Well, I've done it again.
In short, this patch adds:
a) the cutoff and iteration additions to the command line.
b) outputs more information on the setup and the commands to verify they took correctly.
c) breaks the encoding option to be optional, with a default of the system default encoding if not specified.
Note, however, I still belive this really needs to be strongly suggested that it is provided by developers or people who cross international boundries with the models.
Until we get a good map of language to encoding for the input files, we can't be sure the parsers will properly interpret the training files.
Lastly, this patch closes a few todos in the file which were for support for the cutoff and iteration limits.