Do make sure the input is topologically correct if you expect a watertight output
Do [omit_LAS_classes] of points you do not use, this improves speed a lot and can help overcome configuration mistakes
Do make sure all [input_polygons: datasets: lifting] class has a corresponding [lifting_options] class (e.g. [input_polygons: datasets: lifting: Building] & [lifting_options: Building])
DON’T
Don’t expect magic
Don’t combine [simplification] and [simplification_tinsimp] settings, the latter is always preferred
Don’t use lidar [thinning] setting for other then testing to improve speed since this is a simple skip amount while reading points
Don’t forget to configure a value in [use_LAS_classes] when using [use_LAS_classes_within], it defaults to all classes