The reason why Windows is full of UI inconsistencies is (largely) because of one thing: backwards compatibility.

While backwards compatibility has it's perks, it also means having lots of old code in Windows that can't be touched very easily without breaking compatibility, And that means many old UI elements may stay, that's a side effect.

While desktop Linux/BSD isn't exactly well known for it's UIs, I found GNOME 3/40 to be more consistent UI-wise than Windows 10, being a user of GNOME 3 on FreeBSD at home (which I am typing this comment from), and helping port GNOME 40, while (unsurprisingly) using Windows 10 at $DAYJOB.

But then GNOME, KDE, Apple, et al. is more willing to break things than Windows is. They aren't as bound to legacy compatibility, so old UI elements can be shed easily.

This is the answer. Now it could be done to maintain backwards compat and have everything completely up to date, but the juice is not worth the squeeze and there are more important things to focus on (such as making the OS work/build in such a way that there won’t be inconsistencies in the future).

I don't get it. Functional UI changes should be different from visual fluff, yet Windows seems to the only OS where the two aren't separated.

Of course there are limits. iirc Windows 7 RTM had an obviously 3.1 derived file picker make it into the add fonts dialog until it was replaced in SP1 if my memory has the timing right. Despite using win32 widgets which had the correct theming, that didn't look native at all

The newer stuff sits on top of DirectX.

The latest UI framework (XAML/WinUI) sits on top of a better rendering architecture (compositor) that itself wraps DirectX tech.

(TL;DR: the Windows UI mess looks to me like an organizational problem, not a technological problem. From the outside it looks like as the old guard is leaving, the new people coming in discard all the old code and instead prefer to start completely new projects, which of course is destined to fail, because the new people are repeating the same errors which had been identified and fixed decades ago).

So much code is required to do things we now take for granted that the resulting patterns became new frameworks.

Win32 is old and unviable because Microsoft hasn't bothered making comprehensive upgrades for it and never laid out a gap-free plan for how to move off it. They could have upgraded it more smoothly but never did.

WinUI/UWP still uses CreateWindowEx and good old WndProc for event loop message handling at the top level on Windows 10.

It gives you all the building blocks to implement it yourself.

Remember that Win32 is not UI framework but OS API which allow you to build yourself one (wxWidgets, QT, SWT).

You state that as fact, but I believe the poor UI situation is holding windows back from gaining market share. It’s not just the way things look, but uniformly communicate a single experience. Microsoft used to be very integrated and had an entire ecosystem that was uniform. Letting this go has really hurt the brand IMHO, and creates a very low bar for third parties.

It's not so much that microsoft let things go as it was taken from them. When computers became networked we needed ways of easily distributing applications across that network in secure, reliable and always up to date ways. Windows never had a good solution to that (not even today), because every install was fraught with peril, and every app had to roll its own update mechanism. Meanwhile the web was sandboxed and up to date by definition, dramatically lowering the barrier to entry for networked software. Anything that could move to the web did move to the web, which opened the door for chromebooks, which in turn fragmented the OS market, which made it uneconomical for companies to invest in a windows native codebase. Microsoft could have stopped this had they leaned hard into the concept of an app store and sandboxing around 2000, and hadn't stumbled so badly when it came to extending windows to mobile.

I'm pretty sure that modernizing the entire windows UI doesn't change a thing for windows market share. But the reason to do it is not because it makes economic sense, it is because they should have pride in the products they make, and want them to be well-made.

I'm having a hard time believing that. While some Mac users who are used to a more integrated user experience might be deterred, I'm not sure that the os's ui esthetics really affect windows adoption. My feeling is that most people who use Windows do so because they're either (a) told to use it or (b) they choose to use it because it's the shortest path to doing what they want to do.

I'm sure Microsoft know this and understand the tradeoffs very well

I tried to go all Linux many times, gave up, I just wanted something that works.

Microsoft understands keeping their customers. That is why backwards compatibility is king with Ms.

Windows is not growing in adoption. It is under attack by Google (Android killed Windows on phones and Chromebooks are eating away the education market), Apple, and on the server, Linux.

For the cost of the average computer running MacOS, one could buy a TV, a console, a desktop Windows computer, and a laptop. Or whatever it may be. ($1500 on an iMac could be $400 console, $500 TV, $500 computer. Or whatever your break down is. For the lower middle class and working class, trying to 'gear up' their family and make their kids happy and be happy with their home, this is real math they have to do).

While it is true that for those who are wealthy or who can justify the cost, the experience is very nice and justifiable, the cold hard facts of living in the western world is that most people would need to spend irresponsibly to use MacOS regularly.

It is what it is but it's no surprise that the cheapest mass market OS has 90% market share, and anyone who thinks that "UX" is the reason instead of cost doesn't understand the industry fundamentally.

A $900 MacBook Air would probably have lasted far longer and needed less time troubleshooting than a $500 computer made with subpar components. The amortized costs are far lower with Apple products (or other products using higher quality components) compared to buying the cheapest devices made with the lowest quality components, so there is no point in comparing the two.

And most people only need 1 computer, if that. I would actually bet most people can do everything they need to with iOS and an iPad would suffice. And other than the smaller screen size, I can see people make do with an iPhone even, especially if all you are doing is consuming.

I've never seen a more-true statement on this site. Years ago, my wife was using a $500 Acer laptop, which was a cheap plastic POS, which broke after a couple of years. I bit the bullet, and bought a pre-owned Air for $900. She is still using it, 11 years later. I used to have to fix something on her Windows computer every other week. Now I hardly ever have to help her with anything.

A year later, I bought another $900, used Air for my daughter. She is still using that, 10 years later. I bought a new MBP, and gave my son my 7-year-old version. I expect him to get many more years out of it.

People who have never owned a MacBook expect that they "wear out" like a $500 laptop, but they just don't. They cost much less over the long run. It's like hiking boots. You can buy cheap ones every year, or bite the bullet, and pay 3 times more, and have a pair that lasts for 10 years.

