Growing up with MS-DOS, I knew its role in today's Windows' usage of \ to separate directories and / for command-line arguments (choices that sound quirk-y in an Unix-influenced world that uses / and -, respectively.)
I never understood why MSFT - a very Unix-aware shop, having released their XENIX a year before MS-DOS - went with such an odd choice, until I looked at the (recently open-sourced) MS-DOS source code.
The files include documentation for computer manufacturers (so they could write compatible BIOS code, customize distribution, etc.), and this piece on MS-DOS 2.0 (which introduced subdirectories) suggests that - as usual in those times - the party behind the odd decision was none other than IBM:
