[SHR] illume predictive keyboard is too slow - Usability features

[Carsten Haitzler (The Rasterman)]

On Mon, 02 Feb 2009 21:53:26 +0100 "Marco Trevisan (Treviño)" <mail at 3v1n0.net>
said:

> Carsten Haitzler (The Rasterman) wrote:
> > On Wed, 28 Jan 2009 18:59:32 +0100 "Marco Trevisan (Treviño)"
> > <mail at 3v1n0.net> said:
> >> Maybe using something like a "trie" [1] to archive the
> >> words could help (both for words matching and for compressing the
> >> dictionary).
> >> Too hard?
> >>
> >> [1] http://en.wikipedia.org/wiki/Trie
> > 
> > so back to the trie... the trie would only be useful for the ascii matching
> > - i need something more complex. it just combines the data with the match
> > tree (letters are inline). i need a match tree + lookup table to other
> > matches to display - and possibly several match entries (all the matches to
> > display also need to be in the tree pointing to a smaller match list).
> 
> Ok, thanks... I got it. However I hope we could have made something that
>  is based on that idea (the trie) but that can be applied to non
> ascii-chars too.
> 
> However in the past days I sent you privately also a mail about some
> issues of the keyboard in latest e17 svn [1], but I got no answer.
> Maybe the mail wasn't sent correctly?!

got it - i just tend to ignore some of my mailboxes for a while and cycle
around to them... got a lot of email here :) i'll get back to you on it. it
just is that kbd isnt a focus at the moment so it tends to take a back-burner
position.

> However I've wrote there also some features that I'd suggest to
> implement in the Illume keyboard. I'll write them here too to make the
> community aware:
> 
> I use the illume keyboard every day and I'm very happy with it as I've
> said many times in this ML, but sometimes it happens that it performs
> some unwanted actions like:
>  - I involuntarily click on a suggested word while I'm still typing my
>    word (cause I'm not too precise I tap over a word, instead of a top
>    char).

thats a problem. mostly of spacing. its actually hard to figure that out. i
really dont know what to do there - if u reduce the hit area for matches - it
gets harder to select them. if i add more spacing you lose more screen to the
kbd. somewhere someone loses. it's a matter of fine adjustments i guess in the
spacing to add more space.

>  - It happens that I got my keyboard switched while typing (yes, I know
>    that this mainly an hardware-related issue, due to the touchscreen
>    jitters).

hmm thats hard to do. either u make swipes less sensitive and thus make it
harder to change layout and solve your problem, or you live with the occasional
swipe... or we have another way to change layout thats easy.

> They seem unrelated, but why not workarounding them by allowing these
> actions only after a small timeout (i.e. waiting few ms from the latest
> char pressure)?

so > lets say 0.4 sec after the last keyboard key press it will allow for
swipes and match hits etc. that could be done. again - tuning a timing value.
will people then complain that ":i often try and swipe or hit a match and it
doesnt respond. i need to do it again"?. hmmmm.

> Generally you never confirm a word or switch keyboard as fast as you
> type over a char (since typing can be un-precise thanks to the keyboard
> correction, switching a keyboard or selecting a word must be precise)...

correct. it's a fine line to walk tho - as above :)

> And... What about making the horizontal word list (the one over the
> keys) scrollable [right-left] as the configuration toolbar is? Would it
> require more computation? I figure that that could improve the usability.

no - it'd be not much of a problem - i just didnt do it. :)

nb - i can see why you often hit a match word. your kbd layout doesnt have
padding ABOVE the qwerty line like the default does... :)

> Bye.
> 
> 
> [1] http://i43.tinypic.com/i4il2d.png
> 
> -- 
> Treviño's World - Life and Linux
> http://www.3v1n0.net/
> 
> 
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at rasterman.com