Hi,
Continuing on from my earlier experiments with Twiddler key layout
efficiency, I decided to break out the various elements which make up the
"typing time factor" I used to compare layouts. This should help to
clarify how these numbers are made up -- and make it easire to see what is
actually causing the differences in (estimated) efficiency. Here are the
results for the various keyboard layouts and source text combinations I
used yesterday:
Story Story Story Code Code Code
Default Shrdlu T/space Default Shrdlu T/space
------- ------ ------- ------- ------ -------
Time factor: 51442 53633 40205 69621 69085 57270
Key-press factor: 5547 6805 7302 5270 7070 6818
Move 1 space factor: 4972 3849 2155 5522 3974 3903
Move 2 spaces factor: 1669 1908 721 1710 2219 979
Release+press factor: 1518 1318 2196 3533 3009 3237
Thumb-key factor: 337 337 337 1094 1069 765
2-finger adjustments: 402 1809 215 619 1107 511
3-finger adjustments: 0 5 1 35 29 39
4-finger adjustments: 0 0 0 0 0 0
The time factor at the top is the same value I posted yesterday; the
remaining numbers show how that factor is made up:
key-press factor: this is the number of times the action of pressing a key
down was the most time-expensive action required to type a particular
keystroke. Note that the finger was *not* pressing a key down in the
previous keystroke, so it was free to position itself over the key ahead of
time.
Move 1 space factor: this is the number of times the action of lifting a
finger from one key and pressing down an adjacent key (ie, L->M, M->L,
R->M, or M->R) was the most time-expensive action required to type a
particular keystroke.
Move 2 spaces factor: the number of times the action of lifting a finger
and pressing a key two-spaces over (ie, L->R, R->L) was the most expensive
action.
Thumb-key factor: the number of times the thumb had to press down a key.
2-finger adjustments: the number of times two fingers had to move from one
key to another.
3-finger adjustments: the number of times three fingers had to move from
one key to another.
4-finger adjustments: the number of times four fingers had to move from one
key to another.
**********
Okay, so what does all this mean? Well, the effect of two- and
three-finger movements between keystrokes is debatable (see my separate
message to Brandon), so we can discount these for now. But look at the
number of times a finger had to move one or two spaces between keystrokes:
Story Story Story Code Code Code
Default Shrdlu T/space Default Shrdlu T/space
------- ------ ------- ------- ------ -------
Move 1 space factor: 4972 3849 2155 5522 3974 3903
Move 2 spaces factor: 1669 1908 721 1710 2219 979
---- ---- ---- ---- ---- ----
Total: 6641 5757 2876 7232 6193 4882
This is where, I believe, the differences in keyboard layouts really start
to show up. I've summed the results to get a rough measure of the number
of times a finger needs to "jump" from one key to another -- and look at
the differences! The default layout is *worst* in this regard, while the
tabspace layout is by far the most efficient. So, even ignoring things
like the number of keypresses required and the effect of multiple-finger
jumps and the use of thumb-switches, the tabspace layout shows a very large
decrease in the number of times your fingers have to jump from one key to
another. In itself, this could suggest that improved typing speeds over
the default are possible using this layout...
- Erik.
PS: If anyone can see any faults in my logic here, please let me know! I
*think* these results make sense, but I'd definitely be willing to stand
corrected if I've made any mistakes...
--
Subcription/unsubscription/info requests: send e-mail with subject of
"subscribe", "unsubscribe", or "info" to 
Wear-Hard Mailing List Archive (searchable): http://wearables.blu.org
From Wear-Hard Mailing list Archive (WH)
Maintained by R. Paul McCarty
Archive created with babymail