Node.js High Performance

Don't rewrite code unless you need to. Take your time to try some available modules, and choose the one that is right for you. This reduces the probability of writing faulty code and helps published modules that have a bigger user base. Bugs will be spotted earlier, and more people in different environments will test fixes. Moreover, you will be using a more resilient module.

One important and neglected task after starting to use some modules is to track changes and, whenever possible, keep using recent stable versions. If a dependency module has not been updated for a year, you can spot a problem later, but you will have a hard time figuring out what changed between two versions that are a year apart. Node.js modules tend to be improved over time and API changes are not rare. Always upgrade with caution and don't forget to test.

Again, you should always split your code into smaller parts. Node.js helps you do this in a very easy way. You should not have files bigger than 5 kB. If you have, you better think about splitting it. Also, as a good rule, each user-defined object should have its own separate file. Name your files accordingly:

    // MyObject.js
    module.exports = MyObject;

    function MyObject() {
      // …
    }
    MyObject.prototype.myMethod = function () { … };

Another good rule to check whether you have a file bigger than it should be; that is, it should be easy to read and understand in less than 5 minutes by someone new to the application. If not, it means that it's too complex and it will be harder to track and fix bugs later on.

The platform is designed to be asynchronous, so you shouldn't go against it. Sometimes, it can be really hard to make some recursive tasks or even simply cycle through a list of tasks that have to run serially. You should avoid creating a module to handle asynchronous tasks, as there are some used and tested by hundreds of thousands of people out there. For instance, async is a simple and very practical way of helping the developer perform better, and the learning curve is very smooth:

    async.each(users, function (user, next) {
        // do something on each user object
        return next();
    }, function (err) {
        // done!
    });

This module has a lot of methods similar to the ones you find in the array object, such as map, reduce, filter, and each, but for iterating asynchronously. This is extremely useful when your application gets more complex and some user actions require some serialized tasks. Error handling is also done correctly and the execution stop is done as expected. The module helps run serial or parallel tasks.

Also, serial tasks that would usually enforce a developer to nest calls and enter the callback hell can simply be avoided. This is especially useful when, for example, you need to perform a transaction on a database with several queries involved.

Another common mistake when writing asynchronous code is throwing errors. Callbacks are called outside the scope where they are defined, and so you cannot just put the callback inside a try/catch block. Therefore, avoid doing this unless it's a very critical error that should make your application stop and quit. In Node.js, throwing an exception without catching it will trigger an uncaughtException event.

The platform has a rule that is consensual for most developers—the so-called error-first callback style. This rule is of extreme importance, since it allows an easier reuse of your code. Even if you have a function where there's no chance of throwing an error, or when you just don't want it to throw and use some kind of error handling inside the function, your callback should always reserve the first argument for an error event if it's always null. This will allow your function to be used with an async module. Also, other developers will be counting on this style when debugging, so always reverse the first argument as an error object.

Plus, you should always reserve the last argument of the function as the callback. Never define arguments after your callback:

    function mySuperFunction(arg1, ..., argN, next) {
        // do some voodoo
        return next(null, my_result); // 1st argument reserved for error
    }

Library functions are another type of module you should use. They help in handling repetitive tasks, and every developer has to perform such tasks. Some of these repetitive tasks can be done with no effort, just by using a library function from lodash or underscore. They are an important part of your code and have good optimizations that you don't even have to think about. Many cycling tasks, such as finding an object in an array based on an object key, or mapping an array of objects to an array of keys of every object, are one-liners in these libraries. Read the documentation first to avoid using the library and not fully using its potential.

Although these kinds of modules can be useful, they can also downgrade performance if they are not chosen well. Some modules are designed to help developers in some tasks, but do not target performance—just convenience. In other words, these modules can help you develop faster, but you shouldn't forget the complexity of each function. Otherwise, you will be calling the same function several times because you forget about its complexity, instead of calling it once and saving the results.

Functions are very important in this platform. This is no surprise since the language is functional and has first-class functions. There are some rules you should follow when writing functions that will make your life easier when debugging or optimizing it later. They also avoid some errors as they try to enforce some common structure. Once again, you can enforce these rules using, for example, JSCS (http://jscs.info/):

Testing your modules is a hard job and is usually neglected, but it's very important to make tests for your modules. The first ones are the hard ones. Look for a test tool that you like, such as vows, chai, or mocha. If you don't know how to start, read a module's documentation, or another module's test code. But don't give up on testing.

After you start adding one or two tests, more will follow. One big advantage of testing your module from the beginning is that when you spot a bug, you can make a test case for it, to be able to reproduce it and avoid it in the future.

Code coverage is not crucial but can help you see how your tests cover your module code base, and if you're just testing a small part. There are some coverage modules, such as istanbul or jscoverage; choose the one that works best for you. Code coverage is done together with testing, so if you don't test it, you won't be able to see the coverage.

As you might want to improve the performance of an application, every dependency module should be looked at for improvements. This can be done only if you test them. Dependency version management is of great importance, and it can be hard to keep track of new versions and changes, but they might give you some good news. Sometimes, modules are refactored and performance is boosted. A good example of this is database access modules.