1. Unless the underlying field(s) is > 90% unique, a table scan will take place when SQL performs a WHERE. Is that statement a correct generality? If so, I will eliminate many indexes I have that are coming nowhere near > 90% unique. That is, a lot of repetition in the data means no index is helpful.
2. Is there any net plus for a clustered index? I assume the clustered index requires a lot of shuffling of the data if that is required getting the rows in their correct clustered order. If this is the case, the clustered index does create a lot of overhead. That would indicate the clustered index should be fields that are commonly used in the WHERE statements to justify the overhead?