When temperatures and/or densities are specified instead of being calculated from the data, it would be useful to allow the user to also specify an uncertainty. Otherwise, the statistical uncertainties on output quantities will be underestimated. This should be optional though, because I imagine a lot of the time people wouldn't have a meaningful estimate of the uncertainty on the quantity they want to fix.
When temperatures and/or densities are specified instead of being calculated from the data, it would be useful to allow the user to also specify an uncertainty. Otherwise, the statistical uncertainties on output quantities will be underestimated. This should be optional though, because I imagine a lot of the time people wouldn't have a meaningful estimate of the uncertainty on the quantity they want to fix.