The new property "unique" is a great initiative. Unfortunately it's still better to do you own coding to ensure unique data on non-clustered fields. The reason for this is that when the user get's an error because of a unique secondary key the only way to get around it is to refresh the page and the record is lost. When you do the same thing in code and you show an error message the user only has to change the value of the field and the record can be saved preserving all other data on the record.