Even fully up to date there is always another zero day, so for individuals doing nothing of particular importance the best bet is good up to date backups of the important files and when your software tools are working well to help you get whatever it is you do done, don’t change them, and don’t update them, because odds are there will be some regression, some feature removed, a change to a subscription model, some functionality now depends on the Internet, or some level of telemetry/spying.

Software has become hostile to users. Everything is nearly malware now. Easily snap shotted and sandboxed VMs are just about the only way to maintain some level of consistency over a 5-10 year span where you are depending on the software to just keep doing what it is doing.

If it's the Mid-2009 model (the next older model), it can still run El Capitan and therefore the newest Chrome.

I have good experience with Asus laptops, though, I still have an EeePC from 2008, and it's still working fine. Slow as a snail, of course, but still working. ThinkPads, too, although they can be a bit pricey.

There's the software side, of course, macOS is much more pleasant to use than Windows. But I found after using a Mac Mini for about three years that I actually prefer GNU/Linux on my desktop.

I'm reminding some tone-deaf wealthy users that COST is the #1 factor behind OS usage, and the rest is basically irrelevant window dressing except to the upper class.

Is what it is, but you'll struggle to find a more privledged and clueless group of commenters than here. Part of the charm, I suppose.

This actually helps makes new devices more affordable as the cost is amortized over a longer and more useful lifespan, while the secondary market serves others very well.

Interesting. I figured the people here are completely ignoring maybe half of all users, but you seem to be ignoring about 85%!

So given your example which already excludes most people on the planet, I was replying to the fact that Apple hardware usually has a lower total-cost-of-ownership given the longevity and resale value. Of course not everyone sees it this way, nor can they afford the upfront price, but it is what it is. Anyway, the company certainly didn't get to a $2T market cap from a lack of demand.

In Windows’s case the “worse” competitor is ChromeOS and Chromebook, and Microsoft’s basically thrown in the towel when it comes to native apps, so with time everyone who doesn’t need pro Photoshop or AutoCAD could just use a Chromebook.

Note that you can run Word for Android on Chromebooks full-screen already.

I haven’t used android games on chrome os but things like candy crush should work just fine in chrome os by now?

Assuming you have a really low latency to a Google datacentre, the Stadia experience is not particularly compelling enough to ditch running the game on bare metal.

All of the above is moot though if you are not within a few ms latency of a Google datacentre though... which for most of the world is absolutely the case.

Instead they were focusing on Xbox and let their PC ecosystem rot. Now Nvidia, Valve, Epic and other middleware vendors consumed this market. They tried with GFWL, but it was a dumpster fire.

They now try again with Game Pass for PC, but again they just want to push Xbox ecosystem instead of Windows gaming as whole.

And one day Google will release desktop Fuchsia.

I use Linux because it’s a way to run a good Unix-like OS on non-Apple hardware.

The overwhelming majority of people who use Windows (or macOS) is because their computer came with it. As long as it mostly works, people don’t care about which OS they are using.

I like Windows because it's more compatible than other OSes with any hardware that I can throw at it. It's also more stable than any other OS. My Macs have been broken multiple times throughout their life by simply upgrading to the next version of macOS, but Windows has always just kept on trucking along year after year. Another reason I love Windows is that it's got a high degree of flexibility. While Apple hides APIs and does everything they can to prevent you from doing things that they don't approve of, Microsoft does the exact opposite and publishes APIs to help you get it done. Mostly though, I just love the Windows UI and the Windows way of doing things. I find it to be very simple and obvious. The things I don't like are largely unseen, like the forced telemetry and updates.

I use my Mac when I need to build or test something for iOS or macOS. Outside of that, I just can't stand the way they do anything. I find the UI and the UX of macOS to be absolutely hideous. The fact that you have to install third party stuff just to get basic features like "decent window management" is simply beyond the pale to me.

For work, I started using a Linux desktop a few years ago because Windows was annoyingly slow when doing anything with npm/yarn/node_modules or Docker and I was doing more and more work with NodeJS and less with .NET/SQL Server/etc. So, I installed Manjaro with XFCE and that's when I found out that XFCE does a minimalist Windows style desktop better than Microsoft does. They have all the Windows features that I want, like Window snapping and none of the Windows features that I don't want. They even have a feature that I usually installed 7+ taskbar tweaker to get: They let you middle-click taskbar items to close the window (also you can drag/drop them), similar to how Chrome treats browser tabs. I've had Manjaro installed on multiple systems (desktops and laptops) for years now and I haven't seen any of the issues that people complain about here, like having to fiddle with their system to get it to work. Everything just worked for me since day one.

Now I just use Windows for entertainment - I've got a NUC style PC at every TV and I have one gaming PC.

That Microsoft created new APIs, runtimes, and frameworks not for technical reasons but in order to lock customers in or disadvantage competitors. That's a big reason why Windows is an unmanageable mess of inconsistent interfaces and DLLs--they were short-term tactics, not well considered approaches.

No, you.

> they didn't help businesses, they locked them into inferior technology that cost more and undermined productivity.

In order to lock people in to something you have to have something of value to offer... so, your argument defeats itself.

> That's a big reason why Windows is an unmanageable mess of inconsistent interfaces and DLLs...blah blah blah

Pfffft. Yeah it's so unmanageable that every single IT department on earth is running Windows all day every day. Even my personal PCs are running perfectly without so much as a hiccup. What a mess!

That’s one case of “it came with the computer”.

But Microsoft was really clever at giving all the functionality companies want for their corporate issue computers. In the mid-2000’s very few companies offered Macs as desktops because of that. I only have a work MacBook because it can authenticate against AD and has all the management automation required by my employer. That’s also why a Linux laptop wasn’t an option.

No, that's the reason why it comes with the computer in the first place. You don't think that PC manufacturers just decided for everybody do you?

Microsoft won for a one reason - they have products that solve problems better than anything else.

