I was not discussing Snapshots primarily either - I made a COMPARISON how Snapshots work to further clarify my meaning.

I think you answered my question so that only compress and shrink are affected. This does not matter, which was my point, I would say normally, in a heavy use situation. Example from my way:

- create 80 GB C:\ to host all the software & its own data what is best left to C:

- create 150 GB (or larger, it doesn't matter much, because extending is trivial to the second disk) another virtual-disk for data. Put ALL the user data there.

- use non-preallocated, in slices

In this scenario, you never really need to shink anything, because disk space allocation is what you ACTUALLY use. If you delete something, like 5 GB of software delivery data, your VM OS will see that as free space, which you can reuse. You don't need to compress either, because no heavy user has nothing else than SDDs, where disk fragmentation is a non-issue.

It's good that you mentioned the limitations, but I was curious if that affects my workflow - which I would recommend in a normal high-use case. Like I said, if you revamp the VM to some other purpose, then compress/shrinking might be required (a longer discussion why really).