KMeans and Canopies

[Mark]

Should canopy generation and KMeans clustering typically use the same 
distance calculation or is possible to mix and match? Any reason why 
some would mix?

Thanks