UI consistency just isn't that important to most people.

See: electron applications.

Big companies just love to jack around with Windows, and do things like prevent users from changing their desktop backgrounds (in the name of "security," of course), and Microsoft has always bent over backwards to give them the ability to do any stupid thing some corporate IT drone can think of, so they continue to be loved my companies around the globe.

There's a _fundamental_ difference in the way that Microsoft and Apple have approached making an operating system. One treats it like a layer of software in an entire stack that someone else owns and controls, and the other makes PERSONAL devices. And it shows.

And, of course, the Microsoft-slobbering trade press can't imagine a scenario where they don't try to prove that YOU should be running Windows because the numbers -- inflated by the Fortune 1000 -- says it's what "everyone" is running. Meanwhile, more than half the people I know now run NOT-Windows for their personal computing.

Let's break out the corporate purchases from sales data, and then let's talk about actual, personal market share, and whether better UI/UX would help Microsoft in a heads-up battle against macOS or ChromeOS.

In my experience, Windows is still the target for most specialized or business software. Most of these things also demand a Windows Server component. Typical suspects are: "Personalmanagement" timesheet software basically (accounting working hours, vacation, bonuses etc. and making sure the right receipts are printed at the end of each month and more, like when somebody leaves the company, is ill or the is "Kurzarbeit" because the union with the employer settled on this). Another suspect is software for insurance and taxes, energy management (which in a stainless steel foundry is an important piece of software and keeps at least a handful of people fully employed), logistics software (usually also with needle printers that are able to still print "Durchschlag" carbon paper https://en.wikipedia.org/wiki/Carbon_paper which is fast and handy for the drivers). There is special software for measurement and production with complex machines, ovens, mills, spectrometers, microscopes, x-ray machines etc. Oh and then there is the internally developed application, that keeps running the company and will be replaced sometime during the next 10 years with a different Windows native app if it is ever finished. And there is the MS SQL administration in addition to group policy for all the Windows computers everywhere that also is best done from a Windows computer.

People are used to Windows, older generations in middle and slightly eastern european countries (former east Germany/ DDR included) especially. There just were no Macs in any company besides maybe a top design studio in the capital or regional hub. This is a big factor. In most companies, ordering certain behaviour just doesn't work in practise.

In companies, every purchase is oriented towards "Abschreibung" or the amortisation period that is determined by "Abgabenordnung" if I am not mistaken - basically the tax. To be most efficient on paper, you have to use the device at least for the duration of the period determined by the state tax books, else you will have a bad tax audit etc. Anything that hasn't followed the rule will raise an eyebrow and somebody will have to explain otherwise you will likely pay more in taxes or whatever. I don't understand this stuff at all but I was told it works something like that and therefore you do so much accounting stuff such as inventory of basically everything that costs at least X, every piece of software, where you also get Microsoft and Oracle and other audits from time to time...

If you are not in the business of selling/ writing enterprise or specialized software in some capacity, you don't really have to care. Most things work using a web browser nowadays, even some enterprise software some companies use. Some software is just more performant or can access certain APIs that aren't available in the web browser. Also, it is very hard to support anything for 10+ years in the webbrowser. Stuff just breaks too frequently on the web, even compared to Windows 10.

In the end, people usually buy Windows, because it really is the only platform that actually works for 99,9% of things. You can spend a long time to support the CEO's Mac, because it is the CEO. Of course, most things running a web app or supporting infrastructure can be Linux or UNIX-based without an issue, but the OS on people's desks just is Windows and it will stay like that for a long time. There are few companies that have the resources to support something else on the desktop as well. I bet, these are in the software or hardware development or creative/ design/ art business and not much else.

Education is a very special market, really anything goes there. Simple things tend to win, but what simple is changes school to school. :-) There are schools that it seems don't know what to do with their money and have dedicated IT department, supporting anything Apple, Google, Adobe, AutoDesk and Microsoft is not a problem there. There are schools that cannot even equip one IT classroom and students have to share computers each with a different OS and updates barely work because the internet connection is flaky.

The bar for third parties was always nonexistent. That was precisely the advantage of Windows which permitted the boom of development for it in the 90s–2000s.

    I believe the poor UI situation is holding windows back from gaining market share.

I just don't buy that. Everyone I've ever persuaded to buy a Mac or iOS device comes to prefer it to Windows in a couple of weeks, and I stop needing to fix something for them every other month.

Backwards compatibility is not what stops Microsoft from updating the GUIs for their own components. It is not the reason that the components that have been updated failed to carry forward 100% of the features of the components they're replacing, necessitating the ongoing use of the legacy components. It is not the reason reams of GUIs have gone untouched for 20 years. It is not the reason Windows now has at least 4 versions(!) of Performance Monitor, all of which are broken in at least one glaring way. (More on that below.)

The believable reasons for the GUI inconsistencies I've heard are:

1) API churn, because of which the UI teams did not have sufficient time to work on features versus playing catchup.

2) Bad system GUI APIs that are very difficult for even in-house teams to work with.

3) Unwillingness to take ownership of legacy code, with literally noone left at Microsoft willing to touch things like the ODBC connections panel or the Component Services snapin. People prefer to add new things instead of fixing or removing old things because it's "easier".

Actually, let's just stop here for a second. That last point explains the performance monitor views. I mentioned at least 4 copies, written over decades, by different people. Each new team has steadfastly refused to touch (or remove!) the old code, but hasn't replaced the functionality of the old code, so now end-users need 4 different versions to get things done.

These versions are:

1) The performance tab in Task Manager. The only GUI-based view that exposes some metrics such as GPU temperatures. Only shows a small number of fixed metrics.

2) Resource Monitor, which is opened from Task Manager. The only GUI-based view that shows certain per-process metrics, such as the names of files being touched, or per-connection network stats. Has permanent UI issues that will never be fixed, such as not using the system number formatting in some places, making large metrics unreadable as they change faster than users can count the digits. Similarly, the graphs take forever to change their vertical axes, making them useless 90% of the time.

