Re: Latin-script keyboard layout (was RE: Quick Question About Korean Input Methods)

From: Doug Ewell (
Date: Sun Jan 03 2010 - 20:51:21 CST

  • Next message: David J. Perry: "Re: Latin-script keyboard layout (was RE: Quick Question About Korean Input Methods)"

    Marc Durdin <marc dot durdin at tavultesoft dot com> wrote:

    >> 4. can be implemented using Microsoft Keyboard Layout Creator (no
    >> more than 4 shift states; Ctrl+char is not useful)
    > Why point 4? That restricts your options considerably -- with MSKLC
    > you can only implement your requirements with deadkeys and modifiers
    > (shift states). Neither of these are great solutions. Deadkeys are
    > unwieldy and definitely not intuitive, modifiers apart from Shift and
    > AltGr tend to cause conflicts with applications.

    I don't mean to cast any aspersions on Keyman, which I have heard for
    many years is a fine product.

    By mentioning MSKLC, which is just the means to the end, what I meant to
    say is that I can install the keyboard layout on basically any Windows
    PC I may need to work on, without installing an additional executable
    software package. And I can give it to others who also want to be able
    to type "special characters" (I hate that term) without installing new

    I'm well aware that modifiers apart from Shift and AltGr (notably Ctrl)
    are problematic, which is why I said 4 shift states (none, Shift, AltGr,
    Shift+AltGr). If I have misused the terms "level" and "shift state," I
    apologize; hopefully my meaning is clear.

    Karl Pentzlin <karl dash pentzlin at acssoft dot de> wrote:

    > DE> that:
    > DE> 1. is based on the U.S. English layout and does not redefine any
    > DE> of its Level 1 or 2 keystrokes
    > DE> 2. supports as many characters as possible, in an intuitive way
    > Intuitivity is in the eye of the beholder.
    > For people not used to the U.S. English layout, this layout is by no
    > means intuitive. Thus, your requirements 1. and 2. contradict for most
    > people in the world. A realistic requirement, however, is:
    > ... in a way which can be experienced as intuitive when learned
    > and used for some time.

    Sorry, "intuitive" is an infamously loaded word, sort of like
    "user-friendly." I meant roughly that the process of creating an
    "extended" letter bears some reasonable resemblance, as much as
    possible, to creating the base letter and/or its extended component.

    For example, if I want to type Á (A with acute, for digest readers) on
    my current custom keyboard, I type AltGr+' followed by A. The AltGr+'
    is a dead key for U+0301. This may not feel "intuitive" to others, but
    at least I am typing something involving an A and something that looks a
    bit like an acute accent. Typing AltGr+A (or some letter physically
    close to A on the keyboard) would be OK too. If I had to type something
    like Shift+AltGr+7, I would consider this less "intuitive." Obviously
    there may have to be some compromises, especially for lesser-used
    characters like U+030F.

    Other than that, I don't see anything about the U.S. English layout that
    is any more or less "intuitive" than any other national layout. I would
    have as much trouble with a German QWERTZ layout as a German would with
    a U.S. layout. The comment about "intuitive" has to do with typing
    characters for which I have to go to extra effort.

    > DE> 3. can be implemented with existing 101-key hardware (no new
    > DE> physical keys)
    > Please have a look at:
    > The revision of ISO/IEC 9995-3 is now in the FDIS ballot stage and
    > thus stable.
    > It tries to be a compromise between required compatibility to older
    > standards and being intuitive as far as possible for an audience as
    > large as possible.

    The FDIS says: "The standard defines an 'common secondary group layout',
    which is a mapping of characters to the unshifted and shifted positions
    (and, in one case, also to a third level beyond unshifted/shifted) of
    the 48 alphanumeric keys of a standard PC keyboard to be reached by a
    'group selector' (which could be a special key working similar to the
    'Right Alt' or 'AltGr' key of some keyboards, but the exact mechanism is
    not subject of the standard)."

    How do I access this third level using a standard 101-key keyboard, if
    the FDIS leaves the mechanism undefined but suggests a new key? I don't
    have the option of adding a new physical key to my keyboard. The third
    level is critically important because most combining characters live
    there; it would do me no good to implement the FDIS layout without some
    way of getting to the fifth level.

    If you can answer this question for me, I will seriously consider
    switching to this layout. I remember asking for this at least once

    > "4 shift states" is not compatible with the group/level concept of
    > ISO/IEC 9995 which has successful implementations in several keyboard
    > standards in different countries. There are three levels which usually
    > are recognized as shift states (the third one is commonly invoked by
    > the AltGr key, if one is present, or by Ctrl+Alt), and usually one or
    > two groups.

    I mean 4 states: Shift, AltGr (or Ctrl+Alt), both, and neither.
    Ctrl+key, as you mention, is a non-starter.

    > Microsoft Keyboard Layout Creator (V1.4) is definitely an outdated
    > tool.

    My goal here isn't to praise or castigate MSKLC or any version thereof.

    "verdy_p" <verdy underscore p at wanadoo dot fr> wrote:

    > Dead keys (and AltGr==Altr+Ctrl modifiers) are also used in US
    > International keyboards. They are infinitely more friendly than
    > modifier keys (EITHER "Ctrl" OR "RightAlt" OR
    > "AltGr"=="LeftAlt"=="Ctrl+Alt", possibly combined with Shift or
    > CapsLock), as this reduces a lot the number of keys to locate and
    > learn, and they are much easier to type than multiple modifier keys,
    > and many non intuitive assignments all around the keyboard to get the
    > appropriate combinations of letter + diacritic, for which they are
    > first used.

    For my purposes, I insist that the plain U.S. English layout (Shift and
    no-shift) must remain unaltered. I would have a hard enough time
    learning to type (e.g.) an extra space before certain combinations but
    not others; I'll never get any of my colleagues to learn something like
    that. So the dead keys have to be on AltGr or Shift+AltGr or something.

    Marc replied to Philippe:

    > Deadkeys are not 'intuitive', just familiar (kinda like QWERTY or dare
    > I say AZERTY). They are an old compromise from typewriters. When you
    > write diacritics with a pen, do you write them before or after the
    > base letter? When you press a deadkey (on a Windows system), you get
    > no visual feedback. So the naive user then presses the deadkey again.
    > And gets *two* non-combining forms of the character. There is nothing
    > intuitive about that.

    I'm used to them and don't consider them a problem. I did have to climb
    the learning curve, but that was forever ago and it wasn't that hard.

    Getting away from the terminological differences and people's opinions
    of MSKLC, I will accept (offline) and consider any suggested keyboard
    layouts that (a) follow my original points 1 and 3, (b) support as many
    characters as possible in a way that appears easy to learn and memorize,
    and (c) uses no modifier keys other than Shift and AltGr. The 101-key
    physical keyboard requirement is a sine qua non.

    Doug Ewell  |  Thornton, Colorado, USA  |
    RFC 5645, 4645, UTN #14  |  ietf-languages @ ­

    This archive was generated by hypermail 2.1.5 : Sun Jan 03 2010 - 20:53:07 CST