Web page performance is literally a death by a thousand tiny little cuts. Usually, the biggest issue for most is the number of requests a page makes. Keep the requests down and your web page performance increases. No magic just common sense. The mobile world we currently live in makes issuing fewer requests that much more important. While it’s true that mobile and tablet devices are getting more powerful each day, they are still no match for the desktop. Moreover, desktops are connected to broadband. Mobile devices are mostly utilizing 3G and 4G.

alt text

Page Performance

Yes there are more, but some page “cuts” include (in no particular order):

  • not using image sprites
  • not using a cookieless domain for images and other static content
  • not using a CDN
  • blocking javascript - worse if it’s in the head
  • not minifying your javascript and CSS
  • not optimizing caching policies
  • not gzipping your content - most if not all popular web servers support some type of compression

And how many third party javascript partners are doing things on your page too? Are they performant and considerate? What exactly are they downloading (and adding to your DOM)? Hopefully, they are working asynchronously. But I digress (you can sense my disliking of third party javascript vendors).

Javascript Performance

Javascript is an important performance factor too. Sure there is the obvious low-hanging fruit “code smells” that you should keep an eye out for like duplicate or unused code. Using tools (from JetBrains, Microsoft, Google, etc.) definitely help spot mistakes, suggest improvements, and avoid common gotchas while helping to enforce team coding standards. Code reviews and unit testing each lends a big time helping hand. Unit tests help to isolate code which is convenient and critical for performance testing and evaluation. Code reviews help you learn because they invite others to look at your code and tell you it sucks (or not). And sometimes as a developer you need to hear how others would tackle a similar problem. Take a breath. Prepare yourself. You do not know everything. Your code isn’t perfect all the time.

Take a breath. Prepare yourself. You do not know everything. Your code isn’t perfect all the time.

Putting on my manager’s hat for a minute it’s hard to justify spending eight developer hours to shave 50 milliseconds. As focused developers, we sometimes get too sucked in. Over doing it by burning hours to shave a few milliseconds may not be worth it unless it’s a focus area or important feature on your site. Utilizing business owners can help drive your focus by guiding you towards these important areas and features.

What about other areas and under-the-covers code that just needs a rework? Just do what I do, and work on it on your own time. Personally, I sometimes dream about a problem, bug, or performance issue that I am grappling. Sometimes I lay awake at 3 AM thinking about it. I just cannot put it down. It eats away at me. Other developers must experience something similar, right?

Bottom line. Try not to overdo it (on company time). At the end of the day, you need to ship it.

I’ve done some javascript micro-optimizations for Zurb that I believe have been worth it. Worth it not only for my own projects but also for others too. Open source contributions benefit the community.

Sure you might think I’m a little crazy and over the top for micro-optimizing javascript. I don’t care. I want a fast, interactive, and responsive user experience. Here are some examples (with JsPerf links) in no particular order that can impact the UI:

Most browsers have optimizations in place to deal with coercion. It’s not really a performance benefit to using === over ==. It’s more to help with readability and understanding for the next guy. Function expressions are sometimes called anonymous, inline, or unnamed functions. They can make debugging a little more difficult because you might see an error (weeks later in a log) that doesn’t explicitly include a name (it does not have one) or appear missing in the call stack (most browser F12 Dev Tools just display “anonymous function”).

Performance Testing & Code Review

Test using older crappy devices. Throttle your network connection. I use an old 1st generation iPad with iOS 5 to test heavy pages with lots of javascript. If your page is halfway decent on an older device it should be fine on the shiny newer ones. And try to regression test by capturing test completion timing results from QUnit and Mocha working with headless PhantomJs. Track, record, compare and then measure results. Establish key performance indicators (KPI). Yes, small shops can do these too.

Here is some code from a last minute code review that you might let normally go out. It’s not really hurting anything. Maybe that’s a good enough reason to leave it alone? Personally, my javascript is far from perfect at times. “Perfect code” is there even really such a thing? Remember when looking at another developer’s code you usually do not know the context when it was written.

Remember when looking at another developer’s code you usually do not know the context when it was written.

Was there a hard deadline? Was there an emergency bug fix or issue? Was it a show stopper? Was it a junior developer or new hire? Or worse a manager or marketing person? Just kidding. Sorry. I just could not resist.

Code under review:

// Using Zurb foundation.
if (Foundation.utils.is_small_only() && $(".top-strip").outerHeight(true) == 2) 
   $(".top-strip").css("padding-top",0)

Code suggested change:

(function () {
   // Using Zurb foundation.
   if (Foundation.utils.is_small_only()) {
      var topStrip = $(".top-strip");

      if (topStrip.outerHeight(true) === 2) {
         topStrip.css("padding-top", 0);
      }
   }
})();

Ignore the outerHeight call (we can discuss another time). The self-invoking anonymous function might be overkill. Although ideally it should be a named function organized within a namespace. The var caching helps and the code is also more readable. But this could just be personal style and not necessarily wrong. What do you think? Are the tiny javascript cuts I mentioned nit-picking? I am positive there are a lot more out there.