horcom

Welcome

HORCOM is a complete astrology program for Windows and Linux. It casts birth charts in seven house systems, optionally with topocentric positions, follows transits and directions, computes solar, lunar and planetary returns, compares two people in double wheels, composites and combines, walks the seven year phases of the Munich rhythm theory and searches whole chart collections statistically. Robert Rettig wrote and refined it from about 1989 until 2010. He computed his own ephemerides for Chiron, Quaoar, Eris and other bodies when none existed, wrote the commentary texts that come with the program, and wished for HORCOM to live on.

Start here

What you can do

• Cast a birth chart and read it with coordinates, house cusps, aspects and midpoints beside the wheel.
• Set orbs, rulers, house system and aspect colours to your way of working.
• Keep chart collections and enter births with historical time zones.
• Follow transits, returns and progressions and directions as tables, chart rings and linear graphs.
• Work with the Munich rhythm theory and its Septars.
• Compare two people in a double wheel, the 90 degree dial, a composite or a combine.
• Rectify a birth time, cast harmonics and multiple directions.
• Look up fixed stars, Arabic parts, ingresses and eclipses, and search collections with the statistics.
• Print on A5 or A4 or export PDF, SVG and PNG.

Technical reference

The pages below are for developers and for anyone curious how HORCOM calculates. Each describes one module of the C++ rewrite, explains the original mechanics it preserves, and lets Robert speak in his own words wherever his comments still tell the story best. You do not need them to use the program.

// IT IS BY NO MEANS RECOMMENDED, HOWEVER, TO REFORMULATE ROUTINES USED IN THE CALCULATION CHAIN YOURSELF. THE DEVIL IS ALMOST ALWAYS IN THE DETAIL ! // SO: IN THE CALCULATION CHAIN LEAVE EVERYTHING AS FAR AS POSSIBLE AS IT IS, ESPECIALLY THE ORDER OF THE PROCEDURES AS WELL.

Modules

Conventions of the port

Robert's comments travel with the code

Wherever one of his comments still applies to the converted routine it is carried over verbatim, in his German wording, marked with the prefix //RR. One comment per routine names the original HORCOM procedure it was ported from. Code comments stay self contained, they never point at files outside the repository.

Same numbers first

Every ported routine first reproduces the original's output and lands together with its tests, transcription pins plus independent values from the literature Robert himself used, above all Jean Meeus. Only under green tests may a routine be restructured. Where the original and the rewrite disagree, the original is presumed right.

The two integer worlds of GFA BASIC

GFA BASIC has INT, which rounds down, and FIX, which truncates toward zero. The original uses both deliberately. The angle normalisers ng, nb, nh are built on INT and land in [0, period), while wz, the degree within a zodiac sign, is built on FIX and keeps its sign. The port maps them to std::floor and std::trunc and its tests pin both behaviours.

Guards stay where he put them

The epsilon kk = 1e-10 appears in the original exactly where a division or an arctangent argument could hit zero. The port adds it in exactly those places and nowhere else, because the branch behaviour of the aspect logic depends on it.

Verification

The test suite currently holds 131 test cases with 1622 assertions, all green, built with MSVC at warning level 4 and with GCC and Clang on Linux, warnings as errors on all three, and on Linux once more under the address and undefined behaviour sanitizers. The three compilers produce byte identical coordinate tables and wheels, see Installation. The anchors are Jean Meeus's worked examples, the binary data files themselves and cross checks between independent engines, the Chapront theory against the integrated Pluto file among them. The golden reference now exists, tests/golden holds fifty synthetic charts recorded from the original HORCOM7P.EXE running under emulation, dates 1900 to 2012 across both hemispheres, the equator, the polar circle and the calendar edge cases. All 550 planet longitudes agree with the original within one arc second of display rounding, 47 of 50 charts agree on every house cusp, and the three exceptions are a single cusp 1.1 arc minutes off whose cause, the early stop of the original's Placidus iteration in GFA arithmetic, is documented there. tools/compare_golden.py replays the comparison.