Comment 7 for bug 1652762

Alexander, I have seen the previous situation just selected to show either ELP's or DE43x's nDot. Both ephemeris variants should optimally be used with published DeltaT formulae which fit to them directly. If one of the other DeltaT formulae are used, they refer to their own nDots, and their results must be compensated with the correction term. So I find the formulae's nDot values more interesting. The displayed value for DeltaT is independent from this decision, and my changes did not touch numerical/computation results. If you think the ELP or DE43x values should be displayed, we can change it back easily. And we really need a good discussion on this topic in the Guide...

Another observation is the uselessness of some DeltaT formulae which are just approximating polygons through original tables derived from observations (and implemented in other algorithms), they don't bring new information. Some are even problematic, e.g. Schmadel-Zech is a 12th degree polygon which causes absurd faults in antiquity. So there should be also proper handling for dates beyond their range of usability.