As explained in the chapter SSL, a hash is created by applying an algorithm that combines the characters of a message in some way to produce a string that is meaningless but nevertheless can be regarded as characteristic of the message - the chances of any other message, subjected to the same algorithm, giving the same hash are tiny.

It is neither necessary nor desirable that the hash should be unique; the algorithm should be such that the hash does not contain enough information to reconstruct the original message. So there actually will be other messages (though a tiny subset of the set of all possible messages) that would give the same hash. This important property of a good hashing algorithm is called irreversibility. Nevertheless, it should be difficult to deliberately create two messages that produce the same hash. This property is called collision-resistance.

Just as with encryption, there are various well-known algorithms used for hashing, and your SSL software may let you choose which it uses.