qhull.org

What It Is

Qhull is a source-code library and set of command-line tools for computational geometry in arbitrary dimensions. Its core use cases include computing convex hulls, Delaunay triangulations, Voronoi diagrams, halfspace intersections about a point, and furthest-site Delaunay/Voronoi diagrams. It implements the Quickhull algorithm and places particular emphasis on handling floating-point roundoff errors. It can also output volumes, surface areas, and convex hull approximations.

Core Capabilities and Limitations

In terms of functionality, Qhull covers many common geometric construction tasks in research, engineering, statistics, and mathematics, with support for 2D, 3D, 4D, and higher dimensions. The toolset includes standalone programs such as qconvex, qdelaunay, qvoronoi, qhalf, and rbox, and qhull can also be used directly. The documentation explicitly lists what it does not support: triangulation of non-convex surfaces, mesh generation for non-convex objects, medium-sized inputs in 9 dimensions or higher, alpha shapes, weighted Voronoi diagrams, Voronoi volumes, and constrained Delaunay triangulations. As such, it is more of a low-level geometry kernel than a complete mesh-generation platform.

Developer Integration and Ecosystem

Qhull can be called from applications. The official recommendation is to use the reentrant libraries libqhull_r or libqhullstatic_r, and a C++ interface is also provided. Its ecosystem is highly mature: MATLAB’s n-dimensional geometry functions, the R geometry package, GNU Octave, the Mathematica Delaunay interface, and SciPy’s scipy.spatial are all related to Qhull or make use of its capabilities. For visualization, the text mentions that Geomview supports 2D, 3D, and 4D output, and VTK may also be worth considering.

Documentation, Pricing, and Ease of Use

The project provides a manual, README, FAQ, quick references for programs and options, a function index, installation instructions, performance notes, and many command examples, so the documentation is quite substantial. That said, its usage style is fairly traditional and command-line oriented. Options such as Qt, QJ, FS, and FA require some learning, so the entry barrier is not low for beginners. The text does not mention any commercial pricing, and since source code, COPYING.txt, and a GitHub link are provided, it can be understood as an open-source tool; however, the specific license name should be confirmed in the copyright file.

Pros, Cons, and Who It Is For

Its strengths are mature algorithms, coverage of classic geometry tasks, embeddability, and extensive ecosystem validation. Its weaknesses are limited support for non-convex meshes, constrained triangulation, and large-scale high-dimensional triangulation, along with a less modern SDK experience. It is well suited to research software, numerical computing platforms, geometry algorithm experiments, and developers who need a stable low-level geometry kernel.

Access from China

The text does not provide information on network availability, mirrors, or payment, so access from China is unknown. If qhull.org or GitHub access is unstable, users may consider relying on SciPy, MATLAB, R, or Octave distributions that already integrate Qhull, or evaluating alternatives such as CGAL, LEDA, Triangle, and VTK.