Rust Geodesy: a new platform for experiments with geodetic software

Rust Geodesy (RG), is an open source platform for experiments with geodetic software, transformations, and standards. RG vaguely resembles the well-known open source PROJ transformation system, and was built on the basis of experiments with alternative data flow models for PROJ. The actual transformation functionality of RG is, however, minimal: At time of writing, it includes just a few low level operations, including:

• The three, six, seven, and fourteen-parameter versions of the Helmert transformation
• Horizontal and vertical grid shift operations
• Helmert's companion, the cartesian/geographic coordinate conversion
• The full and abridged versions of the Molodensky transformation
• Three widely used conformal projections: The Mercator, the Transverse Mercator, and the Lambert Conformal Conic projection
• The Adapt operator, which mediates between various conventions for coordinate units and axis order
• Also, RG provides access to a large number of primitives from geometrical geodesy, all wrapped as methods on a data model unifying the representation of two- and three-axis ellipsoids.

While this is sufficient to test the architecture, and while supporting the most important transformation primitives and three of the most used map projections makes it surprisingly useful, it is a far cry from PROJ's enormous gamut of supported map projections: RG is a platform for experiments, not for operational setups.

Fundamentally RG is a geodesy, rather than cartography library. And while PROJ benefits from four decades of reality hardening, RG, being a platform for experiments, does not even consider development in the direction of operational robustness. Hence, viewing RG as a PROJ replacement, will lead to bad disappointment.

That said, being written in Rust, with all the memory safety guarantees Rust provides, RG by design avoids a number of pitfalls that are explicitly worked around in the PROJ code base, so the miniscule size of RG (as measured in number of code lines) compared to PROJ, is not just a matter of functional pruning, but also a matter of development using a tool wonderfully suited for the task at hand.

Also, having the advantage of learning from PROJ experience, both from a user's and a developer's perspective, RG is significantly more extensible than PROJ, so perhaps for a number of applications, and despite its limitations, RG may be sufficient, and perhaps even useful. First and foremost, however, RG may be a vehicle for geodetic development work, eventually feeding new functionality and new transformations into the PROJ ecosystem.

Aims

Dataflow experimentation is just one aspect of RG. Overall, the aims are fourfold:

• Support experiments for evolution of geodetic standards.
• Support development of geodetic transformations.
• Hence, provide easy access to a number of basic geodetic operations, not limited to coordinate operations.
• Support experiments with data flow and alternative abstractions. Mostly as a tool for the other 3 aims

All four aims are guided by a wish to amend explicitly identified shortcomings in the existing geodetic system landscape.