A 04:30 97-12-15 -0800, Kent Karlsson keka@im.se a écrit :
>(I'm not sure if this is off-topic or not. But since Alain copied his
>reply to the Unicode list, I suppose it is not too off-topic.)
>
> Your reply makes me worried.
>
> On my Windows (NT 4) (Swedish) keyboard I get a "level 3" as you
>describe. This behaviour I find highly undesirable, since that is not
>what I want. This "level 3" is very sparsely populated with any
>characters, and (plain) shift together with "alt Gr" gives nothing.
>Likewise, no characters are generated via the other "alt" key. This is
>next to useless! I get hold of only 9 characters via this "level 3".
>And I have to switch keyboard mapping or use the char map utility to
>write e.g. Ø.
[Alain] :
You illustrate that we were wise to change terminology.
"Alt Gr" is Level 3 Select, i.e. an additional shift, giving access to more
than 96 characters at a time on all national keyboard layouts throughout
Europe and elsewhere. It has a potential of at least 330 characters if
ISO/IEC 9995-3 is implemented (and more if more than 2 groups are
implemented, athough ISO/IEC 9995-3 does not go beyond 2 groups).
"Alt" has nothing to do with *character data* entry. It is a problem.
Terninology has been kept for the real, initial, and original function of
ALT, which is intended to bring proprietary functionality for the keyboard,
not standardized character data entry.
"Plain shift" used together with any other *function* key is not defined in
ISO/IEC 9995. That is why you get nothing, and that is normal.
I open a parenthesis, you'll see at the end of the paragraph where I lead
it. Group Select, to go beyond, has a symbol and a function, but no
international consensus was never achieved on how to actuate the function.
There are implementations which go from very ergonomic to do that (one key
only, and a toggle à la CapsLock, i.e. you can go for a walk before doing
aytning else, i.e. a function that can even be used by handicapped persons
who only have one hand) to very anti-ergonomic (things which require your
two hands plus a pedal and, who knows, vocal input at the same time (; ).
On my keyboard, the Group Select is of the first kind: it is done by
actuating a key on the PC which has the same scan code as the RIGHT CTRL
key on an American keyboard (and there is a fallback which corresponds to
your expectation of "ALT GR" used in conjunction with the right "plain
shift" to use your terminology).
[Kent] :
> On my MacOS (8) (Swedish) keyboard I get a quite different
>behaviour, which I have interpreted as a "group 2". Both "alt" keys are
>equivalent and they work together with the ordinary shift keys (I stick
>to that terminology for the time being...). So, e.g. , shift-ö
>generates Ö, alt-ö generates ø, alt-shift-ö generates Ø. In addition
>both alt-levels are fully populated with characters. This is the
>behaviour that I want (even though the character assignments could
>sometimes be slightly different). This is a much more useful behaviour!
>
> /Kent Karlsson
[Alain] :
You bear judgement values that are not shared by a lot of people (I respect
your opinion though). Many ergonomists believe that you should not have to
actuate 2 function keys at once in addition with a third key to enter a
character. A pedal would be better. But that said, we can get used to
everything, and once the habit acquired, what is more comfortable when no
habit is taken can be perceived as less comfortable because another way has
become automatically done with precision by our subconscious mind. This is
normal, and also part of ergonomy. One knows, for example, that what is
very stressful for people who are not interested by what they do can be
stressless for others concentrated on an activity which they love, even if
it seems painful to others.
What is important is that both PCs (multiple keyboard implementations, btw,
many of which are based on ISO/IEC 9995), and Macs, can accomplish the same
thing regarding character data entry. Of course the Mac is more consistent
from one machine to another, because its method is proprietary. However
there is one ISO/IEC 9995 implementation at least on the Mac, the one used
in French Canada for all Macs, with some problems with one key for which
the set of national in international layouts lead to at least 5 characters:
3 levels in group 1 and 2 levels in group 2 -- even in combining function
keys to implement the two groups using 4 levels, the Mac can't do it so
far, a pedal is missing (; .
ISO/IEC 9995-3 allows me to enter, on a PC, with no more than 2 keys being
depresseed at a time, *all* of Latin 1, plus OE ligatures, plus Y DIAERESIS
(which are not in Latin 1, as we all know, a problem in DOS, for example,
or under OS/2, or under UNIX systems) under Windows 95 or NT (in one case
using CP 1250, in the other case using UNICODE coding, but I don't need to
know the coding!), and that can be done in all environments in a way that
is code-independent. The Mac in Canada allows me to do that as well with
the same method that Mac users use in other countries for
actuatingalphanumeric keys, although Apple had to modify the ISO/IEC 9995-3
prescription for one key and potentially others if they wanted to implement
all the 330 characters (or so) that ISO/IEC 9995-3 provides, because the
4-level approach is limitative, while being unergonomic up to a certain
point (although we get used to everything, so it is not a big issue).
That said, I want to applaude the Mac even if I am a PC user. The merit of
the Mac in Canada was that it was the first major company to implement
Canadian standard CAN/CSA Z243.200 keyboard standard in 1988, then its
revision in 1992 (this standard incorporates ISO/IEC 9995-3). So we were
tolerant on its slight deviation for one key becuase they did great. Except
for this it is fully conformant to ISO/IEC 9995-3. And that is the only
keyboard you can get on a Mac if you want to work in French in Canada. All
Mac users have but praise for this keyboard, regardless of the source of
the opinion, and that is the reason why Apple kept it to replace its
previous keyboard (pre-1988).
Now the revision of ISO/IEC 9995-3 for the EURO SIGN begins to populate
LEVEL 3 of GROUP 2 with the CURRENCY SYMBOL, *in case one still requires
this symbol*. And the proposed change is the choice of both Apple and IBM,
not forgetting DEC, not mine. I am just the editor of the standard. I had
proposed something else. But I was instructed to produce the FPDAM that you
have seen. It seems reasonable in an international context.
Alain LaBonté
Québec
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:38 EDT