Most letter frequency studies appear to use a corpus of 19th century literature.
1) Does anyone know of studies on letter frequency in addresses and names?
2) What about number frequency in credit cards? I did find this link:
http://blog.lumberlabs.com/2011/04/cred … s-law.html
No one here is going to help you perform credit card fraud with the most efficient keyboard layout.
But seriously, I have been curious about this as well, but don't have a good source.
What about Peter Norvig's analysis? That uses a more recent corpus.
My first thought when I saw that credit card number distribution, was it probably has something to do with Benford's Law. But I see the comments already mention that. Not sure why that would be useful though (or relevant to Colemak come to that!)
Using Colemak-DH with Seniply.
I enter a lot of credit cards. I am currently trying the layout 7531980246 on the bottom row
http://www.keyboard-layout-editor.com/# … 5%3B&=ctrl
I use footpedals for 'missing' keys.
I asked on this forum because there seems to be a lot people here interested in layout, and because I use Colemak.
None of the studies I can find have a corpus of names and addresses.
Making a layout for names seems like a bridge too far to me. You do of course need a good number entry system (maybe an Extend Nav/NumPad?) but, well, names is too far. Consider that you're typing non-names at least as much as you're typing names anyway; will you learn and maintain two layouts then, constantly switching between them? Because i don't think that's very efficient all in all.
*** Learn Colemak in 2–5 steps with Tarmak! ***
*** Check out my Big Bag of Keyboard Tricks for Win/Linux/TMK... ***
The letters j, k, and v seem to appear more frequently than studies would indicate, but I have no data to back it up. If this is correct, it may be due to names (and variants of) 'John', 'Mike', and 'Dave'. However, those keys are fairly well placed already.
I wasn't thinking of a radically different layout, but perhaps a small modification.
The number portion of street addresses likely conforms to Benford's Law. The postal codes may not if business is heavily local - there should be frequent repetition of certain numbers.
The J is exceedingly rare in normal English text (see the Tarmak discussions for instance), so it's no surprise if it's more common in names. This letter could be problematic if you do type it often; Dutch people have reported that for instance (they have a lot of IJ digraphs and more). One suggestion for a minor change if the J feels problematic, is a "Mod-DH"-type switch: Swapping two index-finger keys is seen as a very minor layout change that will only disrupt your typing for a short time in a not too serious way. One major advantage of this kind of swap is that it won't disrupt Colemak's excellently low same-finger bigram frequency. But of course, the other index-finger keys are more common than J in English so choose your poison wisely!
You'll be pleased to know that the V is already placed in a slightly "too good" position for its frequency in Colemak, to keep the the Ctrl+Z/X/C/V shortcuts intact. So if it's more common in names, Colemak should accommodate that just fine.
According to the Mod-DH "crooked line" of thinking, so is the K. In my DH-mod, I switch the H and K from vanilla Colemak to improve the H's position.
*** Learn Colemak in 2–5 steps with Tarmak! ***
*** Check out my Big Bag of Keyboard Tricks for Win/Linux/TMK... ***
Making a layout for names seems like a bridge too far to me.....will you learn and maintain two layouts then, constantly switching between them? Because i don't think that's very efficient all in all.
+1
A different idea. Perhaps it is a good idea to use a script (AutoHotKey for Windows; other solutions for OXS and Linux)? Say, many costomers have a name that ends in "itsch" or that starts with "Van" Than you can program that, e.g. alt + i = itsch; alt + v = van
