Inventing reasons why it can't be done is a bit disingenous; it's like people who explain that putting men on the moon is now too expensive or technologically impossible.
There are probably various ways to do multipoint, and the engineering is doubtless easier than putting men on the moon and probably cheap enough to put in a $60 K (with back) Phase system, but as Doug very correctly points out, in MF there are no drop-in MF subassemblies that can be bought so an AF system implies in-house camera engineering. It is one thing to cost a custom semi-silvered-mirror-submirror-convex submirror-sensor design as a necessity for a $60K system, or for a body that is basically a giveaway to back customers.
As Doug says quite correctly, Pentax is re-using the APS tech they already have, but the specs of their sensor module mean an APS-C sized bunch of focus points. But I disagree about his secondary mirror discussion, I think it is not the weight of the secondary mirror nor the need for a custom sensor, but rather the fixed convex tertiary mirror that becomes a size and cost issue in MF geometries and flange distances, especially since the secondary submirror always has a very convenient mount-point - the back of the semi-silvered primary mirror- but one needs to create enclosed space in the camera bottom to emplace the larger tertiary convex mirror, the sensor and the optical path connecting them, and do so in a way that does not become a dust-trap. But this is just my intuition.
When Sony finally put some phase-detect subpixels on the main sensor (if they aren't already there?) the problem will get solved "for free" as regards physical gear. Of course all MF cameras, indeed all AF cameras with CMOS chips and liveview capability should already be able to do contrast-detect AF for static subjects, as this is basically firmware coding only. At some point Sony will probably also make its full-sensor stabiliser design available too, which will certainly improve handheld MF usage
The above remarks from a senile armchair engineer are probably worth as much as they cost you. Let's hope readers feel the Phase AF system is worth what they need to pay for it
Edmund
Doug & others can speak more to this, but it comes down to a technical decision.
The AF point we're use to seeing, it's a sub-assembly on the camera, with it's own dedicated sensor. So to get a larger AF area - aka more focal points, the Honeycomb sensor has to be larger, and the mirror subassembly has to be larger. With full frame cameras, this means little, but when you look at the mirror size in the XF & H bodies, it's a big a** mirror already, and adding more mass has downsides. Clustering a bunch of AF points in the current Honeycomb setup would actually cause more issue, in that if you didn't do a single point AF, the camera could bounce between two focal points when dealing with shallow DoF, and throw things out of focus. So I BELIEVE** Hass and Phase decided that they would rather have 1 really good focus point, rather than more, less accurate ones. As contrast detection AF gets better in the larger chips, and you are forced to use EVF's, things will get better.
As to the 645D/Z, since it's a body & back in one unit, they can size it to what ever works for that sensor. The XF has to take everything from a 40MP 33x44 CCD to a 100MP+ FF CMOS. Hass could do a bit more since their history has been to sell a back+body, but they still stick their light meter in the viewfinder rather than the AF sensor.