Validity, Accuracy, and Precision

Under construction 🚧

We plan to add information regarding validity of end points, particularly with respect to date ranges, and also expected accuracy.

Generally, calculated astronomical values are presented to 8 decimal places, which may exceed the expected accuracy in some cases. Calculated coordinates (e.g. latitude, longitudes in eclipse paths) are generally given to 6 decimal places.

It's been our experience that excessive precision is generally advantageous - not in a scientific context of course, but definitely in that of day to day software development, where comparability of results is usually most helpful and the ability to detect changes in outputs can often signify an unintended or unexpected change in inputs, assumptions, algorithms or behaviors, even if the digits that changed are not in themselves practically significant.

Last updated