There is nothing to discuss - we should never use anything exported from Poser in commercial products. This also means we cannot do any grouping in Poser, forcing this part to be done in other 3D software. The only ones I know of that support Poser/DS poly groups are Blender, Modo, and Lightwave. All others, including all of Autodesk's lineup (3DSMAX, Maya, etc) do not support unimesh. If Poser creates a new OBJ, we just have to throw it away and replace with the original unimesh one.
In a way it makes sense that it's easier to make DS contents to work in Poser than the other way around. Poser has been making an effort to maintain backwards compatibility in every version, so even ancient DS CR2 plugins still work nowadays because of that. Conversely, DS has been changing with little regards towards backward compatibility, up to the point where both platforms have become mostly incompatible. For many years SMS was asked to keep compatibility with DS, but that was impossible to accomplish simply because DAZ has decided at some point to depart from Poser compatibility, and since then it has never looked back.
The most blatant case was when so many people kept asking SMS to support the Genesis platform in Poser, when not even DS1, 2 or 3 could do that. DS4 was rewritten specifically to support a new proprietary and closed platform, so there was no way Poser could support that - ever. DAZ would sue anyone who tried that because it's their proprietary technology. In addition, people seem to forget that the 2 companies were competitors in the same field. It's like asking Coke to sell Pepsi. Nonetheless, we will still see PLENTY of people asking Poser to support Genesis. I used to be one of them, but at some point I've realized what I was asking for.
Nonetheless, you are right - it is easier to start from DS and then do the Poser version. But mind you, Poser's incompatibility with unimesh pre-dates SMS. It's not a bug, but instead a design choice back from the very first version. It's deeply integrated into the Poser's inner works, and changing this would mean rewriting Poser from scratch. Larry Weinberg told me that himself, and I could tell that's nothing to be taken lightly. The Poser core code is ancient, and it shows. They don't want to change it, mostly because it would be a major overhaul, it would take a long time, it would be super risky, and it could potentially break everything else attached to it (all the other "Rooms"). Larry has considered doing this in Poser 12, but those days were gone after the dev team was fired.
Rendo will certainly avoid such drastic actions for now, when Poser is still new to them. However, I would prefer they released Poser 12 with all the many critical bugs FIXED instead of stuffing shiny new features like SMS used to do, keeping the core under the hood broken. Having that said, maybe the chances are that Poser may never support unimesh internally. That will depend on how seriously Rendo will take this. Maybe it's just me, but I believe that Poser will remain in significant disadvantage for as long as it's internally incompatible with unimesh. This alone will keep it incompatible with everything else, not to mention it breaks part of the content creation pipeline.
Sorry this got long. There is no simple way to put it.