Designing Langmuir probes which can survive the high power fluxes in current and future tokamak diverters is a difficult problem. Conventional probes, which are tired and proud, will no longer be adequate in long pulse machines, even if they are carefully constructed and provided with extra thermal mass, Pop-lip probes represent a fall back solution with generally unsatisfactory duty cycle. Flush mounted probes are the most robust solution, but their interpretation has difficulties which are different from, though not necessarily more severe than those of proud probes, Pop-up probe:, and a 'checkerboard' probe, both recently commissioned on ASDEX Upgrade, verify that flush mounted probes, when properly interpreted, yield the same results as proud probes. The latter also allow detailed studies of the physics of Langmuir probes at grazing field incidence, in particular the separation of angle and size effects. it is found that the angle of incidence has a large effect, independent of the projected area, for which no theoretical explanation is available.