hey
i guess this raster one is for those cases where you want to make sure that no subtle rounding errors can ever occur. For that it works with integers internally which is like having a raster of valid values, with lots of invalid/impossible to reach values in between.
both counters seem to work with non-integer step sizes, but still the raster version promises to work with integers inside. for that to work there is probably some back-end necessary that maps the internal integer representation to the actual value range. i guess having a changing minimum somehow proofed to interfere with that mapping in a way that feels wrong?
But anyway. i guess most of the the time you have a static minimum, not?