I have lost the details but I can recall that PD0-7 drove the address lines, the printer control lines drove D, RAS, CAS and WE,  and a printer status line read the data.

I drove it using my 8088 based Amstrad PC. 

The cells need time to collect light and discharge, so instead of waiting for a row to discharge I alternately charged a row and read a row that had been charged N rows before. This allowed data to be read as fast as rows could be charged. Varying N adjusted the exposure time. 

You can't do this with modern PCs because they don't have parallel printer ports and their operating systems would not allow you to directly read and write to I/O space.