systems read chucks of data in various ways from the media:
fixed length 'records'
variable length 'strings'
whole 'blocks' of data
if the program were to request 80 chars from the media, wait for it to arrive,
process and then repeat for the next 80, you will suffer repeated WAITS
and get slow processing.
if the system or program asks for whole blocks from the media and then
moves thru 80 chars at a time, you suffer only one wait per block.
the latter example is a 'buffered read'