3) The Performance Monitor MMC snapin. Totally legacy, with un-resizable controls that cut off text. Nonetheless, it is absolutely essential because it provides the only live GUI view of 100% of the performance metrics available in the system. It is also the only way to record metrics and view recordings. It is the only GUI for creating metric logging that persists. Etc...

4) The various versions of Server Manager's performance views, which are so useless that I've literally never used it. Nobody can get their job done doing this, we're all still using RDP to connect to servers so that we can simultaneously launch the three tools above. Why RDP? Because 2 of the 3 above do not support remoting.

I could go on and on like this for hours about how bad just this one aspect of Windows is, let alone the hundreds of other GUIs that have been butchered by bad decision making and internal NIH syndrome leaving a trail of half-baked messes behind.

Wait... did I say 4 performance metrics GUIs? I meant 5, because there's also the new Windows Admin Center, which was clearly written by people that have never had to diagnose a performance issue on a server. It's very pretty and utterly useless, which means that: Nobody will use it, and it will be superseded by someone else's half-baked attempt in a few years, 100% guaranteed.

  * Microsoft Mail
  * Windows Messaging
  * Outlook Express
  * Microsoft Outlook
  * Windows Mail
  * Windows Live Mail
  * Mail.app

Windows Live Mail and Windows Mail were essentially newer, Vista-era versions of Outlook Express. It's the same program rebranded. Mail was bundled with Vista, Live Mail was a way to push the Live branding and get the newer program into the hands of those running older versions of the OS (WinXP).

Microsoft Outlook is a different beast with a different lineage, part of MS Office, and can be thought of the "pro" version of the basic mail clients bundled with Windows.

Yes, Microsoft could probably do with some branding discipline but the technology is quite predictable.

I'm not sure what Mail.app is, I always thought that was Apple's. Entourage was IIRC a native 'designed for Apple' PIM that eventually made its way into Office for Mac, eventually they decided to rationalize and have one client -- Outlook.

Hotmail / Outlook.com - Outlook.com, again despite the confused branding, is Hotmail evolved. Over time, Microsoft slowly merged its consumer (Hotmail, Windows Live) UI with its 'pro' UI (Outlook for desktops, Exchange Web) so that these days they are all pretty similar. These days you also have Outlook's mobile apps for Android and iOS and those too have a familiar look and feel.

I don't see any of these as a bad thing. Microsoft's history is linked to the history of personal computing and the ebbs and flows of market forces that shaped the PC biz and continues to shape today's technology. Given all the churn in this space, it's actually kind of awesome how their technology has evolved -- and will continue evolving, e.g. with all the focus on a web version of Office. But yes, their sudden shifts in branding is pretty sucky and they could do a lot better there.

But "none of them 100% work" is pretty harsh. No mail client is right for every use case (except emacs, naturally -- if it doesn't work for you it's because you haven't written enough elisp yet). But these are very widely deployed products and do work for a good segment of their target market.

Mail.app is what I call the new "modern" app shipped with Windows 10. Mostly garbage. Spends most of its time and your CPU to resynchronize your mails.

My experience with these programs is that the predecessor can have feature X but the successor doesn’t but has feature Y, similar what OP described.

I’m always harsh when reviewing software. No need to be gentle about things that doesn’t work well.

I think also I was annoyed over how one email list item was drawn, too big, thus reduces the number I can see at one time. Can’t remember if I found anyway to customize that.

I ended my Office subscription because I didn’t really like the feeling of the applications. It feels like they have put lipstick on a pig. Yes, there is a minimal writing mode in Word, but beneath it is the same old steaming pile that seeps thru.

But I’m somewhat harsh when evaluating software so maybe I’m missing something here.

I do however find it to be the best (on Windows; unsure about on Mac).

I would so much like Microsoft to sell it outside of "Office Pro" as I would like to use it at home too but do not otherwise need "Office Pro" at home.

I am not familiar with Windows APIs but I can see this be the truth if MS made it possible for developers to plug inside this dialogs, like a printer driver would plug soem GUI changes in the panels , changing this would break a lot of devices.

Does logman not working anymore? If not, hasn't it been replaced by PS APIs?

So even there, the unnecessary and counter-productive interface churn continues to the detriment of the users.

I'm always very skeptical of refusal to handle the past because of the promise that the same won't reiterate in the future. What could also happen is that 5 years from now, you or another dev justifies refusal to handle inconsistencies developed "temporarily" and or accidentally in the meantime with exactly the same reasoning.

A more direct and simple observation is that if nobody is working on converging / cleaning up stuff, they don't converge and aren't cleaned up.

As for the amount of backward compat already provided by Windows, it is good but not great anyway. Try to play old games, install/use old VS with a few patches, install/old Matlab, use old devices, etc... Hell even try to continue to use your semi-recent devices across new build of windows, like 20H2 breaking USB-C on my XPS 15 9560 (but now I'm ranting so I'll stop :D )

When I use multiple Explorer Views, most of the time I do it, to interact between them. This is not the case for websites.

So for drag and drop interactions, two windows always beat the drag, hover over the other tab until it catches my intend to switch to this tab, and drop it there workflow.

The same goes for copy and paste. Ctrl+c, alt+tab, ctrl+v. I'm sure there is a corresponding command to cycle through the tabs, but alt+tab is deep muscle memory, that it beats this command always.

I guess it's dependent on the number of applications and windows you have going at any given time.

I have been in love with - and 10x more productive using - TotalCommander!

The ability to >98% of the file related tasks from the keyboard is very addictive!!!

So if you wanted an "unlocked" windows RT, couldn't you just have run windows 8?

Win32 was not removed from Windows RT, but your access to it (if your name wasn't "Microsoft"; because that low-level access was reportedly how they got Office on it) was.

