Elkcub,
You are absolutely correct in your assessment about optical axis parrelism and the need of the eyes to look progressively more off-axis the closer you look. Likewise concerning the differences between porros (except for the rare ones with underslung prism assemblies) and roofs as well as the greater demands for eyepiece design that this places on porros if you want them to perform well at close ranges. I just measured my 10x42 SE which, for my 65mm IP setting has optical axes 130mm apart. For a Leica Ultravid 8x32, 65mm IP gives 65mm spread between the axes. Nikon x32 HG's have optical axes closer than the IPD, and Swaro EL range a little farther. It would be very simple to measure how far off-axis the image will be situated at different distances with a typical porro and a typical roof providing you view stereo with both eyes and attempt to place the image symmetrically for both eyes. I suspect that many people with a pronouncedly dominant eye will tend to place the object of interest closer to the center for the dominant eye and further off-axis for the latent eye to make close range viewing easier.
The SE is one of the few birding porros which has a large enough sweet spot to allow for satisfactory close range viewing, but as John said, even then it is much easier if you adjust the IPD in these situations. Personally, I would prefer the SE to have underslung prisms, even at the cost of reduced stereo perception. This would bring their parallax error down to the same level as it is with roofs, and would also make them easier to hold at least for me.
I suspect that one reason roof prism binoculars have gained such popularity is that a majority of people try out their prospective choices inside a store, where distances are always closer and roof prism binoculars enjoy their inherent parallax advantage to the full.
Actually, the reason for adjusting the IPD is not so much to make the images overlap better - since the effect on this is marginal - but to align the exit pupils with the pupils of your eyes. As you look closer, you progressively cross your eyes more, whereby your IPD gets narrower. I just measured the most comfortable IPD for both the SE and Leica at 3 meters distance, and for both it was down from 65 to about 62mm (the SE does not really focus that close, but even with an out-of-focus image I could determine the best IPD very easily. A three millimeter difference in parallax is almost insignificant with respect to image overlap, but proper alignment of the eye to the exit pupil is important for viewing comfort.
In theory at least, it would be possible to design a binocular where the collimation of the optical axes would adjust together with the focus. However, this would no doubt result in an exceedingly expensive and complex design.
Kimmo