Eclipse Catalog Comparison

This report compares Moira's eclipse calculations against the local NASA catalog fixture already used in the test suite. It is intended as a readable cross-era summary, not as a replacement for the tests themselves.

Maxima Snapshots

Solar

NASA date	NASA type	Moira native type at NASA maximum	Notes
-1797-02-01T21:25:34	H	hybrid	classification at catalog maximum
0500-02-15T11:06:27	H	hybrid	classification at catalog maximum
0500-08-11T00:35:02	A	annular	classification at catalog maximum
2005-04-08T20:36:51	H	hybrid	classification at catalog maximum
2809-02-05T21:20:58	H	hybrid	classification at catalog maximum

Lunar

NASA date	NASA type	Moira native type at NASA maximum	Notes
-1801-04-30T07:38:52	T	total	classification at catalog maximum
-1801-10-23T22:49:56	P	partial	classification at catalog maximum
0499-03-13T12:12:02	T	total	classification at catalog maximum
2000-01-21T04:44:34	T	total	classification at catalog maximum
2800-02-01T23:47:11	T	total	classification at catalog maximum

Search Timing

These are the more meaningful comparison rows, because they compare the catalog's reported greatest-eclipse instant against Moira's own searched maximum.

Solar Search Cases

Case	NASA expected	Moira native	Residual
ancient_hybrid (hybrid)	-1797-02-01T09:51:13.000029Z	-1797-02-01T09:49:52.939670Z	-80.06 s

Lunar Search Cases

Case	NASA expected	Moira native	Native residual	nasa_compat	Compat residual
ancient_total (total)	-1801-04-29T20:03:04.999996Z	-1801-04-29T20:02:15.346372Z	-49.65 s	-1801-04-29T20:08:07.632251Z	+302.63 s
future_penumbral (penumbral)	2801-06-17T19:00:54.999930Z	2801-06-17T19:01:15.757170Z	+20.76 s	2801-06-17T19:01:25.031833Z	+30.03 s

Interpretation

• At catalog maxima, Moira's native classifier agrees cleanly across the representative ancient, classical, modern, and future rows in this local fixture slice.
• The meaningful timing differences appear in searched greatest-eclipse instants, not in simple at-instant classification.
• For lunar ancient cases, the largest single contributor is the Delta T branch choice. In the diagnosed ancient_total case, switching the same native shadow-axis objective from native Delta T to NASA-canon Delta T moves the answer by about 387 seconds.
• Moon treatment matters too. In that same case, switching from a retarded Moon to a geometric Moon inside the native branch moves the result by about 35 seconds.
• Once Delta T branch and Moon treatment are aligned, Moira's native shadow-axis minimum and the canon gamma-minimum objective collapse to essentially the same instant. That means the remaining difference is primarily model basis, not an unstable search algorithm.
• Practical reading: modern and near-modern comparisons are tight; deep ancient and far-future timing comparisons should be read through the lens of Delta T doctrine and event-definition choice, not just raw residual size.