In "How fast are transfers between the C64 and the 1541 floppy drive?" the OP describes an application that needs fairly fast transfers between the C64 and an attached drive, and this application probably also wants to minimize the amount of CPU time it spends on transfers on the C64 side. (The application is to have the drive do co-processing, so each cycle used for communication on the C64 side directly reduces the advantage to be gotten from this coprocessing.)
He mainly discusses using the KERNAL routines, however, and his last sentence is:
If I use the KERNAL routines, what kinds of speeds can I anticipate?
There are a couple of other options for faster and cheaper (in cycles) communication that would suit this application, but both involve not using the KERNAL. For an answer giving an overview of these, is it is better to put the answer on this question or to write a new question explicitly asking about that and put the answer on that new question?