To be technically fair, I believe the Windows 95/Win32 Registry was one of the things backported (from the "sideport") to Windows 3.1 in the strange Win32S back

To be technically fair, I believe the Windows 95/Win32 Registry was one of the things backported (from the "sideport") to Windows 3.1 in the strange Win32S backport, so it is very possible a developer's first experience of the Registry was Windows 3.x, late in the Windows 3.x life cycle. To be fair to third party developers, and to make it clear I don't assign them much blame for "missing the memo that the Registry was not for application data", the mistakes were clearly coming "from inside the house": never forget that the first Win32 developers were dogfooding teams at Microsoft. If even the Office teams never got the memo, there's no expectation that third party developers could have psychically intuited that the memo even existed. Also to be fair to disclosure I'm sure that most of my understanding of all this is badly paraphrased Raymond Chen.