Finally, because the keys are staggered diagonally, anyone who did geometry in school knows that a hypotenuse is longer than the sides, therefore all keys are further away than they need to be from the home row.
This argument only makes sense if you arrange the keys in columns instead of rows (with the columns staggered to accommodate different finger lengths), angle the columns inward to match arm angle, and ideally split and tent the two keyboard halves. Otherwise, the distance that matters is the distance to the key from the natural resting position for each finger when the hand is in its home/center position and relaxed [note that when the hand is relaxed, the fingers do
not naturally fall on the centers of the home row keys]. And euclidian distance isn’t even quite the right metric, since what we care about is how easy it is to reach a finger into the new position, which is not precisely correlated with ruler distance. As long as you have a flat one-piece keyboard with the keys arranged in rows, any change to the specific horizontal displacement between rows is a marginal improvement at best.
In other words, this is an improvement (not too dramatic, but noticeable) for a flat keyboard:
![](http://i.imgur.com/Ucj8Fog.jpg)
This is not really any improvement:
![](http://planckkeyboard.com/assets/jack-alt-top.png)