On Sat, 07 Jul 2018 19:52:54 -0400, Paul wrote:

There is a stark difference, between how Explorer.exe
works when dealing with files, and how a shell-level
FindNext works.

I can do "dir" in a Command Prompt window, and easily list
8 million files in the window. They're sorted in alphabetical
order in three columns. That means the command knew pretty
close to the beginning of the output, what it had to do.

Explorer.exe on the other hand, fails in the simplest
and least taxing of situations. It should be a
"case study for CS class" for somebody.

What I can't figure out, is how "dir" and FindNext
are able to sort a 40GB $MFT in no time at all, and
start outputting file names... in alphabetical order.
It's fast enough that "there's got to be a trick".

Whereas with File Explorer, around 1 million files is
a practical upper limit for "expecting the window
to ever paint". And you can find/create failure
test cases for Explorer with as few as 60,000 files
in a single folder (the window will have a "busy icon"
forever).

I've wondered the same things and I have no real answers. If they really
wanted to, I'm sure they could address each of those performance issues,
but I can only conclude that they don't want to.