My general thought on the chipping of the glass was it didn't really matter as much as you think. It's actually an extremely thin copper trace sandwiched by two pieces of glass. So even though you don't have a clean, flat copper surface exposed like on the flexible cable portion the liquid circuit writer will be pushed up against the glass sandwich and thus the microscopic trace in there. At least that's my theory. :)
I spent 1+ hours trying to chip glass to get a flat trace under there without any luck. Only a couple of the chips resulted in enough copper where I could get the tip of my multimeter to get a signal. I then gave up, did the circuit writer pen, baked it under a 60W bulb for a couple hours to dry it, then tested everything and it was good. You do really have to wait a long time for it to dry and become conductive, particularly if you put it on real thick.