And because of Secure Boot, which (unlike x86, for now) is mandated enabled for Windows on ARM systems, you had to wait for an unpatchable security hole in the booting process (which, while it exists, wasn't relevant until past the end of the Surface RT's lifecycle), you couldn't run full Windows 8 or a tablet-first Linux distro.

Windows RT was just Windows 8 with a locked code signing policy, yup. And yes, Office on Windows RT was Win32, as were File Explorer, MMC, ... it's just recompiled Win8 after all.

Also there is nothing like "bloat". No process needs too loop around or smth like that, so there is literally no way to have a forgotten registry item slowing down your PC.

And last but not least: What are the alternatives? .plist config files like with Apple, scattering around the system?

The critique against the Windows registry stems from the Windows 95 days when the registry could become corrupt and repair or reinstall was necessary. However since then the registry works like a database, it has transaction logs and simultaneous writes are atomic. Registry is also strongly typed, not a bunch of strings. Registry can easily be backed up and restored.

Do you have any evidence of this? I remember in the XP days the German c't computer magazine tested populating the registry with hundreds MBs of garbage and found no performance impact at all.

I do not really think one is better than the other.

I don't see any alternative on where entire OS could be configured other than registry.

Yes, they do break backwards compatibility in many many other ways, but you could put a technical layperson who is used to Mac OS 9 and put them on macOS 11.4 (for those unfamiliar, Apple's OS has been at 10.x for about 20 years) and they would still reasonably be able to find their way around.

One element especially (the global menu bar) is why I got my parents to switch to macOS. They had to learn where 'print' or 'edit' or 'save' was once, and now they can find it for any program, and it'll probably still be in that place in 20 years time.

Its also why I think it's sad that Gnome abandoned the global menu bar. It's an amazing concept that if stuck with makes computers so much easier to use for a large contingent of people.

But generally speaking for them it was still a hassle and hurt to use a computer. iPad with Home Button was so much easier. No support calls for years. It was the computing devices that works. But now they moved on to Smartphone meant iPad get way less usage.

Which also means some of those design decisions that make macOS easier to use might not be relevant as PC transfer to a more professional type usage rather than consumer / everyone computing devices. But even in that case I still think Global Menu Bar is a better design.

In the mid 90s, I could walk my 80-something-year-old grandfather through WordPerfect with ease: "File does basic file operations, like save or print. Edit is for editing, like copying and pasting."

Menu bars work because they're a great feature-discovery tool when done right--with verbs that describe the commands, and with functions grouped by category.

"Modern" UIs don't. Try to explain it: "So, the lines are called a hamburger menu, they're supposed to look like a list of commands. And you can press it, even though it's completely flat. Then you need read the list until you find the command you want."

The Mac GUI's great innovation was combining intuitive visual cues with an consistent, unintrusive, and always-on system for feature discovery (the menu bar).

My perfect OS would have a global menu bar and enforce that all commands can be issued by searching and all ribbons/icon bars can be turned off.

Better than KDE imo - you keep sane defaults, simpler configs, & consistent themes.

They should stay on disk to be used by programs compiled 20 years ago.

They should not be used by OS components, nor visible to end user unless they run these old applications.

The OS already has multiple good ways to achieve that.

1. New programs can opt in to new stuff with a manifest. That’s how you get high DPI support in WinAPI: https://docs.microsoft.com/en-us/windows/win32/hidpi/setting...

2. New programs can call some API functions to enable new stuff. That’s how you get modern versions of Win32 common controls: https://docs.microsoft.com/en-us/windows/win32/api/commctrl/...

3. The OS has a lightweight virtualization layer for file system and registry, that’s how 32-bit apps view things differently: https://docs.microsoft.com/en-us/windows/win32/winprog64/fil... Microsoft could have used that thing not just for 32-bit apps, for backward compatibility too, i.e. old apps could use special compatibility view of the environment.

4. Finally, the kernel supports multiple user-mode environments, Win32 is just one of them. The initial good version of WSL (I still use it every day, too bad they have killed the awesome product moving to virtual machines in WSL 2) runs Linux binaries directly on top of Windows 10 kernel, the subsystem is implemented in lxss.sys and lxcore.sys. Microsoft could have embedded couple popular Windows versions, such as 7 and XP, as different subsystems used for running old apps.

Now, while I can understand the sibling's point about third-party overlays [0] which may break when the windows suddenly become resizable, what I find absolutely infuriating is that even what seem like recent features still inherit this stupid behavior.

The thing that I'm reminded of is "Windows Defender Exceptions" (or similar, don't have a Windows box at hand). Those are configured through a new, Settings App-looking window.

Now I understand this is somehow different, because even though the window vaguely looks like the Settings App (huge icons, tons and tons of whitespace), it behaves very differently.

This "exceptions" panel shows a list of the folders that is fixed, although the outer window can be resized. So if you've got a big high-res screen and maximize it, you'll get 90% of blank space and will still have to scroll around to see the contents. Bonus points for the list being so skinny that I'm only able to see "C:\Users\vlad\..." and have to click on each entry to see it in full.

I sometimes jokingly say that it shows MS has a mouse business, and having to click a thousand times for the simplest of things is by design.

---

[0] Not sure what those are. Do random third-parties "augment" system dialogs by drawing random stuff on top of them? I find that thought horrendous.

The common example would be Dell/Synaptics adding their own tabs to the old Mouse system dialog.

I think the idea was quite sensible. Keeping the configuration in one place while supporting whatever features the manufacturers might have added e.g. touch pad gestures.

The only reason it's unfortunate is that the new dialogs don't support the old extension mechanism, forcing the old dialogs to be kept around.

I agree, but I thought that there was some sort of extension functionality, along the lines of "this is a mouse configuration panel, register it to [something] so that it's added to the mouse settings".

Detecting whether that is being done is challenging. Typical workaround is that, if any callback is registered with a dialoog, to keep things ‘as is’.

