RE: A basic question on encoding Latin characters

From: Frank da Cruz (fdc@watsun.cc.columbia.edu)
Date: Tue Oct 05 1999 - 10:38:03 EDT


> A question, Frank. Whenever I use terminal emulators or a console window, I
> go through just the process that you describe, but I type an ENTER key after
> each line. When the protocol requires that, then it can use final combining
> marks without a problem. Are there lots of situations where both:
>
> a. The command is grabbed without an ENTER key (or equivalent).
> b. The command allows an accented character as the last item.
>
There are two ends to this problem:

 a. The host issues a prompt that does not end with a newline or other
    character or sequence that is not a "starter character". The login
    and password prompts, shell prompts, and the prompts of interactive
    applications. The latter, especially, can end with any character at
    all -- the number of applications is without bound, as is the
    imagination of their developers. Not to mention the fact the prompts
    can be customized by the user.

 b. The host, or an application on the host, accepts commands that do not
    end with a newline or other non-starter. To cite a few: editors like
    emacs and vi; newsreaders like rn and trn; "file browsers" like more,
    less, and info. Plus any application that accepts arrow or function
    keys from a terminal, plus any application that says "press any key
    continue". The list goes on, and it's not restricted to UNIX.

Even in the GUI world, as many have pointed out already, we commonly see
single-letter shortcuts in dropdown menus. Suppose a German application
has:

  []ffnen
  [O]rdnen
    
and the application is reading characters (not keycodes). The application
either waits forever for a combining diaeresis that will not come, or it
gets a false positive on the 'O' when in fact '' was typed.

Again, we can imagine numerous protocols and conventions for avoiding such
situations, but in telecommunications they work only when we have control
of both ends, and this is rarely the case.

- Frank



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT