Responsive design – a simpler approach

If you’ve spent a bit of time working on responsive websites, you’ll have hit two problems or frustrations …

1. Scrolling up and down your css file comparing values between media query sections
2. Using mediaListeners, or checking body width just to work out what media query / display mode your page is currently using.

Inspired by libraries like moderniser or templates like html5 boilerplate that places classes in the head or body tag, I thought it would make life easier if you could place a class on an element (usually the body tag but any element is supported) and then go forth and style and code as usual.

The result is the mediaQuery plug-in for jQuery. If you want to use this plug-in you need to be comfortable with the following: you are already going to use jQuery, and you don’t mind your site being non-responsive for non-javascript users.

On the positive side, the plug-in includes polyfills for css media queries and the matchMedia and addEventListener javascript functions – so you wont need to include any other code to support responsive across all browsers.

Usage is really simple:


$('body').mediaQuery({
  w320: 'only screen and (min-width: 320px)',
  w480: 'only screen and (min-width: 480px) ',
  w768: 'only screen and (min-width: 768px)',
  w1024: 'only screen and (min-width: 1024px)'
});

This code will add the classes w320, w480, w768, w1024 to the body tag depending on each of the corresponding media queries matching.

Your css is then as simple as this eg:


.w320 .content p {
  background-color:#d45354;
}

.w480 .content p {
font-size:3em;
width:360px;
height:360px;
line-height:360px;
background-color:#818b85;
}

Checking for different media queries is now as simple as this line of jQuery code:


if ($(element).hasClass('w1024')) ...

You can view an example of the technique here: http://jameswestgate.github.com/mediaQuery/examples/

Responsive design – a simpler approach

Altering image sizes responsively in all browsers

If you view the responsive web page page by Ethan Marcott in IE8, you’ll notice that it doesn’t work in the same way as modern browsers. This is because IE8 and below doesn’t support media queries – the smart bit of css we use to conditionally style our page based on current or device width. We also need to evaluate media queries to set our img tag through javascript using windows.matchMedia and media query list addListener– this isn’t supported in IE9 or less.

We need another polyfill – to add media queries to IE8 and below and to add media query matching and listening in IE9 (as well as any other unsupported platforms). The responsive polyfill I’ve written combines elements of Scott Jehl’s respond.js polyfill for media queries, Paul Irish’s matchMedia.js polyfill and my own code for media query listeners whilst keeping performance a priority.

This allows us to use code to detect when to change the image or image size in the browser and set the correct image path on initial page load. We thus take the following steps:

– set up a media query to detect changes at a width of 600 pixels
– read the contents of noscript tags and add an image to the dom depending on the media query
– listen to the media query so that when it changes we can modify the image path accordingly.

var mql = window.matchMedia('only screen and (max-width: 600px)');
var elements;
	
//Get contents of noscript tags 
var tags = document.getElementsByTagName("noscript");

//Add image to dom depending on media query
for (var i=0, len=tags.length; i<len; i++) {
  var markup = tags[i].textContent.trim();

  if (mql.matches) markup = markup.replace("/large/", "/small/");
  tags[i].insertAdjacentHTML('afterend', markup);
}

//Scale up to large / down to small when media query changes
mql.addListener(function() {

  if (!elements) elements = document.getElementsByTagName("img");
		
  var from = (mql.matches) ? "/large/" : "/small/";
  var to = (mql.matches) ? "/small/" : "/large/";

  for (var i=0, len=elements.length; i<len; i++) {
    if (elements[i].className.indexOf("responsive") > -1)
      elements[i].src = elements[i].src.replace(from, to);
  }
});

View the example

Thoughts on this technique

It should be pointed out that strictly speaking, once we have the large image we dont need the small image - it is better just to scale it down. However I also think that the smaller image in a responsive design may often be different - perhaps a more cropped version - anyway this page exists to demonstrate the basic technique only.

We also dont use a jQuery style document.ready type approach - this is simply not needed for such a simple example - placing the code in the footer is more than sufficient. However - because the images are loaded fairly late in the day you can see the image 'waterfall' as they are loaded. A more advanced technique may only display them once they are fully loaded, perhaps with css such as

img.responsive {display:none}
noscript img.responsive {display:block}

Is it worth the extra script vs the reduced download? Probably. Certainly the noscript code is small. Making IE9 and below fully responsive may not be worth it however.

PS. The responsive script doesn't currently read style tags in the parent document - it could easily be modified to do so however.

Altering image sizes responsively in all browsers

Progressive enhancement of NOSCRIPT tags containing images

For the 2nd of our 4 part plan to solve the problem of responsive images using current browser standards, we need to put our noscript textContent polyfill to work. What better than to refactor the seminal responsive design original from A List Apart by Ethan Marcotte.

To start with, all we do is move the 6 character images into noscript tags, add our polyfill ot the bottom of the body and run a small amount of cross browser script to append the content of the noscript tag after itself. You can see the results here.


<script type="text/javascript" src="js/noscript.min.js"></script>
<script type="text/javascript">
var tags = document.getElementsByTagName("noscript");
for (var i=0, len=tags.length; i<len; i++) {
  var tag = tags[i];
  tag.insertAdjacentHTML('afterend', tag.textContent);
}
</script>


 
This provides us with the hook we need to dynamically modify image paths before they are added to the DOM.

Progressive enhancement of NOSCRIPT tags containing images

Read noscript tag content reliably in all browsers

The first step in our 4 part plan to solve the problem of responsive images using current browser standards is to reliably access the content inside a noscript tag. Modern browsers support access to the content of the noscript tag through either the innerHTML property or the textContent property. Since innerHTML is encoded, and is also read-only in IE<9, and textContent seems to contain markup perfectly in supported browser, it is the property we will polyfill.

The first job is to feature detect the property, which is done simply with this piece of code:

var div = document.createElement("div"); 
div.innerHTML = "<noscript><span></span></noscript>"; 
var elements = div.getElementsByTagName("noscript"); 
if (elements.length && elements[0].textContent) return; 

 
If we need to add the property, we have to resort to retreiving the markup of the page and parsing it with a regex. I know this may be frowned upon, but in this case it is quick and fairly reliable. In this case we also need to path the XmlHttpRequest object, and then retrieve the markup of the current browser location via ajax.

patchHttpRequest(); 
var xmlhttp = new XMLHttpRequest(); 
xmlhttp.onreadystatechange = function() { 
  if (xmlhttp.readyState === 4 && xmlhttp.status === 200) { 
    var regex = /<script (.|n)*?>(.|n)*?</script>/igm; 
    var tags = document.getElementsByTagName("noscript"); 
..

 
Finally, we loop through each noscript tag, and set the textContent property.

var results = [], response = xmlhttp.responseText; 
var i=0, len=tags.length;
while (i < len && results) {
  results = regex.exec(response); 
  tags[i].textContent = results[1]; 
  i++; 
} 

 
Visit the github page to view the code packed together with some dom and ajax helper code.

Read noscript tag content reliably in all browsers