Growing up in the 90s, I remember an article in Visual Basic Programmer's Journal teaching subclassing in VB by hooking into Notepad and adding a font picker (via a new menu entry that opens the font dialog) to change the text control's typeface. (This was long before MS added that same feature)

So one way to extend UI components, since its inception, is to request the Windowsclass structure of an existing control, replace its event handler function and then call the "base class" handler for the built in behavior.

That typically is for a single program, though, and by that program. System-wide changes, for example for a product that can compress files on the fly, affect every program using open dialogs.

A node you create could be arbitrarily changed or removed.

...completely unjustifiable?

Come on, why does this mean we need 5 different sound dialogues to do 5 different things? That's ridiculous on every level. It used to all be in one place anyway, so it's not even like it needs to be spread out in the first place. Make the old dialogues inaccessible, hide them, and create a new one which does all of those jobs. Then get rid of the old, now unused ones at the first opportunity.

Is the explanation really backwards compatibility, or is it lack of organisation?

Every windows update makes the strips a function from the control panel and ads something not quite as good to the settings app.

After the last update the settings app even has ads for 'bing rewards points' in it. Which appear to be impossible to remove.

I need to look harder at getting one of the commercial versions of windows that only gets security updates, but no feature updates, because I've come to hate and fear upgrades.

I get it, all practical evidence seems to be that they're completely incompetent and this is impossible and any attempts to clean it up will inexplicably make it worse, but the fact that's the state of affairs at a two trillion dollar company is absolutely ridiculous.

I've been a Windows user since Windows 3.11. I suggest another explanation: Microsoft has been a monopoly in desktop OSs for 40 years, and it shows.

So they make control panel 2.0, reach 70% coverage, launch, and never move the remaining features into the new version so they can sunset the old one.

Completely unrelated and a shot in the dark, but if you’re both a Microsoftie and a FreeBSD user/committer, the FreeBSD .NET port is stumbling/regressing somewhat and could definitely use a little help:

https://github.com/dotnet/runtime/issues/14537

(Better start at the end of that thread, rather than the start)

Anything you could do to help (if only pushing the right buttons/people internally), I’m sure would be greatly appreciated.

macOS has actually been pretty good at seamlessly adding new features to old apps, with just a recompile against the latest SDK, like all the NSDocument-derived stuff (auto-save, auto-recover, iCloud etc.), Dark Mode, accessibility etc.

It's bizarre that to do some basic audio or networking configuration I have to wander through 2-3 generations of UI.

And why can Microsoft goes the same route as Apple's Cocoa and Carbon?

All of the Fluent/Metro designs have made things markedly worse. For example, check out the Metro control panel widget for selecting a default program for a file extension. It's a list view with 5 file extensions visible at a time, with no search. That's what I think of when I hear about a Windows UI update, and I'd love to be proven wrong.

It's a "holy shit fire the UX design and product manager leads in charge of this asap". The audio tab is another good example. It's so hard to do anything with the new one, you almost have to know how to get to the old UX to make any kind of change.

For all these people, the new UI is actually pretty good. My parents find the settings they need easily enough. Most business people I work with use Windows and rarely want to configure anything more than the wallpaper. For all these people, the "nerfed" UI is more accessible than a wall of options in the control panel. It's a shame for us power users but most people actually like these changes or at least find them usable enough

Whereas for me, person who has to give phone support, every version of Windows has made important settings harder to get too, which means I can't muscle memory train casual users about what setting they need to review if the problem recurs.

The new UIs are all attempted flash over substance, and idiom breaking - buttons and interactive controls are not clearly highlighted and information density is low.

In day-to-day usage, I have rarely been annoyed by these "inconsistencies".

The best example to illustrate this article is: Installing a driver with the Windows XP layer. That's probably something that will only happen to people installing very old hardware with drivers from over 10 years ago. I consider that a backward compatibility win: It is great that it even works in Windows 10.

Don't you ever have to change sound settings? Because I have to deal with that every week, and have to dig through 3 different style of UIs and several different windows/dialogs to get to what I need.

Everything related to control panel / settings is a giant mess, and it's often really hard to find the settings you want, especially if you don't already know where it is. Some things is in the new UI, some in an old UI.

It's actually one of the main reasons why I really don't want to use Windows for my home laptop. Windows does have it strengths, and I don't mind using it at work, but I really don't want any of that bullsh*t when I get home from work.

I was a Windows user until 2014. Every time I think about switching back (mostly because WSL is so good) I have to think about putting up with the terrible UI.

They really shot themselves in the foot 10 years ago when they decided to go all-in on making their desktop and server OSes a giant advertisement for their tablets.

Which nobody bought, of course. Didn't help that it took them 4 years to get up to feature parity with what you could do on iOS and Android out of the box (like support for SQLite) or that the sandboxed nature of WinRT/UWP meant that you had to jump through a bunch of extra hoops if you so much dared to save a file to the local machine.

They just haven't recovered from that, have had zero incentive to improve, and it shows (MS lost the mobile OS war just as hard as IBM lost the desktop OS war); workplaces are going to buy licenses no matter what, and home users only got the ability to pass through a subset of consumer-level nVidia GPUs (which IIRC doesn't apply to laptops, so you're still stuck there) to a Windows guest VM thanks to a recent driver hack, and that's only for technically competent users who are running Linux on hardware that isn't cutting edge (yep, it's still a problem for Linux) and who are willing to accept the associated 5% performance penalty.

No free solutions to force MS to clean up their act means they just won't, no matter how many collective years of human life is wasted clicking through 5000 different boxes now.

That's not quite correct any longer. Nvidia now supports passthrough on all consumer GPUs that work with their current driver. This also includes laptops with Nvidia chips and they've been actively fixing bugs and compatibility problems in every driver release since then.

Before official support the Nvidia kernel driver just crashed Windows with a BSOD on my laptop. Then it suddenly worked, but I wasn't able to open the control panel and worse, the HDMI output didn't work. The driver after that fixed those problems. The only things still not working are voltage and power monitoring, although temps work.

