Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Image

Book Review: Sams Teach Yourself Node.js In 24 Hours 112

Michael Ross writes "Since its introduction in 1994, JavaScript has largely been utilized within web browsers, which limited JavaScript programmers to client-side development. Yet with the recent introduction of Node.js, those programmers can leverage their skills and experience for server-side efforts. Node.js is an event-based framework for creating network applications — particularly those for the Web. Anyone interested in learning this relatively new technology can begin with one of numerous resources, including Sams Teach Yourself Node.js in 24 Hours." Keep reading for the rest of Michael's review.
Sams Teach Yourself Node.js in 24 Hours
author George Ornbo
pages 464 pages
publisher Sams Publishing
rating 7/10
reviewer Michael J. Ross
ISBN 978-0672335952
summary An introduction to the Node.js framework.
This book, authored by George Ornbo, was released by Sams Publishing on 15 September 2012, under the ISBN 978-0672335952. The recent publication date is promising, because Node.js is evolving rapidly, thus gradually obsoleting books written not that long ago. On the publisher's page, visitors will find a brief description of Ornbo's book, a few customer reviews, the table of contents, a sample chapter (the 14th, "A Streaming Twitter Client"), and links to purchase the print and electronic versions of the book. There is also a link to the companion site, which offers some of the same content as Pearson's page, but also has a link to download an archive file containing all the example code, nicely organized.

The book's material spans 464 pages, and is organized (shoehorned) into 24 "hours" (chapters), grouped into six parts. The first two chapters in "Getting Started" explain how the reader can download Node.js, create a "hello world" web server program, install new modules using npm (Node Packaged Modules), search for modules, locate documentation on them, and indicate module dependencies for an application. Unfortunately, the blocks of source code presented in the first examples (Listing 1.1 and Figure 2.2) are not explained in the narrative (until the fifth chapter) or even commented. Readers would likely appreciate some clues as to the nature of http.createServer, req, res, the "underscore" module, etc. — especially at the beginning of their journey. If readers are not expected to understand these details at this point, then they should be told so, to avoid any concerns that such an understanding is assumed in the subsequent chapters. The author does not explain where Node.js is installed or what changes it makes to the terminal's default path variable. On page 18, the term "project folder" is unclear: should the "underscore" module end up in hour02/example01/node_modules, or nodejs/node_modules, or nodejs/node_modules/npm/node_modules? Only later is this (partially) answered.

Chapter 3 demonstrates the complexity that arises from concurrent input/output in networked applications. This material should arguably have been presented at the beginning of the book, to better establish the purpose of Node.js, and the value to the reader of studying it. The next chapter summarizes jQuery and JavaScript callbacks, and then provides a helpful discussion of how Node.js uses the latter. The author contends that the asynchronous paradigm of Node.js is unsuitable for long-running processes, but does not explain why this is true, which would have provided some substantiation for the claim.

The second part of the book, "Basic Websites With Node.js," encompasses four more chapters. The first one discusses how to: create a simple server (using the core HTTP module), examine the response headers (generated for web pages, in different browsers and on the Linux command line), execute 301 redirects, respond to different types of requests (using the URL module), and create a simple client. Oddly, the author does not explain or even mention the sizable JSON output — the first line of which is "{ domain: null," — displayed in the reader's server terminal when the web page pointing to that server is refreshed or when the "curl -I" command is run. The next two chapters cover how to build websites using the Express framework, and are likely the first point where the reader will see some of the real-world complexity of Node.js. The eighth chapter explains how to persist data between calls to the application, including files, environment variables, and MongoDB.

Debugging, testing, and deploying are all critical topics for any application development, and are covered in the third part of the book. The author illustrates three methods of debugging: STDIO, a core module, is a lightweight method for debugging Node.js code; it allows one to output messages to the console, check the value of any variable or literal, and track function calls and responses from third-party services. Node.js provides access to the more powerful debugger of V8 (the Google Chrome JavaScript engine), which supports breakpoints and code stepping. Node Inspector, compatible with WebKit-based browsers, provides all of the above functionality, and more. The next two chapters present several modules that ease the important process of creating full-coverage tests, and demonstrate how to deploy applications to any one of three Node.js-capable cloud hosting providers (Heroku, Cloud Foundry, and Nodester).

Having covered the basics of Node.js, the author begins the fourth part of his book with two chapters that show how to use Socket.IO, WebSockets, and Express to build real-time web applications. These techniques are illustrated in the development of a chat server as well as a nickname management and messaging system. The aforementioned sample chapter extends these techniques further in working with the Twitter API to consume its real-time data, push it to the browser, and show results in a dynamic graph. This section is wrapped up with coverage primarily of JSON — specifically, how to create, consume, and send JSON-structured data.

