Where and how is the term used WRAPPER in programming, what does it help to do?

The term 'wrapper' gets thrown around a lot. Generally its used to describe a class which contains an instance of another class, but which does not directly expose that instance. The wrapper's main purpose is to provide a 'different' way to use wrapped object (perhaps the wrapper provides a simpler interface, or adds some functionality).

The word 'wrapper' can also be used when describing classic design patterns.

Wrapping an object to provide a simplified interface to it is often described as the 'Facade' pattern. The wrapper is the facade.

Sometimes you may have a class which would suit a specific interface, but you can't change the code for it to make it conform to that interface. You can create a wrapper for that class which implements the interface, but which directs most of the calls to the wrapped object. This is the 'Adapter' pattern. The wrapper is the adapter.

The instance you describe, where you have two classes which can do array sorting using different algorithms sounds like the 'Strategy' pattern, where you provide a way to perform an operation on some object, but the algorithm used for that operation may be different depending upon the structure of that object.

For example, one of your sort algorithms may be great for arrays of length less than 100, but then performance might be an issue for it. The other algorithm might be good for bigger arrays. You can create a 'wrapper' for the two algorithms which supports the sort operation but decides which algorithm to use based on the array length.

The vast majority of wrappers exist to hide some sort of complexity.


Simple Explanation by Analogy: What is wrapping?

Code not used to explain.

Example 1: An Ipad "wraps" an Iphone

What does the ipad do that the iphone doesn't do?

Aren't they basically the same thing, except one has a bigger screen?

The Ipad "wraps" the iphone: meaning that you have all the benefits of the iphone WITHIN the ipad itself, AND have the benefits of a larger screen etc. But fundamentally, inside the ipad, lies an Iphone.

What then is different? The only thing that is different is the exterior: you have a larger screen, and perhaps, you cannot make phone calls on the Ipad.

An iphone is the only thing inside an Ipad. The Ipad literally wraps the Iphone. Except you can't make calls with an Ipad because the ability to make phone calls have not been "exposed". Apple want you to buy x2 products not just one :P

It's literally the same thing with "objects".

.............the inside has all the good stuff which is the same, but a different exterior. A different wrapper.

"Wrappers may expose different features than the underlying object" ......okay, but what does this mean?

Sometimes the wrapper might limit the things you can access inside. For example, the ipad may limit your ability to make phone calls, even though the Iphone buried within it, has that capability.

Example 2: Automatic car acting as a wrapper to a manual car

Think of an automatic car, and a manual car. Behind the scenes, in an automatic car, there is some engineering mechanism which automatically changes the gears for you, but fundamentally, under the surface, the car is still a manual car. In other words, the automatic features of the car "wraps" the manual functionality of the car. If you wanted to manually change the gears by yourself in an automatic car - you simply can't do it. The ability to change gears "is not exposed" in the automatic car. But it is exposed in a manual car. Granted the analogy is a little strained, but I hope you see what I'm getting at.

But what is the purpose of a wrapper?

You'd write a wrapping class if you wanted to simplify things. Leave all the complicated bits inside the wrapper, ensuring that those complicated bits are not 'exposed'.

Examples - wrappers used in code

Pysch: It's a ruby wrapper around libyaml, which I believe is written in C. If you wanted to use a libyaml, but with ruby, what are your options? Pysch allows you to write in Ruby, while getting all the benefits of libyaml underneath. It "wraps" libyaml.

.net wrappers: For example, you could have a .net wrapper, which underneath, makes COM calls. If you didn't have a wrapper, then you'd be forced to make those COM calls yourself. Luckily with the wrapper, you'd just make the calls to the .net wrapping code - which would in turn make those COM calls. And that should hopefully simplify things for you.