dA/dLambda outputs: trajectory file and dispersion

[EzryStIago]

When estimating a restraint free energy by changing the force constant (section 6.5.3 in the user guide), dA/dLambda is accumulated and printed in the log file (without any measures of dispersion).

It would be nice to have:

1. the raw dA/dLambda samples (with some user-specified frequency)
2. a standard deviation (or variance) for the accumulated dA/dLambda

This would streamline my RFEP calculations (and simplify the SAFEP tutorial as well).

[jhenin]

Thank you for the good suggestion. I think the best implementation will be for the bias to output dV/dlambda to the colvars.traj file.