dd should be able to work from the end of a file just as good as from the beginning

[Dan Jacobson]

Ideas for dd:

Add options for counting from the end of a file, instead of the
only the beginning.

Also, let's say a file has a disk error area in the middle of it,
reading that area will derail whatever program we are using.

We can use cat or whatever, to read up until the error, but what about
the portion after the error? (We don't want to learn C, by the way.)

Therefore, a
$ dd --start-from-the-back-of-the-file-and-send-each-byte-one-by-one-to-stdout |
  dd --start-from-the-back-of-the-file-and-send-each-byte-one-by-one-to-stdout
is needed,
sort of like tac file|tac or rev file|rev might work, but those tools
couldn't be made to work simply blindly on any bytes.