I'll go on... the 95/XP UI was meticulously crafted. It solved some real usability problems by drawing on metaphors to aid discovery. It was close to a language. I think what happened was eventually the designer types got their hands in there and started making changes here and there to make specific things look good. Now we're here.
That's exactly what happened, and it all started with the damn ribbon BS in Office.
To me, the "advancement" of the human-computer interface came to a screeching halt at that time. People grok menus. Memorizing sequences of Mnemonic accelerators is natural when overlaid on a haptic path created by mouse movement/menu navigation.
Then you take everything and throw out the baby with the bath water by trying to put every past bit of functionality on screen at the same time when what one needs is to be able to set up a work bench quickly, and just do your work, and if necessary save said work bench state to return to later.
I've seen more and "better" advancements in UI design in exactly 1 place, in all the different software suites I use, and that is in IDEs. Ironically, the one place that programmers have to make due day in, day out.
It makes me wonder whether anyone who presides over these programs is actually proficient in the non-computing version of said task. A Typesetter/document composer for instance. Nothing about the modern interface leads me to believe that any actually invested individual is involved, and that the design was pushed by the need for someone at Microsoft to be seen as needed in order to change something.
Then again... Can"t judge 'em for that. We all want to be wanted.
What I hate is the save / options / preference / file menu of office 2016 upwards. It's a nightmare - just too many steps to save a simple file to disk. Even after using it for years, I still need to pay full attention in this menu rather than muscle memory.
Yep. That thing is a fucking abomination. I believe it's called backstage or something and the idiots who petitioned for, designed and approved it should be dishonorably discharged.
I agree. With the the ribbon they broke their own UI guidelines and never got back. Windows 95 was a huge jump in usability but first with the ribbon and then with Windows 8 they destroyed their own progress.
Even after years I prefer menus and toolbars over the ribbon. With them you can scan in one direction whereas with the ribbon you have to up and down, left and right. I find it really difficult to find something there.
> I think what happened was eventually the designer types got their hands in there
No, what happened is Microsoft placed a bet on us moving to a multi-platform future where there are many fewer PCs and more tablets, hybrid devices, whatever, all with the need to use touch, pen as well as keyboard for input. The existing start menu and taskbar would be hideous on mobile devices.
The problem is they were wrong. But I'm not going to blame "designer types". If you ever used Metro UI on a Windows Phone device you'd see that it was a truly remarkable system... that was bent to fit a task it didn't fit.
I feel their main mistake was lack of choice. The second biggest mistake was pushing too many changes at a time.
For choice, they finally relented with windows 8.1 and allowed people to boot to desktop directly rather than the start screen. Windows 10 took off the start screen completely.
Choice matters. I liked the start screen but the horizontal scrolling killed it. Like pouring salt in Tea.
Too many changes.
Facebook's learned this lesson well. They no longer do large overhauls. Small changes at a time and in a few months, it's a whole different interface. Only infrequent users would notice.
That's what they are doing with Windows 10 - many small interface changes. The little changes are mainly minor inconveniences.
One last thing.
When they released windows 8 - hardware support was almost non existent. Touch screen PC from that era really sucked. Now with detachable screens, Win 8 style pc interactions might be cool.
> a multi-platform future where there are many fewer PCs and more tablets, hybrid devices, whatever, ...
> The problem is they were wrong.
They weren't far off either. While PC sales haven't crashed as many feared, the majority of growth in larger form factor computing has been iPads & Chromebook. That plus the fact that a big percentage of desktop software has moved to Javascript/ Electron.
GNOME 3 actually does a far better job of integrating the "desktop" and "tablet/mobile/hybrid" worlds, and that's with a pretty conventional widget toolkit (Adwaita), albeit running on a loosely Win8-like shell (no "applications menu" or taskbar by default, they're replaced by an "overview" screen). So I'm pretty sure that the traditional Win32 UI would've been up to the task with a few tweaks.
I can't blame MS for trying to retrofit Metro on desktop. They were in a tight spot (apple and google were shining hard) and had to make a move. Market war.
I admit being sad that Metro was wasted though because it really was very nice on phones.
I wouldn’t mind that much, because I can just pin some stuff to the taskbar, press a button to type and find other stuff, and use a few apps. All I do is use a few apps. The OS just needs to make the network connect, show an image on the screen, and keep out of my way.
Unfortunately that’s not what Windows does any more. It lags. It does weird stuff. It keeps moving things around.
This isn’t just one installation. I’ve installed W10 a few times over the past few years and the lag, weirdness and cheese-moving continue in whatever installation I try to get along with.
There are plenty of useful things that can be done below the surface that would make Windows better over time. Breaking the interface due to incessant fiddling isn’t adding value: it’s removing it.
And yes, these are the main reasons I’m currently sticking with MacOS everywhere. I just need the OS to work and let me run my stuff in peace.
Who remembers how snappy the experience was on a fresh XP install?
Sure, everything loaded slow off disk, but once you got your basics into RAM it felt so much faster by comparison to how Win10 responds today. Explorer.exe in particular. What a gem that implementation was. I could play folder tetris at lightspeed with that thing, because it actually felt faster than I could ever go. Today, I am sad to report that despite my reflexes not being what they used to, I now wait upon file explorer to do its thing on a regular basis.
