There are assumptions elsewhere that, when converting an X3F file to DNG, a partial conversion of the raw data takes place, and the original sensor data is not preserved. In effect, the assumption is that the equivalent of "convert to linear image" is performed, even when that option isn't set.

In contrast, I observe that compressed DNGs from X3Fs are significantly different sizes depending on whether that option is used. So presumably the data, or at least the order of the data, is different in those two cases. (Uncompressed DNGs from X3Fs are the same size in the two cases).

What is fact, what is myth? Just curious.