So to step in here as a developer, and one with Unity experience at that...
The "lag monster" would require a server rewrite, not a client rewrite. Granted, we're aware of the stuttering that happens in the client, but that's not where the majority complaints of lag come from - especially on Xanadu.
What he's meaning here is that the Wurm client relies heavily on server communication. Unity isn't very strong in that area, but that doesn't make it impossible to do. I've not had the (dis)pleasure of handling Unity's built-in networking classes, but from other developers I've worked with - it's a beast. It would likely require us to write a lot of custom code just to handle the server communication as-is, which would make it a non-trivial task.
As I said above, it's a non-trivial task, which you've basically mentioned here. It's definitely a resource consideration given our small team and the demands by the player base to keep content fresh and bug fixes coming when we can (and when we make new ones!). The main disagreement I have here is that it'll be 5-10X better. Wurm's client is really performant for being a Java-based 3D application. Yes, Unity would likely be a bit more performant, but not by the degree you're stating here. It would also be at the cost of a major endeavor that I'm not sure Wurm would survive - it would require us to divert all developer time on a new client for quite some time. Likely a year or more. That's not including major fixes that arise from major bugs discovered during that time. Also consider that the majority of our developers are Java developers, not C# developers. So that year can easily turn into far more with just the learning curve of the Unity API and the new language.
I've joked about doing the Wurm client in Unity many times, and I've thought seriously about it more than a few as well. What stops me is knowing the time sink it would become in just recreating our Client/Server API alone. It's a tricky thing to get right to begin with.
I hope that helps to put this notion to rest, at least for now. Maybe when Unity developers volunteer their time freely in the dozens, we can revisit the idea!
OwO notices your reply...
Going to a voxel-based system would definitely be a complete client and server rewrite. Not just something easily hacked up in Unity. Our entire world is far too entrenched in a tile-based map to support voxels.