Change "ctype" to "dtype" in the "buffer" methods? #178
Description
When I created the methods buffer_read(), buffer_read_into() and buffer_write() in #72, I gave them a ctype parameter to specify the data type used in the buffers. The possible values of this parameter were motivated by the corresponding function names in the libsndfile API, which contain them as substrings: 'double', 'float', 'int', 'short'.
Looking back at this now, I think it would probably have been more consistent from an API perspective, to call the parameter dtype and allow the options 'float64', 'float32', 'int32', 'int16'.
Should we change that?
If we decide to go that way, we could at first support both parameters and deprecate the old ctype parameter. After some time, we could remove it completely.