The first fit was made manually, outgoing from well known triple bonds like HCCH or N2.
An iterative program was written based on the penalty function over homonuclear and heteronuclear pairs [Equation (1)].
Putting its derivatives with respect to ri equal to zero, one obtains the iterative algorithm given in Equation (2).
K is the number of data points in each set and N are the iterations.
The standard deviation (given by Equation (3)) was 3.2 pm.