The code for opening Bruker files determines the number of dimensions by looking for acquXs files, but only looks for up to acqu4s.

the loop in findBrukerInitialPoint (Vec.java) excludes complex points whose argument > 0, but the average on the ph value is still calculated on the total # of points in the range.

Processing a Varian 3D file, for example, with acqOrder settings for a an array of phase,phase2 gives the same result as processing it with values for phase2,phase (hence you can't correctly process some datasets).

Gapsmooth operation (like other baseline operations) requires the vector to be real. Other baseline operations automatically convert vector to real. Gapsmooth should also do this.

The dimensionality of Bruker files is determined by checking for acquXs files. But if the dimension size in a file is only 1, perhaps that dimension should be ignored. For example, its common to collect 15N filtered experiments using a 2D HSQC file, with the second dimension set to size 1. This file should probably be interperted in NMRFx as a 1D file. Not doing so causes an issue when using the scanner to combine multiple files into one pseudo-2D file.