Maybe one should discern between reliability and longevity.
During the projected life span of a camera (lets say 10years) i think the electronic ones will be the winner. There is often not that much intricate mechanical stuff that can go wrong. Of course, there are motors, gears, levers etc but the sequence and timing is largely
controlled by microcontroller and not by small moving parts that slow down when lubricant gets old/cold like in an leaf shutter. Also they sometimes use better materials because they are newer. I would bet that the latest leaf shutter blades are far more durable than the ones from the
60s.
But when long term maintenance/preservation comes into play, the mechanical ones could be easier to handle.
A lot of electronic faults can be repaired easily, even if the exact same spare part is not available, one can use functional equivalents. (Diodes, Transistors, plain contact problems, cracked solder joints etc). But it takes specialists to find those faults, the layman usually
wreaks more havoc during repair attempts in electronics than the fault itself
The real showstopper are custom programmed or custom made ICs which are unobtainium
and it would require a sometimes prohibitively huge amount of reverse engineering to duplicate the functionality.
I would feel confident for example to design a replacement PCB board for a Hasselblad 203FE which largely emulates the original behavior. There is not that much stuff involved. Basically light metering, a bit of display controlling (that would most reverse engineering),
some buttons and communication with the lenses (also reverse engineering with a functioning body/lens) and state machine programming. The shutter itself is just 2 small solenoids for curtain 1 and 2 as far as i can tell.
BUT: Nikon F6 replacement electronics? I would never even try that. Far too much peripherals (AF/exposure) to talk with, when some of those are damaged it cannot even be attempted.
There is some technology level up to which one can fully understand all the interrelations between the components.
Mechanical "unreliabilty" very often stems from neglect, lack of service, old lubricants, dirt etc. When a camera is 50 years old, it seems that it has to be stripped down far more to clean out each and everything than normally done by service technicians.
Nice example: Everyone knows about the mirror dampening foam strip against which the mirror slaps. But e.g. in Hasselblads (at least without the GMS) there are 3 foam cushions beneath the mirror. If they deteriorate after some decades, the mirror wont be held properly in place anymore and
random focusing issues arise. So after a long camera life, it starts to be more of an restoration than just service/repair.
Dirt and old grease is a good lapping compound, often to be seen in old pendulum clocks that never got serviced. Some day, they stop. People increase the weight and they run again. Then they stop completely and they go to the clockmaker and say: The clock stopped working, but there
cannot be that much wrong, it worked flawlessly for 60 years! End of the story: Pivots are worn out, bearings are worn out and often worked out to some oblong shape, gear teeth are worn... Totally trashed and it needs a lot of work to get it back to life.
If the clock would have been at the clockmakers for servicing every 5-10 years, the wear would be far less and it could continue its service life for a long time.
Same with cameras.
The good thing with purely mechanical cameras is that they usually consist of only metal and that should be very stable in the long term but they have to be properly CLAd more regularly.
What i hate most about modern cameras is that stupid rubber material that gets sticky and deteriorates. I'd rather have plain metal, even if its not that comfortable.
Regarding 3D Print: Proper machine tools are alien to many people and then 3D printing is brought into the field as solution because its a new trendy technique that is accessible to a broad audience. But they overlook the fact that many camera parts have extremely tight
tolerances and surface roughness to meet and i doubt that it could be achieved by 3D printing even in the mid-term and with professional machines. If someone has more information about what is possible regarding finish and tolerances today, i would appreciate it.
A lot of stuff could be printed even now, for sure. (Cranks, knobs, battery doors, maybe even sprocket wheels) but i just dont see it for those precision machined parts that have to be made to watchmakers tolerances or that need bearing surfaces (need to be reamed).
Also the material has to be suitable for printing.
What also worries me are springs, but even for those there would be solutions to make them. But it needs specialist knowledge and equipment at a company specializing on springs.
@RalphLambrecht: Regarding point 2: I have a Pentax Spotmatic F that sat in the cupboard for nearly 20 years without being used. After that i got it in 2007 and used it for several years. Up to this day, there are no hiccups, the times are spot on, the diaphragm moves quick.
So the japanese seem to know how to lubricate (or rather NOT lubricate, as i suspect that they are using material pairings that dont need it) a camera while european cameras cause far more trouble. One would think that european cameras should be also that reliable, but
if a Hasselblad sits unused for that long, it needs a full overhaul for sure.