1. The following correction is provided by Robert Munafo on 16/02/2014. His message is reproduced here. Thank you Robert!

The standard value of 1/alpha is 137.035999074 with an uncertainty of 44 in the last 2 digits. For your formula (e^(1-q)/(pi/2))^2 to produce this value, q would need to be about 3.00853912384, with an uncertainty of 16 in the last two digits. That's a bit bigger than 3.007...

C1 = alpha = 1/137.035999074(44) 

R1 = alpha: 0.0072973525698(23)

C2 = y = sqrt(alpha)

R2 = y: 0.085424543134(14)

C3 = x = y * pi/2

R3 = x: 0.134184558574(22)

C4 = 1-ln(x)

R4 = 3.00853912384(16

The original statement we reported on http://www.ijspace.org/Alpha/Alpha.html was,
"The exact value of the fine-structure constant corresponds to q ~ 3.007."
---------------------------------------------------------------------------------------