BornAgain 20.2 has been released. It resolves a number of bugs. Full compatibility with 20.0 is maintained for Python scripts and for GUI project files.
Release number 20.1 has been skipped: While we had already published some artifacts of 20.1, another serious bug was reported to us. So we repaired that bug immediately, and restarted the release procedure with a fresh version number.
That long-standing bug caused incorrect vertical positioning when option UseAverageMaterial was applied to multi-particle layouts or particle distributions. It was discovered by visual inspection of SLD profiles.
Another long-standing bug was the incorrect absolute intensity of scattering from rough interfaces, following a misprint (k^2 instead of k^4) in a paper we had relied on. Now, finally, scattering from roughness and from nanoparticles is correctly computed on the same absolute scale.
An important hotfix is the repair of the GUI support for angular reflectivity scans, broken by a bug introduced in 20.0.
We still do not support roughness in polarized computations. At least, if this combination is encountered, we now raise a proper error message.
Most microscopic lengths in BornAgain are internally represented in units of nanometers. This includes probe wavelengths, layer thicknesses, and nanoparticles sizes. However, following established practice, scattering lengths are in angstrom. To avoid confusion, this is now clearly indicated in the GUI.
Thanks to Liz Margarita Montanez Huaman and Timo Fuchs for bug reporting.
Ammar Nejati, Mikhail Svechnikov, Joachim Wuttke
May 4, 2023