Why is there an offset in the Java String implementation?

Question

I've been looking at the implementation and I don't understand why there is an offset. I assume it is important.

I'm taking an Algorithms course taught by Sedgewick, and we're talking about Strings now. In lecture he briefly discussed the String implementation, but he doesn't go over why there is an offset (Note, if lectures were not online, I would definitely have asked).

It seems when one makes a String that within the implementation, it is given an offset, and I can't seem to understand why one is needed. Even for substring purposes I don't quite follow why you would have an offset. For example, apparently if you create a string "David", it is really ['X', 'X', 'D', 'a', 'v', 'i', 'd', 'X'], or something of that nature, where it is offset by the 'X's. Why is this?

Answer 1

This can be useful in cases where strings need to be derived from another longer string, akin to substring().

In this case the same (immutable) backing array may be used, while adjusting the offset and length, to save memory and optimize performance.

This is no longer the case in JDK7.

Answer 2

The offsets typically are present as an optimization. For example, in the character array related items, you can use the same character array multiple times to construct multiple Strings from the array. This is because you would use the same array with different offsets and lengths.

This is a very effective optimization because it prevents the need to construct new character arrays solely for the purpose of having the character you wish to start with at the zero index. For example, after reading in a block of bytes (perhaps from a socket), you could chunk it up into String sized messages without having to do unnecessary byte copying.