An Example Of Overloaded Functions With Very Different Sub-Function Implementations

By Ben Nadel on September 14, 2010

Tags: Javascript / DHTML

A few days ago, I blogged about overloading Javascript functions with a sub-function approach. In that post, I demonstrated that branching logic could be factored-out into the core function in such a way that the individual sub-functions could execute with a very narrow set of concerns. Such a separation might seem overkill when the difference in arguments is not terribly influential. But, sometimes, a difference in invocation arguments can lead to a completely different execution. I wanted to take a quick moment to consider such a scenario in order to look at why separating branching logic is important.

I recently demonstrated how to use jQuery's animate() method to power easing-based iteration. Now, I want to take that concept and extend it to use argument-based branching logic in which the "duration" argument of the ease() method becomes optional. In this overloaded method signature, the duration-based easing will use the animate() implementation (as it did before) while the non-duration-based iteration will use a standard FOR-loop.

<!DOCTYPE html>
<html>
<head>
<title>Overloading The Ease Method Based On Arguments</title>
<script type="text/javascript" src="./jquery-1.4.2.js"></script>
<script type="text/javascript">
// I am the easing iteration funciton. This is built on top
// of the core animate function so that it can leverage the
// built-in timer optimization.
jQuery.ease = function( start, end, duration, easing, callback ){
// Check to see how many arguments we have. If we only
// have four, then we are using standard iteration. If
// we have five, we are using duration-based iteration.
if (arguments.length == 4){
// Standard iteration.
return(
jQuery.ease.iterate.apply( this, arguments )
);
} else {
// Duration-based iteration.
return(
jQuery.ease.iterateWithDuration.apply( this, arguments )
);
}
};
// I am the standard iteration method for easing.
// Essentially, I am just a FOR-LOOP with a given
// callback to execute per iteration.
jQuery.ease.iterate = function(
start,
end,
easing,
callback
){
// Keep track of the iterations.
var stepIndex = 0;
// Use the difference between the start and end values
// in order to mimic duration. We are going to assume
// that the the FOR-Loop we are using has a +=1 nature.
var duration = Math.max(
Math.abs( end - start ),
1
);
// Keep looping until the iteration index equals the
// end index.
for (var i = start ; i != end ; true ){
// Execute the callback for this iteration.
callback(
i,
stepIndex++,
start,
end
);
// Update the index value for our easing.
i = jQuery.easing[ easing ](
(stepIndex / duration),
stepIndex,
start,
(end - start),
duration
);
}
// Execute the callback once for end value (which would
// not have executed in the FOR-LOOP).
callback( i, stepIndex, start, end );
};
// I am the duration-based iteration method for easing.
// I use jQuery's animate() method to power the easing
// over the given time frame.
jQuery.ease.iterateWithDuration = function(
start,
end,
duration,
easing,
callback
){
// Create a jQuery collection containing the one element
// that we will be animating internally.
var easer = $( "<div>" );
// Keep track of the iterations.
var stepIndex = 0;
// Set the start index of the easer.
easer.css( "easingIndex", start );
// Animate the easing index to the final value. For each
// step of the animation, we are going to pass the
// current step value off to the callback.
easer.animate(
{
easingIndex: end
},
{
easing: easing,
duration: duration,
step: function( index ){
// Invoke the callback for each step.
callback(
index,
stepIndex++,
start,
end
);
}
}
);
};
// -------------------------------------------------- //
// -------------------------------------------------- //
// When the DOM is ready, init the scripts.
$(function(){
// Get a reference to the body tag.
var body = $( "body" );
// Get a reference to both lists.
var durationList = $( "#durationList" );
var iterateList = $( "#iterateList" );
// Iterate withOUT druation.
//
// NOTE: We are putting the non-duration iteration first
// because it is processor intensive and will mess up
// the animation-based duration (making it less smooth).
$.ease(
1,
100,
"swing",
function( i ){
// Create the list element for this iteration.
var listItem = $(
"<li />",
{
text: i,
css: {
backgroundColor: "#CC0000",
color: "#FFFFFF",
width: ((i * 10) + "px")
}
}
);
// Add the list item to the list.
iterateList.append( listItem );
}
);
// Iterate with druation.
$.ease(
1,
1000,
1300,
"swing",
function( i ){
// Create the list element for this iteration.
var listItem = $(
"<li />",
{
text: i,
css: {
backgroundColor: "#CC0000",
color: "#FFFFFF",
width: (i + "px")
}
}
);
// Add the list item to the list.
durationList.append( listItem );
}
);
});
</script>
</head>
<body>
<h1>
Overloading The Ease Method Based On Arguments
</h1>
<h2>
Easing With Duration (Animation)
</h2>
<ol id="durationList">
</ol>
<h2>
Easing Without Duration
</h2>
<ol id="iterateList">
</ol>
</body>
</html>

As you can see in the above code, I am using both forms of easing-based iteration to build an Ordered List (OL). Each list item in the ordered list is set to a width that is proportional to the index of the overall iteration. When we run this code, we get the following page output:

The difference between these two approaches (duration and non-duration) is rather substantial; one executes over a given time period (duration) whereas the other executes immediately. This behavioral difference requires a completely different implementation behind the two different invocations. Because of the difference of implementation, it is critical that the branching logic be factored-out of the primary execution function. If it were not, the primary execution function - ease() - would become large and unmaintainable.

I'm sure it's easy to look at this example and argue that these two different implementations shouldn't be two branches of a single method - that they should actually be two entirely different methods (ie. ease() and easeOverDuration()). That might be true; but, this post isn't really about easing - it's a case-in-point of how factoring-out branching logic based on arguments creates clean, cohesive functions with narrow execution concerns.

Tweet This Groovy post by @BenNadel - An Example Of Overloaded Functions With Very Different Sub-Function Implementations Woot woot — you rock the party that rocks the body!

Reader Comments

Nelle Sep 15, 2010 at 8:52 AM

27 Comments

Why not wrap it:
Function = function(){
}
Function.prototype = {
implementations: []
, call: function () {
if (typeof this.implementations[arguments.length] == "function")
{
this.implementations[arguments.length].apply ( null, arguments );
}
}
}

And then it could be set up with:
easing = new Function();
easing.implementations = [ null, null, null, iterate, iterateWithDuration ];

and called with:
easing.call( "arg1", "arg2", "arg3" );
or
easing.call( "arg1", "arg2", "arg3", "arg4" );

Ben Nadel Sep 15, 2010 at 1:03 PM

14,309 Comments

@Nelle,

You could do it that way; ultimately, we're both doing the same thing - factoring out the branching logic such that the "worker" functions can have a very narrow set of concerns. In either case, I think we're on the right track.

Oh my chickens, this post is old!

Hit me up on Twitter if you want to discuss it further.