Comments on: Arduino Pointers and the Ping for Automated High Five http://www.uchobby.com/index.php/2009/08/17/arduino-pointers-and-the-ping-for-automated-high-five/ Microcontrollers Electronics Hobby Tue, 24 Jan 2012 09:58:33 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: signal7 http://www.uchobby.com/index.php/2009/08/17/arduino-pointers-and-the-ping-for-automated-high-five/comment-page-1/#comment-2159 signal7 Tue, 18 Aug 2009 14:19:06 +0000 http://www.uchobby.com/index.php/2009/08/17/arduino-pointers-and-the-ping-for-automated-high-five/#comment-2159 While this is a good introduction to pointers, it doesn't mention the best feature of using pointers: arithmetic. If you declare an array of type int, what you are really doing is creating a pointer to a location in memory that consists of a sequence of 4 byte blocks (because int's are typically 4 bytes to most* compilers). You can then reference the elements of the array using simple addition and subtraction. array[3] is the same as *(array + 3). I know that might seem confusing, but the compiler knows the type of the array and also knows that each element is 4 bytes so when you add some number to the pointer, the compiler internally computes array + (3 * 4) to arrive at the correct address. This might not seem useful - I'll grant you that. However, when you start using arrays of structs or other more complicated data structures, this will be holy grail of understanding the usefulness of pointers. Maybe it's overkill for the arduino environment, but with these AVR parts coming with more and more memory, I'm sure more data processing applications will be coming. * I say most compilers because there is no guarantee that an int is 4 bytes and you can't really depend on it being 4 bytes if you want to write reliable code. If you really need a 4 byte int, use int32 instead. While this is a good introduction to pointers, it doesn’t mention the best feature of using pointers: arithmetic.

If you declare an array of type int, what you are really doing is creating a pointer to a location in memory that consists of a sequence of 4 byte blocks (because int’s are typically 4 bytes to most* compilers). You can then reference the elements of the array using simple addition and subtraction. array[3] is the same as *(array + 3). I know that might seem confusing, but the compiler knows the type of the array and also knows that each element is 4 bytes so when you add some number to the pointer, the compiler internally computes array + (3 * 4) to arrive at the correct address.

This might not seem useful – I’ll grant you that. However, when you start using arrays of structs or other more complicated data structures, this will be holy grail of understanding the usefulness of pointers. Maybe it’s overkill for the arduino environment, but with these AVR parts coming with more and more memory, I’m sure more data processing applications will be coming.

* I say most compilers because there is no guarantee that an int is 4 bytes and you can’t really depend on it being 4 bytes if you want to write reliable code. If you really need a 4 byte int, use int32 instead.

]]>