[derefr]

XP was a rewrite of the shell window decorations + controls system, to support drawing these as collections of bitmaps with fonts on top. Likely they were inspired to do this by the fact that macOS was doing things with buttons, modal backgrounds (pinstripe!), etc. that Windows couldn’t do at the time with its hardcoded GDI-draw-call based theming.

Microsoft knew that they wanted the flexibility to be able to do what macOS was doing with theming. But they didn’t have a clear vision of what they were going to use it for. They likely built the theming engine first, and then handed it to their designers to actually come up with a visual identity.

But the new theme engine needed to be tested while it was being built. And what better test, than to see if it can replicate all the things macOS can do that Windows, until then, could not? Probably they considered the theming engine “done” when it reached parity with macOS’s capabilities.

[specialist]

Microsoft has never understood that appearance (UI) is the last step. They've never been able to understand and therefore seriously address the stuff behind the veneer.

Anecdata:

The 2008 presentation by the guy in charge of the Office Ribbon best demonstrates their organizational blindness. TLDR: They only set out to better organize the complexity, and make it more pretty. They never considered making the underlying products more simple.

One of my classmates worked on Word for ages. One of his subprojects was WordArt. Knowing of my interest in graphics, UI, CADD, he consulted with me. We talked about direct manipulation, affordances, UI flow, etc. Then he went dark. Later, he gave me a preview demo. He was so proud of WordArt. But it was terrible. I was actually angry. To my younger self's credit, I somehow kept my mouth shut and made complimentary grunts. My friend later went on to create EndNote, which wasn't terrible.

One of my early bosses later went on oversee the login and authentication portions of Windows. Whatever that licensing registration step is called. For years. Having been trained as a real world architect (eg buildings, not software), he understood and accepted that he was just putting lipstick on a pig. But the pay was a lot better and he had family, mortgage for a nice house, shorter commute with a newer car. Real world architects don't expect to actually do the work they love and trained for.

[toast0]

> Having been trained as a real world architect (eg buildings, not software), he understood and accepted that he was jut putting lipstick on a pig. But the pay was a lot better and he had family, mortgage for a nice house, shorter commute with a newer car. Real world architects don't expect to actually do the work they love and trained for.

(Building) Architecture as a career sucks. There's too many people going into it, and not enough jobs. Getting licensed is currently very onerous. Wages are low unless you're a partner or owner of a firm. When you're a partner or owner, you don't do very much in the way of drawing buildings (which is what people think architecture is like). Architecture is actually mostly managing lots of groups, the client (sometimes also their clients), engineering contractors, building contractors, permitting offices and inspectors, etc.

[derefr]

> They never considered making the underlying products more simple.

I mean, how could they, really? Their primary-revenue-source customers are enterprises with established workflows, that want those established workflows to continue to work. These enterprises would be very unhappy if Microsoft changed the "architecture" of how any of their products do anything, or even how automation products (e.g. VBA macros) would have to interact with their products to accomplish things. The only reason these companies stick with Microsoft, rather than switching to a competitor like Google, is that Microsoft ensures their 20-year-old VBA-laden Excel workbooks continue to work on new versions of Excel.

Any change Microsoft makes to their UI, has to require at-most-trivial changes to these enterprise workflows. And so Microsoft can only really make at-most-trivial changes. They can't cut any Gordian knots by e.g. merging two features into one that solves both problems, because then the workflows couldn't work in terms of the original two features any more.

I feel like the only way Microsoft could really get out of this hole, is by just creating an entirely-new suite of products, that "replace" their existing products for new users, but aren't meant to replace their existing products for existing users, and so where the existing products wouldn't get deprecated/put on life support, until they're something only stodgy old dinosaur companies are using any more.

If they're willing to do it (a dozen times so far) for their collaboration/groupware suite; then why not do it for Office?