robespierre wrote:
I don't think this "concept" is realistically possible... keep in mind that there is no standard width for 35mm cameras, so the distance of the shutter window from the 135 canister is not a standard either
Easily solvable using some of the heavy-duty flex cables available to stablish a two-parts-body canister.
And of course you know that the flexible image sensor is already on the way to the big market:
Characterization of flexible image sensor arrays with bulk heterojunction organic photodiodes New Light Sensor Mimicks Eye for Superior Camera Performance Scientists develop flexible sensor to allow simple zoom The study of fixing field curvature by flexible image sensor array Okay, it is not the same thing that would be required to mimic a 35mm film, but you could not tell that it is not possible.
robespierre wrote:
There is no way for the camera to command the electronics to begin and end light gathering, so standard image sensor designs would not get the control signals they require.
Are you sure? ...and how do you think works the Sinar's solution?
Curious?

...it is way simple, and almost every camera has one of those "ports": the flash sync connector!
robespierre wrote:
In fact there is no way for any connection to the outside world, for power, status messages, or anything else, it would all be self-contained.
Again wrong! What about Bluetooth? If you REALLY think that such status info is required, you could add Bluetooth capabilities to use some kind of pager-format display/monitor to show the basic parameters.
robespierre wrote:
The cylindrical packaging is about the most difficult to design and assemble circuit boards inside, especially when the device needs its own batteries there too.
Oh!, come on! ...military techies are using such kind of technology since ages!
...but if you don't want to discuss it, just look the commerical park:
Flexible PCBs FPC Laser Cut Flexible PCB Rigid Flexible PCB ...just pick your most comfortable SMT package size, and you're on the way!

robespierre wrote:
And the types of cameras pictured in the re35 site wouldn't work anyway: Nikon F3, for example, requires film or something with the exact reflectance of film in order for its light meter to work (light is bounced off the film surface).
Yes, but this is just the case for Nikon F3, an excellent machine but no too popular and not so big in terms of global sales compared with the rest of 35mm gear if you join every camera model produced by every company along the history...
...all in all, I would not tell that this is an unsolvable problem, but probably just one of the more difficult ones. You know, you could just improve your sensor to mimic the reflectance from a given well known film... which could probably by the way benefit also all the rest of cameras, having the closer thing to the real one.
In fact going even beyond all of that, I'm starting to think that you could even have a use for all those spiral-spring-loaded mechanisms around! You know: use a dynamo with them to store the cinetic energy, like the
BigShot does, reducing the cells/batteries consumption!

Giving another take over this latest point, you could replace the dynamo with a miniature piezo recovering the energy from the film-advance sprocket.
EDIT: several typos. Added more info.