Thank you for that link; there are a lot of creative joints in there that I am sure would be useful for what I'm describing. However, I'm still not seeing quite what I am thinking of; all the examples had either a tie-beam or plate (i.e. horizontal members) sitting on the post, and then the rafter connection on top of that.

Is there a specific reason you wouldn't want the rafters to sit directly on the posts (with an overhang continuing outward), and support them laterally either above (purlins) or below (wall braces) the junction? So many different ideas have been tried throughout history/geography, that it seems strange this combination does not appear, unless there's a good reason for it.

Several of the tie-beam to post connections are along the lines of what I am thinking for the actual joint, but invariably there is a rafter above them (naturally) which is attached by birdsmouth or something similar, before extending into the overhang. It seems like whatever actually interfaces with the post, never extends as an overhang. I suppose this makes some sense, seeing as the joint does weaken the beam, and the post then puts the area in double-shear...but cross sectional area or knee bracing should be fully capable of addressing this. I did see references to joints that offset the rafter from the post/plate junction, so it doesn't seem like all three must necessarily meet at the same point, though it's probably more efficient load transfer that way.

The only other explanation I can think of, is that the beam/plate on post scheme, with rafter on top, makes for a much neater raising process, with individual bent or wall assemblies. My developing structure is octagonal, with the roof elements fanning out radially from the center, onto separate wall frame assemblies, which may explain why the odd concept I've described seems advantageous (I'm not sure it would be, for a typical rectangular frame)

TCB