a new keyboard - discuss and critique

Steve Mosher steve at openmoko.com
Wed Jun 3 20:54:40 CEST 2009

yep.  you have to vary a whole host of test conditions.

 Subjects: those with touch screen experience: those without
 Order of presentation:
 Size of subjects finger
 How they type..for me I've always struggled between being a thumb typer and
and index finger typer.. neither works for me very well since I have big 
fingers, bad eyes,
and shakes from drinking too much coffee

on the accuracy and time  one might find that certain content (say SMS) 
might do better
with one format as opposed to another..who knows?

Could waste a whole summer doing a test like that. Long ago raster I 
actually did test
design for UIs in fighter aircraft. nasty business. fun if you get paid 
for it, but not
how i'd spend my free time. So I aint volunteering.

Carsten Haitzler (The Rasterman) wrote:
> On Wed, 03 Jun 2009 11:08:51 -0700 Steve Mosher <steve at openmoko.com> said:
> indeed. i believe "its not worth bothering". so i am not going to invest any
> time into trying it (to try it i'd have to build it... and that will use time)
> as my theorising says to me "it wont help". if someone else who thinks it
> really is the bees knees... then build it, try it, test it and prove it! get a
> non-biased 3rd party to go use the kbd to enter common input or a phone (emails,
> sms's, etc.) and then time over maybe 50 specific messages how long it takes to
> get a 100% accurate entry of it. of course make the test unbiased by swapping
> which kbd they use first t enter in each test (first test kbd a then b, in test
> 2, b then a etc.).
> if you really think this kbd (on a touchscreen) is the bees knees - prove it! :)
>> before you test you need a good theory. that way you know what data to 
>> collect. raster has a theory why it wont be better. That's a testable 
>> theory, I think.
>> Michal Brzozowski wrote:
>>> 2009/6/3 Carsten Haitzler <raster at rasterman.com>
>>>> but u'll hit dead space (nothing) about 50% more often than the actual key
>>>> you
>>>> intended to hit now (as 50% of the area it would have normally used is now
>>>> dead
>>>> space). for a physical kbd it might help as u have better tactile feel as u
>>>> have
>>>> more edge near the finger, but for a touchscreen - all you do is lose 50%
>>>> of he
>>>> area you had before for hitting keys. (admittedly  for mis-types now where
>>>> you
>>>> would have hit another key there is less area to hit too).
>>>> so you lose more keypresses (you lost 50% of the hit area), but you lost
>>>> mis-hit
>>>> area too.
>>>> one way or another - you press, then quickly press next key, but then
>>>> notice
>>>> that the previous key wasn't hit - u still need to backspace and re-enter
>>>> again
>>>> (possibly missing again). you just give up 1 kind of error for another kind
>>> I'm amazed how much you guys can theorize about the idea before actually
>>> trying it.
>>> Is it because of allergy to patents? I bet we could use the general idea if
>>> we wanted to, without violating the patent. Just change the shape of the
>>> keys or something.
>>> ------------------------------------------------------------------------
>>> _______________________________________________
>>> Openmoko community mailing list
>>> community at lists.openmoko.org
>>> http://lists.openmoko.org/mailman/listinfo/community
>> _______________________________________________
>> Openmoko community mailing list
>> community at lists.openmoko.org
>> http://lists.openmoko.org/mailman/listinfo/community

More information about the community mailing list