It may be that we end up focusing on one of these wallet related features. I suspect “almost” lock free transactions would provide the most impact and immediate benefit. And would push us down the path of starting to think about more flexible wallet interactions.
If we can do this then we can take in progress transactions and “translate” them to use alternative inputs and potentially outputs. We will also be able to adjust the kernel offset per participant as necessary.
I know several people are already thinking about this so I can definitely make some of my time available to help here.
There is also potentially node related changes here in terms of identifying a “locked” transaction during initial push to the node, to allow the wallet to adjust a tx and re-push if necessary. I have not thought this through so this may not actually be the case, maybe all this should be isolated in the wallet itself.
Avoiding taking on too much at the wallet layer would free time up to continue to focus on node as you say.