From APMR's and other replies, I believe that the problem is less about procedure design and all about aircraft equipment. Aircraft with basic displays do not have any form of vertical representation of glidepath, while the suggestion of a continuous (digital) display of correct procedural altitude would probably add more to pilot workload than a simple glideslope indicator.
Problems associated with incorrect setting of altimeter only strengthens the case for the FAF to remain:
1. To act as a final crosscheck of altitude/glide path and highlight altimeter setting errors prior to the MAP, bearing in mind that altimeter errors are not linear and a final crossheck as close as possible to the MAP is desirable.
2. To provide another critical step for obstacle clearance, where obstacles in the approach path may be limiting.
On both these points, I fail to understand how the FAF in an RNAV/GNSS approach differs from any limiting altitude step in a DME/GPS Arrival procedure or from the FAF/glideslope height check in an ILS approach.
On Point 2, should the FAF be removed from an RNAV/GNSS approach where there are obstacle limitations, the minima may be raised and the benefits of this type of approach unnecessarily restricted.