How to cancel optimal extraction ?
With version 2.0.x of calfuse, you have to change the key_word ``OPT_EXTR'' to 0 in the parm*.fit files.
With the version 2.1.x, you just have to change the key_word ``RUN_OPTI'' to ``NO''.
How do calfuse get the estimation of the flux Fx'?
By default, optimal extraction is processed 3 times. For the first optimal extraction, It is using an estimation of the flux given by a box extraction (the boxes are coming from the spex*.fit files). For the others, It is using the flux obtained by the previous extraction.
How do calfuse correct bad pixels?
When there is a bad pixel, the pipeline change 
to 0. As a result, the normalization factor ( 
) is decreased, so the flux is adjusted to compensate the dead pixels.
How do calfuse correct worms?
The first part is to detect where the worms is. This is done by a 2D cross correlation between the data (after subtracting the flux) and an image of a worm stored in the calibration file.
The second part is to correct it. To do that, we decrease the probability ( ) on the pixels were the worms is.
Do optimal extraction emphasize fixed-pattern noise?
Theoretically, yes. But fixed pattern noise is a big issue only with high signal to noise data, and the optimal extraction is not weighting anything in that case. The weighting become important when the S/N ratio is very low (close to 1 or 2).