Talk:Read a file character by character/UTF8

Why?

What is the point here?

Specifically: are we trying to disable operating system buffers? Are we trying to disable language-maintained buffers? Do we actually not care if buffers are used? Is this about using sequential file reads to accomplish inter-process communication? (Are we thinking that an indexed file read can simulate the use of a device file?)

For example, let us say we memory map a file - does that count as "reading" the file? Or can we claim that the instructions access the mapped memory represent the "character read"?

(I am inclined to implement this as an abstraction layer on top of "reading the entire file", but wanted to make sure I was clear on the purpose before attempting the implementation.)