I think that I should correct myself. In the first alternative, the
space-like geodesic should be normal to the world-line of the
observer, too. Otherwise it will not be unique.