The non-intuitive nature of dead keys that I first noted back in 2004 comes back to haunt us all...

by Michael S. Kaplan, published on 2013/10/21 07:01 -04:00, original URI:

Through all of the brave talk about chained dead keys that has been going on, one question has been left unanswered.

It is around one of those times when Apple got it right in the early days, and Microsoft as it happens got it wrong.

When you are typing the diacritic piece of the dead key, Apple let the uncommitted character appear as an amorphous chunk, while Microsoft showed nothing.

And it has always been this way, despite how unintuitive the behavior appears to the user.

Not to mention how much harder it is for keyboards like the US International keyboard layout.

This was true even before chained dead keys became a reality with the Cherokee phonetic keyboard layout.

i have been assured many times by native Cherokee informants that the layout's behavior is completely intuitive to the user.

Perhaps informants and native speakers of Indic languages or Korean would feel the same way if appropriate layouts could be given to them.

However, it is less clear whether behavior of Fidel script languages like Amharic, Tigrinya, or Ge'ez was involved.

I have been told that I am almost certainly not going to ever see this behavior change irregardless of how desirable the result might be, so I shouldn't even bother to ask for the change.

And I believe what I have been told here.

I am resigned to wondering whether knowledge gained from the Cherokee Phonetic layout could be used to subtly improve the US International layout.

And yes, I have asked for this very change many times since I first noticed the behavior back in the middle of 1996.

You are welcome to start a petition up on if you think it might help; I'll vote for it!

For the foreseeable future, however, this is the behavior we will be required to live with...

Though in all seriousness, PLEASE let me know if you start that petition; I will not only vote for it, I will also blog about it encouraging others to do the same!

Anubhav Chattoraj on 21 Oct 2013 7:53 AM:


Looks like my multi-keystroke Indic keyboard will have to be implemented in Google IME.

Azarien on 22 Oct 2013 3:58 AM:

Perhaps a new mechanism, some kind of IME-lite, would be ideal.

Brendan Elliott on 24 Oct 2013 1:21 AM:

Yeah, moving more languages to be IMEs seems like the obvious way to fix it within the existing input framework. Put the dead char in a composition, update it in-place on the next key press. Don't even need candidate UI at that point for most languages and it would probably end up quite similar in design to the Korean IME's Hangul input.

Doug Ewell on 24 Oct 2013 7:42 AM:

I like Azarien's idea of an "IME-lite." The existing model seems to divide scripts into the really simple, sticking an accent on top of an 'a', and the really complex, visually picking one Han character out of dozens of candidates. Something in-between is needed.

Please consider a donation to keep this archive running, maintained and free of advertising.
Donate €20 or more to receive an offline copy of the whole archive including all images.

go to newer or older post, or back to index or month or day