This argument, as far as I know, has come and gone. I believe the concensus amongst aviation professionals (NOT just pilots) is that while it may sound good, in practice it's not that great. Here's an example:
Prior to a fatal accident, a crewmember reaches over, puts his hand on a switch/button relevant to the cause of the accident. He pauses, seemingly unsure of what to do, then takes his hand away. Now - did he actually flick the switch/push the button or not? Is the camera going to be have sufficient resolution (ie be expensive enough) to pick it up? Will the camera be able to see whether a light is on or off in direct sunlight (a tough job for a camera, given the large contrast between light and shadow in an airliner cockpit). Will that system be tied into the DFDR, so the switch position/light could be independantly verified? If the DFDR is good enough, do you need a camera? A seemingly trivial point perhaps, but possibly adding more unknowns.
Plus, along the lines of what I already stated, I don't want MY family seeing me get ripped apart if I make a total botch of it, thank you very much. Not that I plan on doing that of course, indeed quite the opposite, but then neither did so many dead pilots before me.
In any case, there are VERY few totally unsolved airline accidents. Adding a video recorder (or two) that can record at enough detail to be useful, while still having a crash and fire resistant recording component, will just add an enormous cost (this being the aviation world after all) for less than zero benefit. IMHO. I'm not close minded about the issue, I'm open to arguements for it - but so far I remain unconvinced.