Dear ZmnSCPxj,
I think it is very important to start thinking about extending the protocol to support multiparty channel constructs. As you know I recently pointed out the advantages of multiparty channel constructions for payment reliability and service level guarantees of payment channel networks. However I have not quantified the downsides of such constructs with respect to on chain bandwidth and costs of unilateral exists yet. So I am not sure yet about the tradeoffs that come with such constructs.
Additionally I have one concern about your proposal
While this makes sense from an engineering perspective I fear this proposed solution is not optimal. From a payment delivery perspective one wishes to forward sats from one multiparty construct to another but does not virtually want to fix 2 party channel(capacitie)s. It is my understanding that virtually creating two party channels would mitigate some of the advantages for routing abilities of the network that we gain from using multiparty constructs.
In any case I am looking forward to see progress on your proposal.