Moray, a pinion can be spring loaded into a rack somewhat to help counteract the inherent separation forces occurring due to the gear pressure angle, but you can't let the pinion be the sole support for the gantry and have the full weight of the gantry jamming the pinion into the rack and get any kind of satisfactory operation whatsoever. The tooth engagement depth will be way past the design amount, and the gears will literally eat one another until they get all the clearance they want. The R&P in a plasma cutter has to run dry because of all the flying gritty debris from the cut. My guess would be if you build a machine in this manner and put it into a test loop where the bridge is traveling continuously and the gearing is dry, simulating whats going on when cutting, your rack and pinion will both be substantially destroyed before 4 hours is up. The motion, right from the start, will be so rough as to be unusable.