Additionally, and I know this is a special case, performance is actually better than on bare metal. The reason is that in a VM I can inject a modded VBIOS with higher power limits and that gives me a nice 15-20% boost. I can't do that on bare metal because the VBIOS is stored in the system firmware and accessed via an ACPI method by the driver. Sadly only signed firmware can be flashed directly and even a flash programmer wouldn't help because at boot the firmware is verified and overwritten by a clean if it doesn't pass. This doesn't matter for desktop chips, but on a laptop you can actually get way more FPS if the power limits are ridiculously low by default! In my case I went from 30W or so to 75W (PCIe limit) and even though clocks stayed the same, games ran much faster. I guess it was hitting power limits and throttled, but didn't down lock because clocks only depend on voltage and available voltage level on temperature. I also ever saw power limits (via AfterBurner OSD) in furmark, but apparently it was also limited in other workloads without indicating it. With the modded VBIOS I don't even see power limits in furmark. It's either voltage or temps.

Sure, there might not be any more Code 43 (which wasn't the hardest thing to bypass, but you did have to hide the fact that the VM was a VM), but Code 43 was the relative tip of the iceberg compared to all the other setup you had to do to even get there in the first place.

Passing through an nVidia GPU the way that data centers do it, where you can run several hardware-accelerated VMs on the same GPU (as well as the host, on only one set of outputs), to my knowledge still requires this hack (https://github.com/DualCoder/vgpu_unlock); this is specifically what I'm referring to and is really the only workable solution as it's the closest to "press button get seamless performance from your gaming Windows VM" as we can get right now.

I'd be delighted to now be wrong about my assessment of how difficult trying to set that nonsense up was, because then I could do it and stop worrying about how awful the Windows 11 experience is going to be (even GNOME, for all its user-hostility, still isn't stupid enough to ditch the taskbar labels for the full OS X dock; Windows 11 has so far made that shit mandatory).

It's true that setting up everything can take some time to get it working, but it depends very much on the hardware. In most cases on a desktop system all you need to do is select the right PCI device in libvirt's GUI and it should work. On laptops you need to hack together a custom OVMF UEFI firmware imagine with the VBIOS included, but there are ways to auto-generate it [1].

Input can be handled in multiple ways, which makes most sense depends on if you want to use dedicated output, or if you want to stream the output back to the host using something like Looking Glass. I just have a virtual qxl GPU in the VM which lets me use the qemu window for input and that works fine.

For audio there is scream[2], which gives me super-low latency 7.1 audio. According to pulseaudio latency is under 2ms in addition to whatever latency Windows itself has. I use a -ck kernel with MuQSS scheduler, the default scheduler won't be able to go that low.

I mean it can be tricky to set up, but if the hardware is capable it shouldn't be more than a weekend project. Once set up it's rock solid stability wise and compatibility is the same as bare metal. Caveat is some online games consider running in a VM cheating, but that doesn't affect me. It's not for everyone, but neither is Linux as a desktop OS

Overall I consider this the holy grail for those that want to primarily run Linux but don't want to give up games or deal with wine/proton issues. Actual dual booting is completely obsolete with GPU passthrough in my opinion. YMMV, depending on what you play and your exact hardware.

[1]: https://github.com/bitthief/ovmf-with-vbios-patch [2]: https://github.com/duncanthrax/scream

Until reading this article I didn't know Fluent Design, Metro and Windows 8 Win32 were supposed to be separate design languages; I'm still not convinced, and in any case I can only see those with the sharpest of eyes noticing any differences here.

Calling out things like MMC, winver and screensaver settings as relics from Windows 95 also seems like a huge stretch to me. Even though their layout has remained the same (which is a good thing; they do their job well), their design language is quite clearly different from their Windows 95 counterparts.

It makes sense why it’s not much of an issue on other GUI environments: in GNU/Linux and macOS, only what’s available on GUI is available on GUI, and the rest goes to the unified scary black screen with very task specific procedures.

Only in Windows the GUI is primary means of configuration and only Windows carry over old GUI associated with each configuration items.

Alternative would be knowing that a lot of 'old' control panel UIs are accessible through the new one, so: go into the new settings UI to network-related stuff (Win-x, w or Win-X, n and start typing 'network') and there will be 'change adapter settings' somewhere. Or possibly you can change the setting you want from the new UI.

Which indeed shows this particular par was turned into a mess. But as the OP says: I haven't actually been bothered by it a lot myself. Likewise for some UI parts looking different: it has been like that since as long as I can remember on computers, no matter which OS or whether it's a CLI, TUI or GUI: if you use more than a couple of tools there are always going to be tools which look and/or do things in a different way. So people kinda get used to that anyway. Not that it's an excuse for doing this to an OS, but just looking at it from the practical side.

Also -- it looks great! Why isn't that an option somewhere in Control Panels?

The wifi and hotspots are in metro ui setting app.

Interfaces are in its own standalone app. (win7 style)

However... The firewall are in a xp style standalone app.

Can't they combine them into one app or at least make a folder that contains link to them?

I just want a version of Windows that I am in control of. I don't want my machines (or my Customers' machines, for whom I am responsible to support) torpedoing themselves with shoddy updates that try to install at inappropriate times (need to shut down your laptop to put it into a bag and head home-- tough-- we're installing updates!).

I can deal with UI. I've been learning new UI my whole life. The whole "it's not your computer anymore" tack Microsoft took with Windows 10 is condescending and abusive.

So much of this is self-inflicted on the part of Microsoft.

Minor things like Win10 vs. Metro are less obvious, though.

Think however about people who are new to computers, whether adults or kids growing. They find it all confusing.

I actually spent 15 minutes the other day trying to remove an additional keyboard layout that somehow found itself on my machine. I had to give up as I could not find the option in the new Settings screen and it seemed to have disappeared (or moved?) from the usual place in the Control Panel...

