Dynamic RAM cells store electrons which can be knocked out their capacitors by photons. They can thus be used as simple photo-sensors. Light can also interfere with the rest of a DRAM chip. 

There are several designs based on this idea. 

• The Cromemco Cyclops, circa 1975, used lots of TTL to drive the chip and produce an image that could be viewed on an oscilloscope. The computer interface was S100 and the code was in 8080 assembler.
• A company called Idetix produced a slightly modified DRAM with metal layers to shield the part of the DRAM where light-sensitivity would cause malfunctions. Their design used lots of TTL and interfaced to the 6502-based Apple.
• Electronics Today International produced a 4164-based design interfaced to the IBM PC ISA bus, and used lots of TTL.
• Your Robot magazine produced a design interfaced to the Sinclair Spectrum bus. It used a Z80 PIO and a 4116 chip which required -5V and +12V rails.

I had moved on to the IBM PC by this time, but did not want to wire up the E&WW magazine design and put it into my Amstrad PC. I had the idea of driving the DRAM chip with the parallel I/O of my printer port. Obviously software is slower than hardware, but this was never going to be a high-performance camera.

I got my design detecting light and dark, with a 'moving shutter' method reading rows while waiting for other rows to discharge. However, the pixel pattern did not match the E&WW design logic. I did not have the time or motivation to find the correct unscrambling, as I felt I had proved the principle.