Module geodesy.scalar
Scalar policy shared by geodesy-d numerical value types and operations.
The public numerical surface deliberately supports the built-in floating scalar types float, double, and real. Public scalar choice does not force every algorithm to use that same type internally: numerically sensitive float operations may promote working arithmetic to double while preserving float at the public boundary.
Numerics
Double is the normative reference precision unless an operation-specific contract states otherwise. Real is platform-dependent. There is no library-wide epsilon; convergence and validation tolerances belong to individual algorithms.
See Also
Operation-specific Numerics: and Validation: documentation.
Date
September 26, 2026
Manifest constants
| Name | Type | Description |
|---|---|---|
isGeodesyScalar
|
True for the supported built-in floating-point scalar types. |