To this day, every time I need to uninstall an application I still do: win+R -> control appwiz.cpl -> enter and it brings me to the place I want to be.

Very often I don’t bother opening either of them, instead using start menu search. For instance, to open the old “network and sharing center” GUI I press Win key, type “netw”, and press enter.

And lastly, while it always stays responsive, control of the background processes is also lacking. I spent quite some time restarting the download of a system language, just because I started it while having DNS issues (and to fix those... see above).

I personally don't mind the older designs being mixed in, especially since most of them look good enough with the updated buttons, but the settings panel is, in its current state, still a sidegrade at best.

It will continue to suck until they've actually moved all the settings from the control panel into settings.

But then when I fire up Windows, it is just this bizarre mix of all these different UI frameworks. I guess Windows isn't the big cash cow it once was for Microsoft and maybe it's not the highest development priority anymore, but still it's remarkably off-putting and I have no idea why they would let it stay in this state.

I read they were introducing a new UI framework to unify everything (Project Reunion), but I can't help but think that will mean it's just another different-looking toolkit thrown into the mix.

With all that said, Microsoft's commitment to backwards compatibility is fairly legendary so there could be an actual technical reason some things can't be redesigned with a newer toolkit.

This is probably to encourage developers to build software using the new Windows 10 APIs. Which is something they definitely have a problem with - I've been using Windows 10 quite heavily since the first beta and I don't think I've ever seen or used a third party Windows 10 (fluent / modern / metro) application.

It is also a way to bring into Win32 side most of the API that really matter on the UWP side of the fence, most likely (although they are yet to confirm this) as migration path to UWP developers.

.NET Native, C++/CX are pretty much dead, and C++/WinRT keeps missing many of the tooling available on C++/CX for XAML applications.

So I expect them to migrate the APIs into an OS version independent layer, and then give the mercy shot on UWP apps.

[1] Edit: quarter of a century, not quarter of a decade.

Enterprises expect to run programs developed a quarter century ago... if a quarter decade was all that was required, MS could have abandoned a lot more cruft.

yes, I understand - bring on the downvotes

I think I've seen computers with more uninterrupted uptime than that, even if servers. Or at least approaching this figure.

Perhaps you meant ‘quarter of a millennium’ or something like that.

https://github.com/jfinkels/analyticalengine

https://www.fourmilab.ch/babbage/emulator.html

I had the good fortune to see the Babbage Engine in action when it was at the Computer History Museum. Only got a few photos, but here they are (plus an unrelated photo of a core memory plane):

https://geary.smugmug.com/Computers/History/

https://www.computerhistory.org/babbage/

I'd say it's still not as bad as Windows, in that at least it's not pure white that burns your eyes, but I still hate how half of my laptop screen is taken up by useless UI elements.

[1] https://www.macworld.com/article/234382/macos-big-sur-learn-...

After using BS for a year+ now (since the early betas), the biggest UI/UX issue IMO is around notifications. After making other UI elements larger and 'touch friendly', clearing notifications require hitting a tiny X or a hover over to clear all. I'm not on the Monterey beta yet, but hopefully it is addressed.

While macOS has consistency issues (why can't I resize the preference pane?!), and windowing problems (when plugging in an external monitor sometimes my windows don't go to the right spots), they are nothing compared to the inconsistencies in Windows. I'm sure Windows users would love to mainly complain about the size of a close button vs. wondering why they still see dialogs from the 3.11 days.

Personally I'm thinking of trying BS in a virtual machine to see if I'm ok with it or will stay away until apps that I want drop support for Mojave (likely Brew will begin to compile lots of stuff instead of just installing binaries). Though, seeing as Catalina already makes life harder for developers outside of the App Store, Mojave is probably gonna stay with me for a while.

;) Sadly I think this is the group Microsoft is least likely to attempt to cater towards.

https://www.reddit.com/r/windows/comments/o12no7/still_rocki...

The analog in Unix apps are GTK's file dialogs, which always look jarring and out-of-place anywhere other than Gnome (and maybe XFCE and KDE, dunno)—making me feel like I stepped into 90s Unix with those Motif-derived concrete slabs. Plus they're un-integrated with non-Linux OSes, missing features. And the dialogs keep being dragged around by GTK apps—open Gimp on Mac or Windows and there you go: GTK dialogs.

I'm worried with 11 it'll be a repeat of Windows 10 where at the start you're constantly using "yet another UI". At this point it's nearly gotten decent to the point if you aren't a sysadmin (MMC, regedit, task manager, advanced control panel items not nowadays part of Settings) you rarely run into the "old" style stuff with the one exception of File Explorer. Hopefully it doesn't take another 6 years to be saying that about Fluent/Metro stuff being updated to match 11's new look.

It would be an acceptable tradeoff if these new interfaces added more functionality or visibility but it seems they're just trying to move to the new settings ux and not thinking about adding better function.

Try adding a new local user account to latest windows, compare the functionality in new the network interface settings page vs the legacy adapter properties prompt.

Seems most of the "this just works" stuff that I respect is from windows 2000 and the NT team.

Wish Microsoft would head back that way with a nod to modern usability instead of dumbing things down and worse: changing it every release.

I did some work on Windows servers in the XP era and it was a continuous hassle that the menus were always different in NT 4, Win 2k, Xp, Vista, and all the different versions of Windows server.

Contrast that to command-line admin on Linux that stays the same or the perpetually busted Linux admin GUIs that should just be taken behind the woodshed and shot.

Or is it that the UI elements somehow get baked into the program at the time you build it, so if you didn't update you are just stuck with that particular version of UI you built the program ?

Under Windows, the rule has been that basically anything goes, and that's been true for upwards of two decades. Anybody who used Windows 98/2K/XP can probably name several apps they remember using totally custom UI off the top of their heads. That never really changed. Custom UI has become a lot less wild looking, but it's still custom (which means no "free" updates), and on top of that there's no incentive for developers to update their apps to match the latest look.

Those rules were set by example by microsoft.