Ever wondered what the text editor vi looked like on a Cray supercomputer?
As a core part of the UNIX operating system, used by programmers and system administrators for editing and adjusting text files, the vi editor looked worked and behaves much as it does everywhere else.
Under the covers, however, some clever programming was required in order to make it work. Supercomputers and other large mainframes really do not want to be bothered every time a user types a single character on the keyboard. However, vi and other visual tools require constant updating of the screen image to reflect what has been typed. On an essentially text based machines the terminal driver has to have both a line mode input and character based input. The default for big machines is usually line mode where an entire commandline is constructed before sending off as a single unit, but when switching into a text editor, character mode is used to capture individual characters sent immediately with updates received instantaneously. On a heavily loaded machine where it takes considerable effort to switch between active tasks, if those terminal sessions have been swapped out, extended delays between characters can be expected. Certain machines that had very large memories such as the Cray -2 were particularly good at handling very large numbers of interactive sessions. In the early days of supercomputers, it was encouraged that files and decks of information were prepared on an alternative machine and submitted to a batch process into the supercomputer.
The vi reference card for Unicos the Cray operating systems in use from the introduction of the Cray-2 ( about 1985 ) is shown here.
Download here … SQ-2054 Vi Reference Card
Also available on PC/Mac and many many other systems from Vim.org