Overcoming 25+ years of muscle-memory

Is it worth the initial investment of time?

Maybe not to you now, but clearly it would be to all the Mac users who have raised this issue here (and those who haven't but agree).

I understand that individual Mac users are no longer a priority. All those corporate customers for whom the close button has always been on the top-right in their OS aren't having to change their behaviour. But Mac users choose macOS because they care about the user experience, about consistency of design and are prepared to pay extra for that. Maybe that's why you originally chose a Mac all those years ago, but perhaps your priorities have changed?

Wouldn't it be wiser to invest this time in a new feature or fixing a bug?

Of course you should add features or fix bugs, but this is neither of those, this is something you've changed for the worse - it is different to previous versions, breaks Mac users muscle memory, and is a move away from macOS conventions.

Making the Preferences window different only for Mac means that it will also require on-going effort to keep it this way as the app evolves,

You already have many more complicated platform-specify features to maintain, however, I'm referring specifically to the title bar / window controls (I'm less concerned about whether it floats or not). You don't seem to have any problems maintaining different title bars for the main 1Password window between platforms which are all quite different:

It is also inconsistent design within the same application on the same platform. The main 1Password window has standard macOS controls as you'd expect, but the Preferences and New Item windows don't.

However, I see from the release notes that you do invest time fixing many tiny UI inconsistencies (down to the pixel level), yet this major, in-your-face inconsistency is left untouched.

Today, our cost of writing and maintaining documentation is huge because of so many differences between the platforms

I would not expect anything else. It's a direct consequence of the business decision you and Dave took to expand and release 1Password on so many platforms, and why your workforce has ballooned in recent years to keep up with the workload.

But the location and appearance of a window close button isn't something you would cover in your apps documentation - that's an OS-level feature which users of each OS would already be familiar with.

However, by actually moving and changing the appearance of a standard Mac window close button, you are forcing Mac users to jarringly leave subconscious behaviour and consciously make the effort to do a common action differently.

I was sad to see that many of our docs have removed the screenshots simply because it was so difficult to maintain them

Do you not have an automated screenshot creation process?

Here is a list of apps in my Mac that either do not have a floating window or do not have the traffic light controls or both

I too can find exceptions on my Mac. But this is about 1Password, an app that is much more important and frequently used than all of the exceptions on your list. It is one of the few apps that runs constantly on my Mac and I interact with it hundreds of time a day. Previous versions of 1Password didn't have this issue either.

It always surprises me that you often resort to the "school playground" style response of "Well, so-and-so does it too!", when these kind of topics come up, rather than trying to set an example of exemplary cross-platform support, whilst still honouring OS-specific UI/UX conventions.