Glenn writes:

> Ugh.  Workable, but hardly "typist-friendly."

>  I'm not convinced this would help.

>  I don't know which DS app was tried, but chances are it wasn't
>  DS Word 97.
>  
>  Fortunately, DS has improved the technique over the years, so that
>  now there's actually some thought behind it.  While it always takes
>  me several aborted attempts to determine which menu option is needed
>  (it's Insert -> Symbol...), once I find the char (say, o-umlaut)
>  it is relatively easy to remember for when I need it again (in the
>  short term).  In this example, it's Control+: o (control-colon oh,
>  where the colon is like a tipsy umlaut).
>  
>  Of course, diacriticals have been on the Mac since day 1, but you
>  knew that. :)  Their philosophy is slightly different.  For example,
>  since the tilde [~] is "most often" used over 'n', the you get this
>  diacritical with Option+n n.  To get the tilde over other characters
>  where it makes sense, you still start with Option-n.  Use "Key Caps"
>  under the Apple menu for experimenting.  Note the wide border displayed
>  around some keys when pressing Option.

>  For the intended use of diacriticals in the AICS app, this may well
>  be true.  But for more than extremely-casual use, it would be a pain
>  (IMHO, of course).

To which Ewart added:

> I'm inclined to agree with Glenn (and others) that the AICS
>  methodology will infuriate most touch typists and is best for casual
>  use only.
>  
>  Another method which seems not to slow typists down is that
>  developed by Prof Donold Knuth at Stanford for his typesetting
>  program TeX.  Here all diacritcals are introduced by the "\"
>  character and apply to the next character, thus \"o places an
>  umlaut over the o and \~n puts ~ over the n. It allows \i to get the
>  "dot-less" i (spanish? I can't remember).  Typists seem to like this
>  approach because  they don't have to stop or even slow down.  I
>  don't have my TeX books to hand but I'm certain that similar "\"
>  commands are used for various ligatures and non-english
>  characters.

To which Ferenc added:

> This F11 trick is a typical twisted up English or American solution. It is
>  unacceptable for Hungarian (and Polish, Czech, ...) typists, too, because
>  the the accented characters are very frequent in these languages.

To which, I can only reply: Absolute Nonsense.

Lordy, it appears that you have to break a few eggs to make an umlaut.

There will be four ways to generate a marked character glyph in QCTerm. They 
are:

     o The highly praised AICS method
 
     o The use of a character map palette

     o The use of the ALT+numkey four digit number method

     o The use of a nationalized keyboard

The bottom three work for all Windows-based products that use ISO Latin-1 
encoded character sets. If you are unfamiliar with these bottom three 
techniques, see: http://www.bris.ac.uk/Depts/Education/ml/accents.htm and 
then try these methods with a very simple program such as Notepad.

The original complaint was that we didn't support Windows-extended ISO 
Latin-1. Rather, we encoded the terminal's display function and line drawing 
sets into the upper register of the ASCII character map. That was done to 
significantly increase the speed of screen renderings (and it worked very 
well), but it was probably a mistake, one that we're now correcting. When 
that process is completed, the bottom three methods will also work with 
QCTerm.

I am very familiar with all of the other methods mentioned of imposing 
diacriticals and foreign marks on text. Indeed, while Jim was here, I showed 
him the three basic Microsoft ways of marking characters (the bottom three 
mentioned above, but using only printouts of the European nationalized 
keyboards as we have none on the premises). I also showed him Word 97, 
including Glenn's "improved" method. And I showed him an HP terminal's method 
[and mentioned that a Macintosh is philosophically identical to an HP 
terminal, albeit differently implemented (we have plenty of Macs, we just 
didn't get up and walk to the next building to view one)]. I am also familiar 
with Professor Donald Knuth's method. I am also very familiar with 
PostScript's method, which is the method that we use in QueryCalc.

Of all of the methods listed above, the one we're putting into QCTerm is the 
one that I like the best by far. I am a touch-typist and I can type Hungarian 
at high speed with the best of them (so long as it's not a nationalized 
keyboard).

It is true that we originally biased the cycle-through list towards the 
romantic languages, but that's being changed as we speak. The method we're 
adopting requires only one modifier key; you know where it is -- and if you 
were a touch typist, you should be able to blaze new trails through smoking 
keyboards.

While I would have preferred to use a modifier key in the lower-left of the 
keyboard, there are no keys available for us to use there. The control key is 
already doing double duty and the alt key is similarly well spoken for.

Moreover, this method won't require the memorization that every other method 
requires. Most foreign glyphs can be generated by a SINGLE additional 
keystroke, in this manner:

     o  a vowel followed by F11 -> most common mark in chosen language. In 
Spanish, French and Italian, that will be an accent. In German, that will be 
a diaeresis, etc. 

     o  consonants followed by an F11 will similarly be appropriately 
modified, most often only by one alternate character. N's become Ñ's, M's 
become mu's, S's become ß's and Š's, and T's become eth's and thorn's (to 
support the small but increasingly raucous Icelandic contingent of HP3000 
users).

     o  all of the currency symbols appear with the $. A $ followed by an F11 
gets you, in this presumed order of frequency-of-use, $ -> euro -> lira/pound 
-> cent -> yen -> florin.

     o  all of the commercial symbols appear with the @. A commercial at 
followed by the appropriate number of  F11's gets you @ -> copyright -> 
registered mark -> trade mark.

     o nationalized punctuation marks are similarly done. ? followed by F11 
-> ¿, ! followed by F11 -> ¡, and so on.

I have typed for years using every method mentioned so far, and QCTerm's is 
not only the fastest, it's also by far and away the easiest to remember. If 
you want to advocate using some other method, compare in a dry-run this 
method to the one you're advocating (use the inverted ?'s, for example).

Wirt Atmar