While 'proof' with regards to alcohol is "equivalent to .5 % alcohol content by volume," that definition does not match the historical origin of the term. Also, proof, as a measure of alcoholic strength is really only used in the United States now. Elsewhere, such as in Canada, a simple measure of % alcohol by volume is used.
Prior to 1816, alcohol strength was determined by whether or not a mixture of the spirit and gunpowder would explode when ignited with a match. If the mixture exploded, then there was "proof" that the spirit contained a significant amount of alcohol. If it didn't explode, the spirit was deemed weak. This method of testing is reminiscent of the Middle Ages where ale was tested by having someone sit in a pool of it until it dried up and checking whether the unfortunate person conducting the test had their pants stuck to the wooden board they were sitting on.
Proof, and similar measures, are interesting insofar as they represent the attempts of societies before the advent of government testing and standardization to ensure the quality and contents of a product. Originally, brands played an important role in ensuring quality and consistency of a product; when a producer had invested money and reputation in a brand, they had a strong incentive to ensure a high level of quality.
Source: The Case for Brands. The Economist
. Sep 6th 2001