would adding 2 1.5u change anything?
The 1.5u keys we're
missing are rare keys like F1-F12, arrows, or SUPER, which it wouldn't hurt to add if it didn't drive up the price too much. We could also add in already-existing keys, such as a vertical 1.5u TAB, or more of the other keys in the set. If I look at the results from my optimizer, however, it tends to want to add 2u keys rather than 1.5u keys at this point.
Just did a simulation run with no FN, and here's the suggested 21-key layout: as you can see, the only 1.5u key it added is GUI (SUPER), and for the rest it added 2u keys.
More
21 keys, with:
mean coverage 78.0%
std. deviation 18.3%
median coverage 79.2%
minimum coverage 16.7%
maximum coverage 100.0%
13 mean extra keys
('1x2', u'KEY_Spacebar')
('1x2', u'KEY_Spacebar')
('1x2', u'KEY_Control')
('1x2', u'KEY_Shift')
('1x2', u'KEY_ReturnEnter')
('1x2', u'KEY_DeleteForward')
('1x2', u'KEY_Escape')
('1.5x1', u'KEY_Escape')
('1.5x1', u'KEY_Shift')
('1.5x1', u'KEY_Shift')
('1.5x1', u'KEY_ReturnEnter')
('1.5x1', u'KEY_GraveAccent_Tilde')
('1.5x1', u'KEY_GUI')
('1.5x1', u'KEY_Dash_Underscore')
('1.5x1', u'KEY_Equal_Plus')
('1.5x1', u'KEY_SingleQuote_DoubleQuote')
('1x1.5', u'KEY_LeftBracket_LeftBrace')
('1x1.5', u'KEY_RightBracket_RightBrace')
('1x1.5', u'KEY_Slash_Question')
('1x1.5', u'KEY_Tab')
('1x1.5', u'KEY_DeleteForward')
Assuming we're not trying to reduce the size of the set, are there 1.5u keys that people feel are missing and would like to see instead of FN or SHIFT?
EDIT: How about a 2u blue enter?