Speech recognition is a technology that turns spoken words into written ones, and stores them in a digital document such as a word processing file. Although around for decades, it is only fairly recently that it has become mature enough to enter the mainstream. Previously, speech recognition systems had to be "trained" to recognize a particular person's voice even for the most limited applications. Today, the algorithms driving the systems have become sophisticated enough to accept voice commands without training. Some of the more common uses recently include voice commands that drive telephones to do simple tasks; voice commands that help a phone system route calls appropriately; simple data entry; and easy completion of highly structured documents. But the thing that most people think of as true speech recognition – where they can simply talk naturally and have their every word transcribed accurately in real time – still requires a substantial investment in time, effort – and patience.