we could use a BSVD, since XT and BU have failed us.
But even if you assume the most optimistic case and bsvd has better developers, a more scalable design, and less bugs, all of the same problems still persist and Bitcoin SV would still be limited by the existence of a "reference client".
No matter how well done a bsvd design could be, it still would have bugs and it still has some scaling limitations. Knowing this miners would all choose to run bsvd to maintain compatibility with other miners, rather than run their own clients. As a result miners themselves become limited by any scaling limitations in bsvd. Worse there is only one group that can scale the reference client and economic incentives to develop other clients are removed because miners won't run them due to the first point.
A published spec and lack of a reference client changes all of that. We'd see the emergence of multiple clients competing for miner adoption and miners being willing to run different clients. The existence of bugs would remain, but they wouldn't come to define the coin. Instead when found there would be a rush to fix them in a client to remain compatible with the published spec.
Personally I think the best thing nChain could do for BSV is to clean up the reference client just to the point that it is compatible with the v0.1 spec again, and then publish the spec and invest no further. At that point it is up to other miners to develop their own clients and we'd see a the emergence of a diverse ecosystem, something we talked about in 2013, but was proven not possible when both XT and later BU were rejected for some of the reasons above.