Author Topic: If you build a better numpad...  (Read 4242 times)

0 Members and 1 Guest are viewing this topic.

Offline Snarfangel

  • Thread Starter
  • Posts: 288
If you build a better numpad...
« on: Sat, 23 May 2015, 09:12:23 »
I think I've seen this mentioned on Geekhack, but back in July 1960, an article was published in The Bell System Technical Journal titled
"Human Factors Engineering Studies of the Design and Use of Pushbutton Telephone Sets, By R. L. DEININGER."

This article (available here: http://www.beatriceco.com/bti/porticus/bell/pdf/touchtone_hf.pdf ) gives a pretty in-depth description of the experiments done that resulted in the familiar 3x3 + 1 number format used in touch tone phones (and later, smartphones).

Several interesting facts emerge. First, they tested a lot of formats:
101434-0

Second, the calculator number pad layout
101436-1


 is measurably inferior to the touchtone layout
101437-2


though the difference was small.

Third, and most interesting to me, *the 3x3 plus 1 touchtone layout was NOT the layout most preferred.* Out of the five top designs, two horizontal rows of five keys
101438-3


had the most preferred votes (the touchtone layout was third), and forth in "least preferred" votes (the touchtone format had the second highest "least preferred" votes). Interestingly enough, while two horizontal rows was the most preferred, two vertical rows were the least preferred.

I do wish they had tested formats with 0 next to 1, such as
101439-4
 because having the most common numbers on the numeric keypad "home row" might be advantageous for data entry, especially with Enter or Shift/Function under the thumb instead of an enlarged 0 key.

Other possibilities I think would be better test subjects than "The Cross," "The Stairstep," or "The Rainbow" would be things like:
101441-5

and
101442-6

Finally, I'd love to see some testing done on a full-size numeric keypad
101444-7
Preferably with all of the functions of a Goldtouch GTC-0077


I'd be interested to see what other ideas Geekhackers might come up with. I have a Cherry matrix POS board to play on, and it might be fun to try them out. :)
« Last Edit: Sat, 23 May 2015, 09:15:16 by Snarfangel »

Offline davkol

  •  Post Editing Timeout
  • Posts: 4994
Re: If you build a better numpad...
« Reply #1 on: Sat, 23 May 2015, 11:35:35 »
Original/programmer Dvorak. It's based on Benford's law and maximizes hand alternation.

Offline Kaiteokid

  • Posts: 10
Re: If you build a better numpad...
« Reply #2 on: Sun, 24 May 2015, 11:19:17 »
I actually prefer the standard numpad that most of us are fond of ^^

Offline samwisekoi

  • MAWG since 1997
  • * Administrator
  • Posts: 2480
  • Location: Mt. View, California
  • Sorry, moving houses. Be back ASAP.
    • Tweet samwisekoi
Re: If you build a better numpad...
« Reply #3 on: Sun, 24 May 2015, 11:47:33 »
You could make any of those using a GH36 Matrix Keypad.

https://instagram.com/p/3CLiAIJpVB/?taken-by=samwisekoi

 - Ron | samwisekoi
Auto-typed by my JD45 keyboard.
I like keyboards and case modding.  Everything about a computer should be silent -- except the KEYBOARD!

'85 IBM F-122/Soarer Keyboard |  Leopold FC200 TKL (Browns) + GH36 Keypad (Browns/Greens) | GH-122 (Whites/Greens) with Nuclear Data Green keycaps in a Unicomp case

Offline dorkvader

  • Posts: 6289
  • Location: Boston area
  • all about the "hack" in "geekhack"
Re: If you build a better numpad...
« Reply #4 on: Sun, 24 May 2015, 18:25:44 »
one aspect of the standard keypad is that you can press 0 and 1 (probably the two most common numbers) with your thumb. This speeds things up greatly than the traditional touch-type method of using the forefinger for 741 middle for 852 and ring for 963

Offline Oobly

  • * Esteemed Elder
  • Posts: 3929
  • Location: Finland
Re: If you build a better numpad...
« Reply #5 on: Mon, 25 May 2015, 03:41:49 »
Very interesting!

One thing to consider, though, is that this was for keying in telephone numbers, probably with local numbers being higher focus than long distance, meaning that 0 and 1 are not the most common digits used in testing.

I like the 2 horizontal rows and find I'm fast enough on a single horizontal row, I'm slower on a numpad than using the top row :) Just never needed to use the numpad, so never got used to it.

