Can anyone point me to a good reference to help me understand typed datasets?

My understanding to this point is that a typed dataset simply has an XML Schema associated with it. I'm doing this at runtime by using the ReadXmlSchema() method of the dataset.

I'm a little vague as to the utility and advantages of using typed datasets, and it would seem that maintenance of the XML Schemas for each dataset used in a complex application could become a bit of a headache. I want to learn more.

Thanks.