The advantage of including everything, or the problem with trying to cherry-pick, is that many problems are only the result of the partial but incomplete sets of interoperating components.
If service A is always trying to use service B, but you removed service B, you can have annoying problems caused by service A hanging and timing out behind the scenes or presenting buttons that don't work etc.
One way to avoid that is never remove anything, but another way is to remove both A and B.
One way to get both A and B without knowing what they are, is get everything.
And this particular method, using an actually fairly managed and graceful command (managed/graceful meaning not just removing files and editing the registry, but only running a user-facing command and interface) is like an automatic filter to include only things that don't actually break the OS.
Have you tried this method yet? Appx packages tend to be very good about declaring their dependencies so that removal is easy and safe.
They’re not suggesting deleting individual files or mucking with the registry, they’re talking about using a clean, user-facing tool just in a more surgical and thoughtful manner.
That's what I said? Did you mean to reply to the same parent comment I did? I fully agree this is a quite good trick, or at least is a solid theory, not claiming to have done it.
If service A is always trying to use service B, but you removed service B, you can have annoying problems caused by service A hanging and timing out behind the scenes or presenting buttons that don't work etc.
One way to avoid that is never remove anything, but another way is to remove both A and B.
One way to get both A and B without knowing what they are, is get everything.
And this particular method, using an actually fairly managed and graceful command (managed/graceful meaning not just removing files and editing the registry, but only running a user-facing command and interface) is like an automatic filter to include only things that don't actually break the OS.