by the numerous cross-references and the list of performance-related Items at the end of the Introduction. Moving on to the specifics of the material covered in this volume: Meyers is at his best when addressing a host of topics that elsewhere are often jumbled together. The finest examples of this skill are Items 31, 41, and 45, which deal with sorting options, function adapters, and search algorithms, respectively. Meyers manages to methodically explain the commonalities and distinctions between related concepts. The overwhelming majority of the topics covered in this volume have to do with real-life programming in C++, e.g. how to pass vector and string data to legacy APIs (Item 16). In this vein, some of the Items in "Effective STL" can safely be said to constitute required reading, especially the following on: a) range member functions versus their single-element counterparts (Item 5), b) C++'s most vexing parse (Item 6), c) the swap trick used to eliminate excess capacity (Item 17), d) the erase-remove idiom that allows one to really remove elements from a container (Item 32), e) the superiority of algorithm calls in comparison with hand-written loops (Item 43), and f) the advantages of container member functions over STL algorithms with the same name (Item 44). Other Items feel like digressions (e.g. Item 10 on allocators, Item 18 on vector of bools, or Item 23 on replacing associative containers with sorted vectors) though that doesn't make them any less interesting.The Bad: this volume is not introductory but it isn't a very good reference either. For example, in the early Items Meyers talks about list's splice member functions matter-of-factly. Of course, "Effective C++" and "More Effective C++" are also not introductory or reference works, but there's a major difference: in those books when Meyers discusses a new construct he first explains it briefly. In "Effective STL", on the other hand, he often implicitly assumes the reader will go look stuff up in something like Josuttis' book "The C++ Standard Library" (regarding which Meyers writes: "[e]very C++ programmer should have ready access to a copy"). Obviously, a short hints-and-tips type book cannot contain everything, but the reader's life would have been made much easier had Meyers simply included a few reference tables (containing function signatures) in an Appendix. Moreover, what the Appendices do include is not uniformly interesting: Appendix B addresses developers using Microsoft Visual C++ versions 4-6 and is therefore largely irrelevant today. Another point where "Effective STL" diverges slightly from the earlier volumes' practices is cross-referencing. Meyers mentions other Items extensively in those books, too, but here he's gone overboard: for example, by the time he introduces unary_function and binary_function in Item 40, he has referred to them 10 times already, which probably means that they should have been covered (way) earlier. On a different note, given the non-exhaustive nature of the book, there's important stuff that's missing, e.g. there's absolutely no mention of the at() member function of vectors, deques, and strings. Also, a few of the examples feel somewhat contrived (e.g. anyone paying attention up to that point can tell that the initially proposed solution in Item 42 is obviously unnecessary), though this is more the exception than the rule. Finally, the book's age is starting to show: "Effective STL" was published in 2001 so (despite being more up to date than "More Effective C++") a number of things would be done differently today. For example, Meyers uses for_each in many code snippets, most of which would look much cleaner using lambdas.In a nutshell, I would say that this book is not as indispensable as "Effective C++, Third Edition", but is probably more useful than "More Effective C++". Meyers is a talented author and instructor, so all three volumes are worth reading. Were "Effective STL" to be re-written today (using C++11 and its library) the result would certainly be more interesting, but the advice contained in this book is solid, nonetheless. Overall, four and a half stars.Alex Gezerlis

I'm a professional software engineer. I write code all day long and have lots of experience with C++, but I hadn't used STL much until recently. If you're in a similar situation--decent C++ knowledge but not an STL expert--this book is for you. I haven't even read the whole thing yet, and already I am using patterns from the book to write more effective code.Before I started this book, I thought STL was kind of neat. It had some useful containers. It was nice to be able to use a list or map or string class that had already been tested.Boy, was I underestimating the power of STL. This book has made me a big STL fan, but I'm not reviewing the STL now so I'll leave that topic alone... Thanks to Scott Meyers, I have a much better grasp of the capabilities and limitations of STL. I can use it to do a lot more. I write more concise code that's easier to read and debug. I make better choices about which containers to use. I recognize situations where I can use an STL algorithm instead of many lines of my own code.In short, I look at the STL code I wrote before and laugh... I mean, it all works, but the Meyers book has taken my use of (and appreciation for) the STL to a whole new level. I recommend this book for any C++ developer who isn't already an STL expert.An update, 2 years after the above text was written: I still recommend this book to people and still think it's the best STL book I've read.

I bought this book in a hurry for the purpose of using STL functions for my programming project. It does have some very good and useful advises and tips for dealing with various problems, and is not assuming the reader of it is too much of an entry level programmer.However, my biggest complaint of it is the print layout, including its font/size/color, the font is too heavy and dark (almost all Bold like) and make it harder to read through, especially some C++ code section which should definitely be printed in smaller, lighter and program code like style which is unfortunately not.

I was brought up on STL with C++, and can't imagine the language without it. This book, however, opened my eyes to many of the strengths, weaknesses, and issues of the Standard Template Library. It tells the reader why to use a deque of booleans instead of vector<bool>. More importantly, it provides warnings about features to avoid, as well as how to make better use of this now standard library. Open source software contributors should also check this book out to understand best practices. There's a temptation in C++ to treat iterators like pointers, but Scott Meyers explains clearly why you shouldn't. He briefly discusses the 100 algorithms that are offered in STL, and several key Items, like "Item 30: Make sure destination ranges are big enough" and "Item 32: Follow remove-like algorithms by erase if you really want to remove something."Readers even get some of the back story on the decisions that went into the development of STL and why they made some of the compromises that they did. The book is so clear, so lucid, and so interesting that even the novice C++ programmer will enjoy its prose. It is thoroughly informative even for programmers at all levels of experience with C++.