When data is input from a file (eg from a magnetic disc) it is transferred one block at a time. Several records may be stored in one block - or one record may need a number of blocks. The block is transferred and stored in a buffer in memory, from where it is processed.

In the same way, when data is to be output to a file, it is stored in a buffer, and is only transferred to the file when the buffer is full.

This process will keep the number of transfers to a minimum and save time - but it needs some memory space for the buffer.

Buffers are also used to overcome the problem of different speeds of peripherals and processors - processors operate thousands of times faster than printers. When data is sent to a printer, it is handled independently of the processor which is left free to carry on with other processing. The data to be printed is stored in a print buffer. This buffer may be in memory or actually in the printer. The hardware takes over and prints the data. When the buffer needs re-filling, an interrupt is sent to the processor.

To speed up processing sometimes two buffers are used (double buffering) - so for example, one buffer can be filled while the other is processed.