I like the programmer Dvorak layout, but for a numpad style area I think I'd go for a double row, either like this:
Code: [Select]
0 2 4 6 8
9 1 3 5 7

or like this:
Code: [Select]
0 2 4 6 8
1 3 5 7 9

I think that'd be pretty efficient, with .*/ above, + as a thumb key and a double height = to the bottom right. Like this:

Code: [Select]
  . * /
0 2 4 6 8
1 3 5 7 9
 +     =

Or this:

Code: [Select]
  . * /
0 2 4 6 8
9 1 3 5 7
 +     =

That's what I'd consider a "better" numpad.
Buying more keycaps,
it really hacks my wallet,
but I must have them.

Offline pyro

  • Posts: 177
Re: If you build a better numpad...
« Reply #6 on: Mon, 25 May 2015, 06:35:17 »
Code: [Select]
0 2 4 6 8
9 1 3 5 7

Seems this puts frequent digraphs (12, 34) on the same finger.

Was a bit disappointed they didn't test a two row layout that starts at the bottom(*). Still nice to have confirmation, that two row layouts are more comfortable to people than regular numpads, at least in pure number input without */+-.

(*)
Code: [Select]
67890
12345
I don't think its inferiority may be inferred from the numpad/touchtone comparison. You have to move down to reach 123 on the numpad layout, whereas here 12345 would be on the resting position/~homerow.
« Last Edit: Mon, 25 May 2015, 06:38:04 by pyro »

Offline Oobly

  • * Esteemed Elder
  • Posts: 3929
  • Location: Finland
Re: If you build a better numpad...
« Reply #7 on: Mon, 25 May 2015, 09:37:58 »
...
Seems this puts frequent digraphs (12, 34) on the same finger.

...

Are those really frequent digraphs? I suspect it's very dependent on the type of data you're entering. I would think 10 and 99 would be most common, not sure what would follow them.

Anyway, for pure number entry speed, it's best to use a single row and both hands like with programmer's Dvorak as Davkol posted.
Buying more keycaps,
it really hacks my wallet,
but I must have them.

Offline davkol

  •  Post Editing Timeout
  • Posts: 4994
Re: If you build a better numpad...
« Reply #8 on: Mon, 25 May 2015, 11:09:19 »
And they aren't even digraphs…

Offline samwisekoi

  • MAWG since 1997
  • * Administrator
  • Posts: 2480
  • Location: Mt. View, California
  • Sorry, moving houses. Be back ASAP.
    • Tweet samwisekoi
Re: If you build a better numpad...
« Reply #9 on: Mon, 25 May 2015, 11:14:12 »
I think a key question is "what is it good for?"  Those studies were indeed used for dialing phone numbers, and they were made using people whose prior experience was in using a rotary dial to "enter" phone numbers.

Had they done a similar study with bookkeepers to enter columns of numbers, the answer might have been different.  (Who knows, maybe there is another such study from the real dawn of time.)  Nobody today uses either an adding machine OR a rotary phone, but people did, and they used the two devices to enter different things.

So now that we have smart phones, calulators, and computers, what are we entering?  Phone numbers?  Maybe once per number per device.  Less than that with cloud storage of contact lists and Caller ID.  Strings of numbers for addition?  Not even at checkout counters, really, now that we have UPC bar codes.

Speaking for myself, I enter numbers and formulae into Libre Calc.  And hex color values into various things.  Credit card numbers because browsers don't cache those.  Weapon load-out and selection when gaming.  And for those purposes, really all I need for drastic improvement over a TKL layout is to have all of the numbers in one place on a number pad.  That, IMHO, is the real comparison now.  String of ten keys with zero at the "wrong" end OR a cluster of numbers that I can build muscle memory to use.

But if we are slicing the efficiency numbers finer than that, then I'd say the biggest problem with the two standards is the fact that there ARE two standards.

Going beyond that, I'd go for a double arc of numbers from 0 to 9  plus an optional extra row for hex from A to F.  Possibly with zero being 2x tall for use with monetary entry.  And a decimal point.

101686-0
Hex Keypad Possibility

I might flatten that out and give it a try.

 - Ron | samwisekoi
Auto-typed by my JD45 keyboard.
I like keyboards and case modding.  Everything about a computer should be silent -- except the KEYBOARD!

'85 IBM F-122/Soarer Keyboard |  Leopold FC200 TKL (Browns) + GH36 Keypad (Browns/Greens) | GH-122 (Whites/Greens) with Nuclear Data Green keycaps in a Unicomp case