The thing that worries me about digital is that once you've lost highlight detail there's no way to get it back.
That's true for the moment, but engineers at a number of companies are no doubt working quite hard at figuring out solutions for that, and other technical issues. New technologies will expand the dynamic range (digital version of latitude) of sensor technology. I can imagine one theoretical scenario, for example, in which a prism or half silvered mirror splits the incoming image from the lens sending the image to two sensors simultaneously within the camera. One sensor could be optimized to capture highlights only while blocking up the shadow detail, and the other sensor could have the opposite duty. Later, the best elements of each sensor's capture would be married for output in the data stream in camera, or perhaps done in post. Who knows, maybe the same idea could be achieved using a single sensor, and done via processing algorithms. It's a crude idea that just popped in my mind, but the point is I feel confident that some new technology will eventually tackle this current limitation of digital origination.
When magazine photojournalism went from black-and-white negative to color transparency (the mags went color for the ads and slides were easier to edit than neg) still photographers finally learned to bracket exposure. Maybe the most journalistically important image wasn't exposed properly, but you probably had an acceptable back up. Movie guys can't bracket
That's true, but when shooting digital/HD, you have the benefit of seeing a quite accurate image of what you are recording with a calibrated and tented in monitor on set. You also have quite a bit of room to adjust all manner of parameters in a tape to tape color correct session (especially with cameras like the Panavision Genesis/Sony F-23, Thompson Viper, Arri D-20/21 and presumably Red). With film, you have the benefit of telecine transfer, but on set you are using your experience to gauge how you want to expose the shot. No monitor on set showing you a videotap image will give you a real sense of how your film will come back in dailies.
I don't worry about DP's on features, but I can see young documentary shooters blowing an important take when something unexpected but important happens. (It's not just dropping film costs and, sometimes, smaller crews that will move them to digital. With Red bringing in their new 5 meg body at 40G and selling the old one for 17, Red could become the digital version of the old Bolex or used Arri that a lot of documentary shooters started with.) What is the exposure latitude for something like a Red raw file?
Again, anyone young DP coming up shooting a documentary today will probably be shooting it digitally, and will therefore have a reasonable method of seeing an accurate representation of what they are recording (assuming they choose a work flow that allows for it). One key is to know what your distribution is going to be. I set up HD cameras and expose differently for 35mm blow-up for theatrical distribution than I do for video broadcast. I've been fortunate in that I have so far always known before shooting a project what the end distribution will be. A lot of projects don't know in advance and hope for theatrical 35mm distribution, but don't have that locked in at the start of principal photography.
As for the exposure latitude of the Red, I am not sure. I think there are claims of between 9 and 11 or 12 stops, but I think even 11 is perhaps wishful thinking. I did shoot some tests with it late last year in preparation for a feature film. We were deciding on formats, and that was a contender. I have to say that the post house will have a huge affect on the results you can get from that camera platform. In our case, our post house was locked in, and they had never seen Red files before. I was the first to bring files to that particular house with my tests. They really didn't know how to handle the files, and as a result, the test footage was uninspiring to say the least when we screened it at the post house (screened both digitally and blown up to 35mm). I was disappointed because the images looked very promising while shooting the tests. I feel certain that a different post house with experience handling the Red files would have been able to make the files look the way we saw the camera perform while shooting the tests. Alas, time constraints and the locked in post deal prevented me from seriously considering that system for that project.