APIs were addressed briefly in the previous section, but are explored much more deeply in the subsequent five chapters. Readers may initially conclude that the discussion of processes is elementary, but the author then shows how one could utilize that knowledge to interact with Node.js scripts, including detecting script exits and errors, sending signals and arguments to a script, generating child processes if needed, and sending messages among them. In the 18th chapter, the author goes into greater detail about Node.js's Events module, best practices, and how to generate event listeners dynamically. The buffer API may be low-level, but it is essential for storing raw binary data, as opposed to the Unicode-encoded strings that JavaScript uses within a browser. The Buffer and Stream modules are presented with plenty of helpful examples.

The last part of the book addresses miscellaneous topics, starting with CoffeeScript (a JavaScript precompiler). While CoffeeScript affords numerous benefits, it is not clear why it would deserve an entire chapter in a book dedicated to Node.js. In the next chapter, readers learn how to verify their Node.js code, add command-line executables, and then package it all up into portable modules that can be contributed to the npm registry or GitHub. The last two chapters explain how to create and configure middleware using the Connect module, and how to use Backbone.js (a front-end JavaScript framework) in conjunction with Node.js to build browser-based web applications.

Each chapter concludes with a summary (invariably a waste of space), a Q&A section, a workshop comprising quiz questions (with the answers presented immediately below it, for almost instant spoiling), and several exercises for the reader.The index at the end is missing several of the important topics discussed in the text.

The book contains many errata: "EBay" (page 1; should read "eBay"), "OSX" (page 9; presumably Mac OS X), "yaml" (page 15; should read "YAML"), "irc" (19), "led to [a] great deal" (27), "to solve Concurrency" (37), "process" (54; should read "processes"), "try and" (55; should read "try to"), "This goal" (56; should read "The goal"), "how [a] class" (56), "You will [see] the" (62), "status of [a] web server" (70), "javascripts" (77), and "then [the] name" (87). At this point, less than 20 percent into the book, it was clear that the copyeditors had done a sloppy job, so I stopped recording these flaws that should have been caught. Those first four errata suggest that "textese" is even pervading the world of technical publishing. (Strangely, there does not appear to be a place on the publisher's website for reporting errata.)

The production team should have been looking for places to cut down on the heft of the print edition. The "Try It Yourself " sections sometimes duplicate what is found in the regular text nearby — especially in the third and fourth chapters. For instance, three sets of HTML markup are repeated, as well as the surrounding discussion (pages 42 through 47).

In general, the text does not appear to have been carefully scrutinized by technical reviewers and copyeditors. Occasionally the reader is given critical information later than would be optimal, e.g., the "Watch Out" warning on page 18, provided after the reader installs a module. The writing style is noticeably awkward in countless places in the book, including several run-on sentences. (Technical authors should not be bashful in using commas when doing so would help readability.) Also, the text is littered with too many exclamation marks — as if that is going to make any narrative more exciting.

In terms of the production quality of the book, a lay-flat binding would have made it much easier to read when using both hands on the keyboard. Also, in my review copy (kindly provided by the publisher), a disappointingly large number of the pages had small black splotches of ink; fortunately, none made the text unreadable.

On the other hand, Node.js is certainly not a simple subject area, and this book is able to convey a lot of information about it. This book's forte is the extensive use of example code to illustrate the concepts being presented. Incidentally, kudos to the author for inviting the reader to contribute to the Node.js community, such as adding new modules to GitHub or updating the documentation of existing modules. Overall, readers new to Node.js would certainly benefit from working their way through this volume.

Michael J. Ross is a freelance web developer and writer.

You can purchase Sams Teach Yourself Node.js in 24 Hours from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.

*

This discussion has been archived. No new comments can be posted.

Book Review: Sams Teach Yourself Node.js In 24 Hours

Comments Filter:
  • Re:Benefits (Score:5, Interesting)

    by Kergan ( 780543 ) on Monday December 10, 2012 @05:37PM (#42246213)

    What are the benefits of using javascript on the server side over a more typical server side programming language?

    You get to write every possible thing you need to use for your app, including basic things that aren't functional in the handful existing (albeit very dysfunctional) libraries.

    *Ducks*

    In all seriousness, there's a sweet spot when you're writing some app that needs a real-time, state-full, non-blocking environment, such as a chat server. There are of course alternatives, but this is what node.js was designed for.

