Currently Landmark is in early alpha. And based on how things have been going, they have a lot of work to do to get it ready for large amounts of people.
First up was world crashing, and people losing their claims, items they built on the claims, crafting stations assembled and put there. The reason? Oh in our internal testing we never had more than a world and 1 island setup. Scaling it out to 2 worlds and multiple islands failed because their code had no clue what to do with islands beyond one. So anything created on the 2nd world or any island beyond the original 1 on the first world, would get lost, or disappear.
So they fixed that, or so they thought. At least peoples claims were being held. But still anything people created on their claim("voxels") would end up disappearing when they logged off.
The reason was that creating the claim db for each world, they introduced a new argument to db access routines, but the "voxel" routines hadn't been made to use separate dbs for each world. So when they went to update the db, it was using random indexes for which db to put data in. This of course caused huge db corruption, so they of course did a wipe of the db, tried to give people back the pre-order items lost, and let people start over again.
They decided to use a single db for all the voxel data. But what happens when you get the initial 500k people there, with 10k-20k voxel items on their claims? That's 10billion items a single database is going to be managing and trying to access in realtime. With out separate dbs for each world, or a good database with partitioning to make look ups faster, it's going to suck.
I submitted some ideas about database design to help alleviate that issue, lets see if they stop to think about it.
But overall, being able to EQ:minecraft has been fun. And when the servers are up long enough to actually collect and build, it's rewarding.