Re: Much better Latin-1 keyboard for Windows

From: Alain LaBonté (
Date: Fri Jul 23 2004 - 07:38:36 CDT

  • Next message: Philippe Verdy: "Re: CGJ for variant combining marks, in Hebrew as well as German?"

    À 17:16 2004-07-22, Michael Everson a écrit:
    >I've never understood this keyboard philosophy. Its "groups and planes"
    >terminology just doesn't make sense to me (as someone who has designed
    >keyboard layouts for well over a decade). I like good old-fashioned
    >dead-keys and four keyboard states (plain, shift, alt, and alt-shift.

    [Alain] There is no « plane » at all in ISO/IEC 9995. This is ISO/IEC
    10646 terminology, which also has a term called "group", but it is not the
    same thing (and yet, you do not find the notion of plane, group, row and
    cell complicated while it is indeed multiple enough to make it more
    difficult to remember). I think you did not try hard to understand the
    concept of keyboard groups, even if I have explained it to you many times (^;

    The keyboard group concept is not new (a decade is relatively new in the
    world of keyboards, and the notion is a bit older than that), it was
    designed before ISO/IEC 10646 (as early as 1988) and it is not difficult to
    understand (you should have tried, as you heard about it before designing
    keyboards, as you say (^; )... I too, designed keyboards, since more than
    two decades¹, and I also have written keyboard drivers implementing group
    selection on PCs, as soon as I heard about the layer shifting concept.

    A group is just a keyboard layout of up two 3 levels (in general only 2, as
    for the US keyboard).

    The concept of group and group selection (called "layer shifting" by its
    two designers, Dr Umamahwesaran and one of his IBM colleagues, in 1986, if
    my memory is good) was taken into consideration by ISO with the intent to
    extend it to multiple groups. However the multiple group model, if it
    exists, has not been standarized yet and deployed fully in its modalities,
    but time may have come for this. For this we must rely on international
    standarization, not on the will of only one individual (everybody has ideas
    about keyboards, as I hardly learned since I began to work on our Canadian
    keyboard standard in 1985 -- it is a prowess to come to consensus on
    keyboard issues, but we did it in Canada [we adopted our standard
    unanimously, after long "battles"], and internationally, with success [also
    after long "battles"] -- however it needs everybody to try to understand
    each other's ideas and integrate them in harmony).

    >>With UNICODE/UCS now of age, this in our opinion would be highly
    >>desirable to go beyond international standardization of the Latin script
    >>support limited to some languages as now.
    >[Michael] Please see the specification of the Irish Extended keyboard for
    >Unicode, at

    [Alain] Every layout can be considered a group in the ISO model. What is
    lacking is standardization (taking all platforms into consideration) in
    what you write.



    ¹ the first keyboard driver I developed was on an 8K (yes, 8192 octets of
    RAM only, not one more) Commodore PET, in machine language (6502 processor,
    I had to make my own assembler program before, and it too had to fit within
    8K) where I had to care in real time about the row and column of the wires
    intersecting each key switch, to determine the keys that were being hit...
    Nowadays with PCs, the keyboard microprocessor does only this, and just
    sends a code (called scan code) to the main processor indicating that a key
    well identified has been hit (no need for the PC to watch in real time,
    since al this is put into a buffer before an interruption signal is sent to
    the PC). I made my first PC keyboard driver in 1982, a few months after the
    first IBM PC had been released with an Intel 8080 processor under the hood
    (August 1981 [they made an Assembler program at once, and I bought it

    This archive was generated by hypermail 2.1.5 : Fri Jul 23 2004 - 07:42:17 CDT