Ralph Alexander wrote:

> There are 7 German special characters:
> 
>  äÄöÖüÜß

Actually, this kind of information would do us a great deal of good if we 
knew the special characters associated with each language group.

We're programming up the special character key to be F11. Each time the key 
is pressed, you move through a state transition table associated with a 
finite-state machine that we've programmed into QCTerm. The consequence of 
repeatedly pressing the F11 key, if you were on a "e", would be this sequence:

    unmarked e -> e w/accent -> e w/grave -> e w/caret -> e -> w/diaeresis
         -> unmarked e

The longest sequence is for the "o":

     unmarked o -> accent -> grave -> caret -> diaeresis -> tilde -> slash
          -> oe -> masculine ordinal -> unmarked o

The shortest tend to be the few marked consonants, such as the "n":

     unmarked n -> n w/tilde -> unmarked n

The way that we've programmed this up tends to heavily favor the romantic 
typist. Spanish virtually always works with one key press, and because I 
fleetingly know Spanish and can type it with some speed, the single press F11 
key is a very fast and very easy way to mark the few characters that need 
altering. French, because it also requires the use of the grave and caret, 
may require two or three pushes of the F11 key, but at least the diacriticals 
always appear in the same sequence and it doesn't seem all that much worse.

However, Boris suggests that the German or Scandanavian typist will be slowed 
by this approach. Ted Ashton wrote me privately last night and suggested that 
we allow the user to select his language group and start the state transition 
loop at different places for different languages. While I dismissed his idea 
initially, I've reconsidered.

By letting the user select his language as German, we could change the user's 
first F11 key press to bring up the diaeresis (umlatted) characters. All of 
the characters would still be there, in the circular loop; it's just that a 
different starting place would be defined, perhaps in this manner:

     unmarked o -> diaeresis -> accent -> grave -> caret -> tilde -> slash
          -> oe -> masculine ordinal -> unmarked o

so that a German typist could still type Spanish, Czech, or Swedish, but the 
German characters would be the first on the list, requiring the least number 
of keystrokes.

Another person wrote me and mentioned that:

   In Danish, we have 3 special letters, æ ø å upshifted to Æ Ø Å
   The Norwegian language uses the same diacriticals except for the AE.

   The Swedes use the German model, using the two dots over the a
   and the o (ä ö / Ä Ö if visible?) PLUS the Danish/Norwegian
   double a (å / Å)!

   In Finland they use the Swedish alphabet except for the double a
   (å / Å).

It's this kind of information, language-by-language, that would do us a great 
deal of good. We could let you set your language in Terminal Preferences and 
make it so that a native typist would not have to press the F11 key more than 
once or twice for your specific language. 

I would greatly appreciate it -- if your language isn't German, Swedish, 
Danish, Norwegian, Finnish, Spanish or French -- if you would e-mail me the 
special characters that are used in your language, preferably by frequency if 
one letter has multiple diacritical markings.

Wirt Atmar

PS: We've also been asked so far if we were intending on programming up 
Arabic, Hebrew, Chinese, Japanese, Russian and Greek. The answer is 
unfortunately no. We're simply sticking to the superset of ISO Latin-1 that 
is the standard Windows encoding.