Notes

Let me elaborate in more detail upon some line items listed above.

1. Smaller, lighter body. Even now, after two years with Mark II, I'm enjoying the Mk. I more in direct comparisons. Understandably, some users will feel just the opposite. (Note, however, that almost 30% of what Mk. II gained in weight, is due to the new, larger battery.)

   If it turns out that shrinking Mk. III back to the old size cannot be easily done, I won't be complaining.

2. Swiveling monitor. There is one class of people who like the swivel type (two degrees of freedom): those who write camera reviews based on counted tick marks. Swivel and tilt must be better than just tilt, right?

   Not quite. In the landscape orientation the swiveling screen makes the whole setup just too wide, sticking all the way to the left. It can be used in portrait, though (unlike the tilt-only type), but I've never seen anyone doing that, and for a good reason: it becomes quite awkward. No, thank you; I'd rather do without.

   A double-tilt solution, allowing also for tilting the screen in portrait orientation, is also available (check some Fujifilm models). Given the right practicability and durability it might work here. Anything but the tilt & swivel type for me, please. (Your mileage may vary.)

3. Higher pixel count. Let's keep in mind that the increase from the current 20 to expected 24 MP (20%) means just a 9.5% higher linear resolution (nominal). A 40% increase (to 28 MP) translates into 18.3% linear. Still whatever choice Olympus makes here, is fine with me.
4. Global Shutter: this, I believe, requires each row of photosites to have its own charge-pickup circuitry; a complex and expensive solution. What we gain in this deal is getting rid of geometric distortion effects caused by subject or camera movement, and faster serial frame rates with electronic shutter (now limited just by the shutter speed, no longer by the frame scan rate).

   Higher serial rate and no shutter roll are nice to have, but not essential in still photography (more useful in video), except that it affects another feature, see the next note.

5. Olympus High Res mode (E-M5 II, E-M1 II) uses electronic shutter to capture eight frames, applying a pixel shift in two square patterns and then combines them into one RGB image with pixel count of 50 or 80 MP. Currently this requires using a tripod (or a similar support). A global electronic shutter will make handheld High Res a realistic option.

   The camera and the subject must hold steady during those eight exposures, and in order to see any benefits of the HR Mode, the magnitude of camera shake must be, roughly, half of that allowed for non-HR pictures.

   With the 16 ms frame scan time in Mk. II, this means handholding an effective exposure of 128 ms (1/8 s) added on top of the actual total exposure (say, 8 ms or 1/125 s for a shutter speed of 1/1000 s). That's a total of 136 ms, without the benefits of image stabilization (the sensor-shift motor is busy doing its square dance).

   Using the global shutter reduces the first component from 128 ms to zero, while not affecting the second one. The total drops from 136 down to 8 ms (roughly 1/8 to 1/125 s). This is much more handholdable, and it gets even better at higher shutter speeds.

   I'm simplifying things a bit (for example, neglecting the time elapsed between exposures), but this does not affect the general pattern. With some caveats, handheld High Res becomes possible.

6. With autobracketing active, It is up to the user to make sure that the right number of frames is shot. or that serial drive mode is set on before and off after the bracketed sequence (which will terminate after the preset number of frames).

   If there exists a more cumbersome and unsafe way, I'm not aware of it. Olympus introduced it with the E-500 in 2005; the original reasons are lost on obscurity (limited buffer size?).

   What makes this even more annoying, is how simply it can be fixed. Each kind of bracketing uses some parameters (like number of steps and step value. Adding just one more: time spacing between steps (or Off) would do the trick.

7. Focus Bracketing, or rather option known under this name, is really no bracketing at all. The term "bracketing" implies using values [ [ [ on both sides ] ] ] of that set by camera's circuitry (or chosen by the user), while here the camera starts from the focus "as set" and follows with a number of frames (up to a thousand) focused increasingly further.

   This is exactly what Focus Stacking does, but the final stage: merging the individual frames into a composite (which can be done in postprocessing, so that the outcome will be the same).

   A real Focus Bracketing function does actual bracketing: shooting a frame or more on each side of the measured/chosen focusing distance.

   Last time Olympus did that was (again!) the E-500. Interestingly, it worked only for AF activated in MF mode ("on-demand" button), but it offered some protection from mis-focusing in either direction.

   I've used this function a few times on the E-500 and, indeed, it is worth having. I am missing it on current Olympus cameras.

8. There are seven auto-bracketing modes in current Olympus cameras. Two of them deal with White Balance (in different color dimensions). I consider them useless: only one frame is really taken, and "bracketing" is done by multiple raw-to-RGB conversion. This is much better done in Olympus Viewer postprocessing, and I would be offending your intelligence, explaining why.

   Art Bracketing is neither art nor bracketing: again, one frame is shot and then up to 20 Art Filters applied in RGB conversion (all except ones you disabled). People, don't you have anything better to do but polluting my camera's interface?

   I would be happy to see WB Bracketing and Art Bracketing to be taken out of their misery, or, at least, being able to remove them from the list of selectable bracketing types. The latter option would be nice for all types of bracketing.

9. Better battery. The numbers quoted in the table (see also here) seem to indicate that it should be possible to make an improved version, increasing the charge from 1720 to 2200 mAh (and the weight from 74 to 95 g), within the same form factor.
10. Art Modes — are you serious, Olympus? In a pro flagship camera? Yes, I've seen some writersce, jumping up and down in joy with any new Art Filter; no comment on that (I'm desperately trying to be nice).

    More, this feature pollutes not just one, but thee areas of user interface: Picture Modes, and bracketing. The designers have all bases covered.

    The best way to fix this feature is to kill it. It does nothing which cannot be done better in postprocessing with Olympus Viewer or third-party software.

11. Diffraction Correction by Deconvolution. If done right, this would be HUGE, as diffraction is a lens flaw which cannot be fixed by proper lens design : it is defined only by the focal length and aperture used. Also, it affects the μFT format twice as much as it does the full frame.

    S very little of what you can read about it in general-audience sources is written with any understanding. Don't blame the writer; it is just one of those subjects.

    Being myself semi-literate in the subject (well, that's way better than illiterate but more dangerous), I wrote a small   article on this subject (or, at the moment just the first part). At least I've tried.

12. The extra 4 bits result in 16-fold increase in number of distinct signal levels per photosite (or per RGB channel). Some of this gain will be used to accommodate higher dynamic range of the sensor); the remainder — to reduce signal level granularity.

    16 bits-per-channel RGB files will be able to make a better use of upcoming HDR displays (as opposed to HDR emulation on non-HDR displays which is the norm now).

    Among a few 16-bit RGB image standards, JPEG XR (.JXR), originally by Microsoft, seems to be a major candidate, being versatile, efficient, licence-free and open-source.

13. The size of this menu is due mostly to the huge number of options and/or parameters, accessible to the user in Olympus cameras. At least some of its complexity and inconvenience, however, is due just to a sub-par tree design. A better job can be done.
14. To provide a more "serious look", even some compact super-zoom cameras have a fake "prism hump". I expect it to take another ten years before this fashion disappears.

Most of the list items marked as firmware changes do not actually have to wait until Mk.III — they could be included into one of firmware updates. So far, however, the updates did not bring anything significant to Mk. II, with one notable exception: the small AF target option. Under development...