As a professional grade MS microphone it will be preferable to have Mid and Side output directly. You might consider offering an optional MS>LR matrix adapter in a short amplifier body which would be inserted between the the microphone and mic-cable if direct LR output is desired, as that approach would eliminate the complexity of housing a switch and matrix circuitry in the microphone amplifier body while providing the same functionality.
End-address is the prefered format for the type of live music recording most folks at Taperssection are doing (and I presume would also be prefered for boom use), because that orientation is easier to aim appropriately with a windscreen in place and can be considerably less visually intrusive atop a microphone stand. However, side address achieves closer coincidence in the stereo plane and may be preferred for studio and spot use. I'm not sure to what extent the offset between the capsules in end-address orientation actually affects the hf polar patterns and resulting sound in a practical sense. You probably have a better sense of the significance or lack of if of that.
CSM28..... kind of puzzle idea..... Omni center..... or..... can it be a variable directional mic when facing omni and figure-8 on the same direction? User can get any kind of directivity, and it will be a true omni and figure-8, not back to back cardioid.... Especially the omni one, the low frequency response will not be affected by distance.
Yes, some very interesting applications with that one.
Such a microphone could be quite useful in the multichannel array techniques I tend to champion around here, which are simple enough to be setup using a single microphone stand. The approach builds upon a basic foundation of a spaced pair of omnis plus a center directional microphone. One of the first improvements on this the substitution of a coincident stereo pair for the single center microphone. An end-address CSM78 or CSM58 would work nicely in that position in most cases.
A step further would be substitution of pair of CSM28 for the spaced omnis, with the end-address omni Mids facing directly outwards to either side, such that the omnis are pointed 180 degrees away from one another and the fig-8s are facing fore/aft. In that configuration the microphones are setup as sideways-facing end-address Mid/Side pairs, but would be used more like the variable-pattern technique you describe, providing front/back pattern control over the omni pair. The recordest gains greater control over the direct/reverberant pickup ratio in the spaced omni pair, and can adjust that balance to taste after the recording has been made.
In total, that's three Mid/Side microphones feeding 6 recorded channels, providing control over the coincident stereo spread across the center of playback image in combination with control over the nature of the ambient pickup from the wide omnis. This relatively simple setup is also leverageable for surround by forming a rear-facing, wide-spaced cardioid pair from the omnis, subsequently routed to the surround channels. This is somewhat similar to your VR ball microphone, excluding Z-axis and substituting the wider A-B omni spacing for a spherical baffle.
Interchangeable Mid capsules would make these microphones fantastically flexible and attractive, and would be a unique feature as far as I'm aware, but is obviously only applicable if your system is designed as such. I don't see the lack of that feature as being a problem as long as the pricing for individual complete microphones remains attractive.
Thanks for the ear and best of luck moving forward!