Comments on: New Keyboard Layout Analyzer, Less Lines of Code

By: patorjk.com » New Keyboard Layout Analyzer (Preview)

patorjk.com » New Keyboard Layout Analyzer (Preview) — Fri, 03 Feb 2012 00:06:30 +0000

[...] Posts Three Things to SayAsciiMo – Where have I seen this before?What patorjk.com isMonk-E-GodNew Keyboard Layout Analyzer, Less Lines of CodeJavascript Snakepatrickgillespie.com – What Should I Do With It?Plotting and Analyzing US Baby [...]

By: patorjk

patorjk — Mon, 19 Sep 2011 03:21:15 +0000

albedo – Thanks for the information on the Neo2 layout. I tried a few pieces of sample text but I didn’t see the Neo2 outperforming the other layouts (though I only used English text).

By: patorjk

patorjk — Mon, 19 Sep 2011 03:14:22 +0000

hi Fireandmirth – You’re right about #1. A lot of foreign keyboards also have different shift set ups, which is something I didn’t consider when I first wrote the app. When I get the time, I’ll probably set it up to make it much more configurable this way. #2 is an interesting idea, and something I’ll consider. For my Typing Speed Test, I made the keys users pressed with each finger configurable, which is something I’ve always meant to go back and add into this app – though I didn’t think about the starting position.

Anyway, these are both great ideas. Thank you for the compliments and I’m glad you liked the app .

By: Fireandmirth

Fireandmirth — Sun, 18 Sep 2011 21:31:11 +0000

I love your analyser. Love it!
I’ve been tweaking keyboard development for years and wish I had found this sight before.

Two quick suggestions for the next iteration:
1) Please, pretty please, allow shift states to be user defined (i.e. delink the characters on a key). Current shift states on all of the standard keyboards you offer are not optimal. Classic errors include (1) / on the same key as ?, with / unshifted and ? shifted, and (2) a dedicated key for ‘; :’.
You mention this above in comment 4, and I think it would be well worth it.

2) This is a bit trickier, but allowing the user to state where the hand is placed on the keyboard would also be a huge plus. In your bio you talk about Dvorak overworking your right pinky. Moving your entire right hand one key to the right reduces the strain on the right pinky from having to support 13 keys (7 of which are relatively high use) to having to support only 9. The index fingers can pick up two each of the middle keys, and the index fingers can handle it .

Anyway, those are my two cents. Thanks for your great work.

By: albedo

albedo — Fri, 05 Aug 2011 18:52:22 +0000

Neo2 always wins – even in English texts — although it’s primarily made for German.

Original Neo2 layout (the small o’s are only placeholders for the diacritical marks:

ô123456 7890-ò
xvlcw khgfqßó
uiaeo snrtdy
üöäpz bm,.j

Shift (mod 2):

?°§?»«$ €„“”—?
XVLCW KHGFQ?õ
UIAEO SNRTDY
ÜÖÄPZ BM–•J

Mod 3 (caps lock)

¹²³›‹¢ ¥‚‘’
…_[]^ !=&?ø
\/{}* ?()-:@
#$|~` +%”‘;

Mod 4 ()

contains a numeric block and wasd like cursor controls, tabs etc.

Try this approximation out, it works for text without special characters, unfortunately you can’t add more layers that you would need for a true Neo2 layout, at least for the Shifted characters:

0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,43,45,37,44,16,36,34,33,15,32,25,26,27,28,21,22,29,17,23,30,47,18,19,31,20,40,41,51,38,39,24,42,46,48,49,50,35,52,53,54,55,56,57,58,59,60

By: patorjk

patorjk — Wed, 02 Mar 2011 17:29:51 +0000

Hug – I’ll certainly look more into bigrams for the next version of the app.

By: Hug

Hug — Tue, 01 Mar 2011 17:00:32 +0000

Very nice script.
Next step is to integrate statistics about bigram management?

An exemple of what it could be :
http://bvofrak.blogspot.com/p/statistiques_31.html

An once again great job.

By: patorjk

patorjk — Tue, 01 Mar 2011 14:24:28 +0000

Tony – Yes, horizontal movement is factored in. The distance is based on your fingers’ home row positions, unless you’re using the same finger that you used previously (for example – if you’re typing T and just typed R, the distance calculated will be from R to T. If the next move you make doesn’t include the index finger, that finger returns to its home row position).

By: Tony

Tony — Tue, 01 Mar 2011 10:01:42 +0000

Thank you for your time and effort in coding this analyzer.

I want to know whether you have calculated the horizontal movement travel.

In Qwerty, I mean when you type G or H? These keys in the home row, yet it takes time and index finger travel to reach them. But less distant than top or bottom row, of course.

And to move up to upper row is easier than bottom row, so you should have a more key distance on bottom row too, in comparison with the upper row.

By: bsdhacker

bsdhacker — Tue, 09 Nov 2010 03:17:55 +0000

Looks like my custom ZBULP layout beat them all, including Colemak & the generated “optimal” layout for my given corpus, which was just a couple of ebooks.

zbulpkmyfj[]\
riaohdtens’
;xcvqgw,./

Now to throw some source code at it…