    It does a good job at it, too, as long as you don't mind the tons of callbacks (read: impossible to follow spaghetti code), and the lack of decent debugging tools (unless this changed since last year?), the scarce (and occasionally unmaintained, and frequently buggy) libraries related to much about anything useful.

    My (admittedly scathing) take on node.js is this: it was created by a someone with a specific need, lots of love of javascript, and time to spend on reinventing the wheel instead of using more mature tools for the job. Many credits to him for being such an enthusiastic hobbyist who then open sourced his work. The issue is, it then got picked up by an army of web designers, as opposed to actual programmers. They thought "Oh great I can now code server stuff too!" and went off to write a library or three. And it shows, because nearly all of the libraries in question are borderline unusable except in very specific cases.

    In practice, node.js is a horror story littered with writing your own libraries for much about everything because testing existing ones revealed that they suffered from either or all of a missing feature, too many bugs, no maintainer, etc. As a hobbyist and potential library contributor, go there; you're much needed, and you'll even have fun if javascript turns you on. For actual development, just don't -- not even with a 10-foot pole.

  • Re:Benefits (Score:5, Interesting)

    by VoidEngineer ( 633446 ) on Monday December 10, 2012 @06:05PM (#42246479)
    It's actually a really huge advantage, if done correctly. A trivial example: I'm working on an app that has a file import functionality, so I wrote a half dozen helper functions to parse a spreadsheet into JSON. Later wound up using those same functions on the server to implement an XML interface. Business logic is the same in both instances, and having a single language for both server and client allows that business logic to be stored in a single place, maintaining the Don't Repeat Yourself (DRY) rule of thumb. Having a single language on both server and client is a godsend for keeping your app DRY.
  • by epine ( 68316 ) on Monday December 10, 2012 @07:31PM (#42247179)

    I would think that the resurgent interest in functional programming would ameliorate this to some extent but most of the JS I've worked on is crap and I would strongly recommend against using it unless you have a clear picture of why you would want to do so.

    I'm taking an online course on MongoDB. This caused me to look into JavaScript and Node after a long hiatus concerning the web application stack.

    We're deep into the territory of people confusing a tool with the culture of its use. JavaScript appears to be a fairly decent language of Scheme derrivation, minus one horrible blunder on scoping rules (to be addressed, perhaps in the worst possible way, in the next language standard). I'm given to understand that many of the horrendous web applications out there involve heaps of disorganized JavaScript maintaining application state via DOM edits. What could possibly go wrong? Is the language at fault, or the culture of its use? Hint: one of these answers takes more than five seconds to justify, so you might find it time efficient to run with the first answer that comes into your head. No one ever got fired for spewing blame in the widest possible arc.

    Node.js itself is a strange beast. There's something to be said for language plasticity, of not having to prematurely bind code to server or client. Adopting an entire world-view to make this possible is the domain of fools and zealots. Today's zealots are tomorrow's gurus. Gurus don't grow on trees, after all.

    I write a fair amount of code in R, also from the Scheme family tree. Are there any people out there who think that JavaScript sucks, but R doesn't? Wow. I'd be interested to hear that argument. No wait, with an infinite number of monkeys, all languages suck, by the simple statistical principle that natural languages are denser in idioms of derision than admiration. It's almost as if our languages know something about us. Hey, no one ever got fired for spewing blame in the widest possible arc.

    I've said this before and I guess it's my fate to keep repeating this observation. There's a litmus test across the language like/despise spectrum of attitude toward one's fellow humans. If you hate your fellow humans, you most favour languages that offer facilities of orthodoxy, control and prevention. On the other hand, if you believe that your project is so difficult that only by the miracle of collaboration and teamwork can the project be completed at all, then you tend to value languages with the fewest expressive impediments, whether the nail-gun interlock works or not.

    I've read many times about small teams doing big things with C++ and remarking about the language's unbearable legacy of self-harm "we're professionals: we rarely have a problem with it". If you're an ambitious code monkey and the coder beside you working on the same project was fathered by a dangling pointer, you really hate C++. No power under the sun can prevent the dangling pointer from making you also look bad.

    I would certainly use Node.js, but with great caution. It won't maim your appendages like C++, but it will paint you into a corner (a certain asynchronous world view) if you don't understand its long term limitations. This is why I ultimately rejected my online MongoDB course. It taught you enough to get yourself into trouble (deep trouble), but not enough to get yourself back out again.

With your bare hands?!?

Working...