Tuesday, May 8, 2012

14 Reasons Why Nobody Used Your jQuery Plugin

With so many folks developing jQuery plugins, it’s not uncommon to come across one that just plain – for lack of better words – sucks. There’s no examples or documentation, the plugin doesn’t follow best practices, etc. But you’re one of the lucky ones: this article will detail the pitfalls that you must avoid.

  1. You Aren’t Making a jQuery Plugin
    There are some patterns that are, more or less, universally accepted as “The Right Way” to create jQuery plugins. If you aren’t following these conventions, your plugin may… suck! $ is passed as jQuery; we do it this way to ensure that, outside of the anonymous function, $ can still refer to something else entirely, such as Prototype. Passing the variable for the globally accessible window object allows for more compressed code through the minification processes (which you should be doing, as well).
  2. You Aren’t Documenting Your Code (Correctly)
    Arguably, the most important thing you can do when publishing your code is add the necessary documentation. The gap between what you explain to developers and what the code actually does or can do is the time that users don’t want to waste figuring out the ins and outs of your code. Documentation is a practice that doesn’t have any hard-fast rules; however, it is generally accepted that the more (well organized) documentation you have, the better. This process should be both an internal practice (within/interspersed throughout your code) as well as an external practice (explaining every public method, option, and multiple use cases thoroughly in a wiki or readme).
  3. You Aren’t Providing Enough Flexibility or Customizability
    The most popular plugins offer full access to variables (what most plugins refer to as “options” objects) that a user may want to control. They also may offer many different configurations of the plugin so that it is reusable in many different contexts. For instance, let’s consider a simple slider plugin. Options that the user might wish to control include the speed, type, and delay of the animation. It’s good practice to also give the user access to classnames/ID names which are added to the DOM elements inserted or manipulated by the plugin. But beyond this, they may also want to have access to a callback function every time the slide transitions, or perhaps when the slide transitions back to the beginning (one full “cycle”).
  4. You’re Requiring Too Much Configuration
    Ok, so tip number three suggested that the more complex actions your plugins have, the more complex control that should be available. A big mistake, however, is making too many options required for plugin functionality. For instance, it is ideal for UI based plugins to have a no-arguments default behavior. Certainly, sometimes this isn’t realistic (as users may be fetching a specific feed, for instance). In this case, you should do some of the heavy lifting for them. Have multiple ways of passing options to the plugin. For instance, let’s say we have a simple Tweet fetcher plugin. There should be a default behavior of that Tweet fetcher with a single required option (the username you want to fetch from). The default may, for instance, grab a single tweet, wrap it in a paragraph tag, and fill the selector element with that html. This is the kind of behavior that most developers expect and appreciate. The granular options should be just that: options.
  5. You’re Mixing External CSS Rules and Inline CSS Rules
    It’s inevitable, depending upon the type of plugin, of course, that you will have to include a CSS file if it is highly based on UI manipulations. This is an acceptable solution to the problem, generally speaking; most plugins come bundled with images and CSS. But don’t forget tip number two – documentation should also include how to use/reference the stylesheet(s) and images. Developers won’t want to waste time looking through your source code to figure these things out.
  6. You Don’t Offer Examples
    The proof is in the pudding: if you can’t provide a practical example of what your plugin does with accompanying code, people will quickly be turned off to using your plugin. Simple as that. Don’t be lazy.
  7. Your Code Doesn’t Match Their jQuery Version
    jQuery, like any good code library, grows with every release. Most methods are kept even after support is deprecated. However, new methods are added on; a perfect example of this is the .on() method, which is jQuery’s new all-in-one solution for event delegation. If you write a plugin that uses .on(), people using jQuery 1.6 or earlier will be out of luck. Now I’m not suggesting that you code for the lowest common denominator, but, in your documentation, be sure to explain which version of jQuery your plugin supports. If you introduce a plugin with support for jQuery 1.7, you should strongly consider maintaining support for 1.7 even once 1.8 comes out. You should also consider taking advantage of new/better/faster features in jQuery as they come out.
  8. Where’s the Changelog?
    Along with keeping your jQuery version support/compatibility a part of your documentation, you should also be working in version control. Version control (specifically, via GitHub) is largely the home of social coding. If you are developing a plugin for jQuery that you want to eventually publish in the official repository, it must be stored in a GitHub repository anyway; it’s time to bite the bullet if you haven’t learned how to use version control. There are countless benefits to version control, all of which are beyond the scope of this article. But one of the core benefits is that it allows people to view the changes, improvements, and compatibility fixes you make, and when you make them. This also opens the floor for contribution and customization/extension of the plugins you write.
  9. Nobody Needs Your Plugin
    Ok, we’ve ignored it long enough here: some “plugins” are useless or too shallow to warrant being called a plugin. The world doesn’t need another slider plugin! It should be noted, however, that internal teams may develop their own plugins for their own uses, which is perfectly fine. However, if you’re hoping to push your plugin into the social coding sphere, find a reason to write more code. As the saying goes, there’s no reason to reinvent the wheel. Instead, take someone else’s wheel, and build a racecar. Of course, sometimes there are new and better ways of doing the same things that have already been done. For instance, you very well might write a new slider plugin if you are using faster or new technology.
  10. You Aren’t Providing a Minified Version
    This one is fairly simple: offer a minified version of your code. This makes it smaller and faster. It also ensures that your Javascript is error free when compiled. When you minify your code, don’t forget to offer the uncompressed version as well, so that your peers can review the underlying code. Free and cheap tools exist for front-end developers of all levels of experience.
  11. Your Code is Too Clever
    When you write a plugin, it is meant to be used by others, right? For this reason, the most effective source code is highly readable. If you’re writing countless clever one-liner lambda style functions, or your variable names aren’t semantic, it will be difficult to debug errors when they inevitably occur. Instead of writing short variable names to save space, follow the advice in tip number nine (minify!). This is another part of good documentation; decent developers should be able to review your code and understand what it does without having to expend too much energy.
  12. You Don’t Need jQuery
    As much as we all love using jQuery, it is important to understand that it is a library, and that comes with a small cost. In general, you don’t need to worry too much about things like jQuery selector performance. Don’t be obnoxious, and you’ll be just fine. jQuery is highly optimized. That said, if the sole reason why you need jQuery (or a plugin) is to perform a few queries on the DOM, you might consider removing the abstraction entirely, and, instead, sticking with vanilla JavaScript, or Zepto.
  13. You’re Not Automating the Process
    Grunt is a “task-based command line build tool for JavaScript projects”. It allows you to do things like this:
    grunt init:jquery
    This line (executed in the command line) will prompt you with a set of questions, such as the title, description, version, git repository, licenses, etcetera. These pieces of information help to automate the process of setting up your documentation, licensing, etc.
  14. You’re Not Testing
    Oh, by the way – you do test your code, right? If not, how can you ensure/declare that your code works as expected? Manual testing has its place, but, if you find yourself refreshing the browser countless times every hour, you’re doing it wrong. Consider using tools, such as QUnitJasmine, or even Mocha. Testing is particularly useful when merging in pull requests on GitHub. You can require that all requests provide tests to ensure that the new/modified code does not break your existing plugin.
If you are writing a jQuery plugin, it is vital that you stray away from the pitfalls listed above.
Original Article : 14 Reasons Why Nobody Used Your jQuery Plugin

0 comments:

 
back to top