I had posted the benefits of the O/EVF, including no needs for adjustments, plus self-illumination, dataset displays etc. etc....in the mega-thread long ago. The technical points got lost in the excitements and many posters rather argued about JPEG v. RAW; or whether the X100 is "professional" or not...
The X100 VF optical diagram, particularly the [likely cemented] prisms immediately reminds me of typical set up in photogrammetric instruments...complicated enough even without adjustment problems. [Since you are located in the Netherlands, you must have heard or known of ITC in Enschede, once the mecca of photogrammetry.]
In photogrammetry, an image is first conceptualized as "a bundle" of light rays gathered through the perspective centre of a lens...as if the lens is a pin-hole... The science has a century of developments behind it.
[Many fundamental breakthroughs in the analogue era were actually made by Otto von Gruber of Zeiss, while Wild-Heerbrugg [renamed Leica in 1992 by its owner who also long owned E. Leitz Co.] is a more successful instrument builder until digital photogrammetry wiped out its mechanical product line...sounds ominous to Solms?]
In digital photogrammetry [based on analytical
geometry first applied by a Canadian...in 1957], image tip/tilt [from being perfectly vertical or down-looking] can be re-sampled into normal or reverted from normal into having tip/tilt [think vertical and horizontal rotation in conventional cameras], given imaging distance [flying height]...as if a prism is virtually there. An example well-known enough in RFF is keystone
correction for architecture photography in Photoshop.
The focusing ring [also a focal distance encoder or the focal distance LUT dataset display will not be possible] or AF can supply the imaging distance, which is then used to re-sample and re-injecting parallax...wholly unnecessary
if AF is used to begin with but useful as a manual focusing aid for RF aficionados.
In this "parallax wedge" application, the OVF is deemed "normal" while the EVF is treated as if tilted. The y-parallax neutralized but x-parallax injected EVF image is then superimposed onto the OVF. The baseline is the principal axes offset of the independent O/E VF optical trains.
Re-sampling the whole or partial image is only 12Mp of work, at most...thus real-time, in a science long overcame image files sizes in the hundreds of megabytes.