IIRC, Frescox output provides a qualitative summary in stdout of the quality of the numerical computations. If this is true, should we parse it out and make it available to users? Also, performance metrics and parallelization setup information in stdout should be parsed and made available to users.
While these are in stdout, we should determine if they're also in other output files and, if so, which is the preferred source.