From: Marc Durdin (marc.durdin@tavultesoft.com)
Date: Mon Jan 04 2010 - 14:55:50 CST
Michael Everson wrote:
> On 4 Jan 2010, at 00:44, Asmus Freytag wrote:
> > On the contrary. The lack of visual input is a definite drawback in
> > designing keyboard input for rarely used characters and by less than
> > expert touch typists. Whether the comparison between using a pen or
> > using a keyboard is salient or not, the fact remains that hidden
> > states in user interfaces are suboptimal.
>
> That is a fault of the Windows interface, then; the Mac OS *does* give
> visual indication when a dead-key state has been invoked. (For
> example, the grave state is indicated by displaying the grave accent
> in a yellow field until the following letter is pressed.)
Yes, I have always liked the way Mac OS did that. Adding that type of visual feedback to Windows today would be a big improvement (hard to see how it could be done across all applications...) But remember the entire point of the original discussion was about the limitations of using MSKLC for a large scale Latin keyboard.
>> I made use of a variety of non-diacritic deadkeys in the Irish
>> keyboard I designed: A key I call "specials" allows me to access
>> ƣƿǝʀθȝʊɪɸ (qwertyuip) and əſẟɡɣƕƞĸɬ (asdfghjkl)
>> and ʒχɔʌβŋɯ (zxcvbnm). I use a number of these very
>> frequently in medievalist and linguistic work. Another Key I call
>> hooks is there for pseudo-diacritics: ʠỽꜥʈƭƴʋɩi҆ƥ
>> (qwertyuiop) ꜣʃɗƒɠɦƙ (asdfghk) ȥɖƈɓɲ (zxcbn). And with
>> these two keys and alt I can access even more.
>> ɑʍɹꞃꞇʏᵿᵻɤɸ ɐꞅꝺꝼᵹʰʲ№˚ꞁ ɒᵈˠᵇꝿ.
>> This is useful, mnemonic, and a great use of keyboard technology.
OK, I take your point. However I would ask how you determine that e.g. "specials" followed by e gives you ə and not ə or ɛ or ɞ or etc. Diacritics on dead keys are reasonably self-documenting. For any other situation that really doesn't hold. And there reaches a point where you have so many dead keys and combinations to cover all the different variations of each letter that memorising the combinations becomes difficult. And it is especially difficult without visual feedback.
> Keyboards using this design work well, and are widely implemented. I
> think it is wrong for Marc to dismiss them as he has done. (I have
> used his software, by the way, for development on the Linux platform,
> and will continue to use it. And I am happy it supports deadkeys.)
I will continue to hold onto my unsupported assertion that typing diacritics after the vowel is more intuitive than typing them before. But Keyman has supported deadkeys since 1993 and I've written plenty of keyboards that use deadkeys. Interestingly, in Keyman keyboards, the deadkey support has been used by many keyboard authors more for (primitive) state management than for the original deadkey purpose.
Marc
This archive was generated by hypermail 2.1.5 : Mon Jan 04 2010 - 14:57:59 CST