LW's VDB appears a thrown-together thin wrapper/integration of the existing OpenVDB library APIs in most cases. Of COURSE it's got a horrible UI/UX, it has all the hallmarks of a "last-minute bullet-point add", likely when someone with a shred of sense realized offering _just_ a new render engine alone after years of near-silence would lead to an outright customer revolt.
A deserved revolt, considering how litte of what was delivered in LW2018 a) worked, and b) reconciled with Newtek's own communications about what was coming. Large swathes of VDB weren't really usable through most of LW2018 product cycle (and well into LW2019's product cycle), just as large sections of the "new render engine" weren't all that usable in the same periods. Whether the issues were later fixed in LW2019 is meaningless, it's quite depressing to see what all "delivered" in LW2018 actually _worked in a stable and reliable manner_ by the final update of LW2018.
It´s really a shame, I bought in to the upgrade from 2015 to 2019 version, primarely because I felt the VDB tools was exciting though crude, but figured that would be continued to be improved, is realy really a huge shame, since had they implemented some other stuff with proper multiscatter for primitive volumes, and GPU, and a bit better tools for the fluids, then they could have had something really strong and competitive.
Had I known about Lightwave getting stuck within a "no tell or confirm development, and a by by dev team wink off"..such as the current status seem to be, then I would perhaps not have invested in that upgrade.
Lightwaves new primitive volumes and how you work with them, is despite some minor detachment from the easy access philosophy with the old legacy hvs´s, Still one of the best tools to get up and running with volumetric items, it´s just the GI slow renderspeed and lack of multiscattering that is the big issue.
VDB´s a lot more missing though, and unlike primitive volume items, VDB´s would take much more effort to fix/improve.