To see what I'm talking about, just run something extremely lightweight like mspaint, then right click it in your task bar. Be sure to have a stopwatch handy. It might take a little while for the context menu to appear and you don't want to lose track of your seconds. It's the perfect amount of delay right between complete disengagement (e.g. walking off to get coffee) and proper instantaneous feedback. I sometimes wonder if psychologists working at Microsoft have specifically tuned this delay to be as misanthropic as possible. I used XP not too long ago and I recall a visceral emotional experience due to the realization that we will never have a UI that feels that snappy again branded by Microsoft.
I realize I can probably get my low-latency UI fix with various Linux installs... Anyone have any recommendations? Who has the fastest window manager these days? With .NET Core working on Linux and any code targeting it being largely portable... I could maybe get used to a new normal on Linux for purposes of software development, which is also when I'm looking for a snappy UI. Visual studio on windows being the bloated nightmare it is... Some days I spend a few minutes writing specifications for a new IDE just as a means of catharsis. There is zero excuse for the kinds of delays I regularly see between keypress and UI feedback.
> I realize I can probably get my low-latency UI fix with various Linux installs... Anyone have any recommendations? Who has the fastest window manager these days?
LXDE (the original, not LxQT) is still well-supported on Debian Stable, and that's probably the closest thing to a 'XP-like or better' experience on Linux these days. Xfce is not far behind, though. Or you can experiment with more niche stuff such as WindowMaker, Fvwm95, IceWM and whatnot, but you're more likely to run into unimplemented stuff and these are simple window managers, so utility programs must be installed separately.
Of course, you should be mindful of RAM requirements. A modern Linux install will need at least 1GB to feel really snappy, which is far above the typical XP machine.
Even the nicest Linux desktop has more lag than ~KDE3 days. I think that's a global shift in how rendering is done, it used to be crude and glitchy but extremely cheap cycle wise, now you have compositing ... smooth matrices but maybe (I never wrote code for composited DE) this cause a fixed offset in response ?
The compositing step is done entirely on Vblank and is typically hardware-accelerated (hence no CPU overhead), but this means that updates to the screen only happen at next Vblank. With a single, non-composited framebuffer, an update happens as soon as the updated content is "scanned out" to the display, so you will save half a Vblank interval on average. (You will also incur some tearing but that only matters when showing media content, not in typical use.)
Most of the latency you see in modern Linux desktops comes from the tech stack, not the compositor. E.g. Plasma, and lots of other KDE components, are QML-based. You can see ths if you disable compositing. For example, you can run KDE with Openbox, without compositing, and lots of things are still laggy. Same thing with XFCE, where it's possible to disable xfwm4's compositor.
Not picking on KDE specifically -- I'm mentioning it because I use it on a few machines and it's the one that I'm most familiar with.
Compositing will inevitably give you some lag, but there's cap on how much, depending on the Vblank time. Extra lag is entirely self-inflicted :-).
I usually disable it on my work machine, too. I don't use flat themes anywhere, so I don't really need shadows, and I'm more than happy about the reduced latency.
>To see what I'm talking about, just run something extremely lightweight like mspaint, then right click it in your task bar...
>I'm looking for a snappy UI.
Problems like these I think stems from poor design decisions made at Microsoft as a whole.
Take ARC compared to Garbage Collection as an example. Not sure why Microsoft chose a Garbage Collector design when implementing .NET, but given how Garbage Collection works, there's always risk of it causing UI stutter when it's doing its thing. Apple in comparison chose an ARC design when implementing Swift to avoid potential responsive UI issues... even if ARC is overall less performant.
Then you also have the decisions to support backwards compatibility and you have a mess on your hands.
Frankly, I hate it somedays when I click on my windows calculator app and it opens a blank window and just sits there. I have to kill it and restart it 2-3 more times before it actually opens up. Stuff like this should never happen!
Sorry you're right, I meant tracing garbage collection.
>Same UI issues can happen in reference counting with deeply nested data structures.
Only if done poorly. The difference is this is deterministic and therefore an issue you can control and avoid with ARC (albeit at a cost to overall performance). The point was that ARC doesn't have the potential for a uncontrollable random process kicking in, stopping your app and causing UI stutter as it cleans up memory, whereas tracing GC does.
it makes me mad when non technical friends or family members of mine give me their Windows machines to "fix" something and I see all the garbage they are exposed to in the start menu.
Invasion of privacy and leeching of attention.
Not as bad as the boot times and the overall performance I suppose, but still ...
I'm glad I'm not the only one. You have reminded me I need to go see my aunt to see if a fresh install of Linux mint will solve her usual computing needs though. Or at least to figure out how to kill some of that trash.
God, I remember the near instantaneous boot of 7... I was giddy when I had that way back when. Now? I'm just angry at all the garbage.
Now the struggle will be to not forget how much these sucked, so that they don't come back in slightly different guise in five years, when the wheel has turned and a new generation of designers starts to think it is bright idea.
I wouldn't mind seeing the Desktop Gadgets from Vista and Windows 7, which were killed off in favor of Live Tiles, make a comeback.
There was a period of UI design in the late-2000's (late 90's, if you include BeOS) where desktop gadgets caught on, and I thought that was a pretty good implementation of the general idea.
That's a little generous, but the rise of smartphones (and their data connections) brought widgets and gadgets to the handheld, with-you-everywhere device and a little advertising on top didn't hurt. The motivated persons who developed the gadgets for free or donation moved to apps because it worked better for them _and_ they were able to get some extra money on top.
Then you have MS playing catch up and trying to merge mobile and desktop in all the wrong ways with 8 and the userbase shrunk. What was left was OS X and I'm not sure what happened there.
Windows 98 had its own flavor of network-powered "desktop gadgets", with its Active Desktop. It was supposed to download updated Web content in the background, with a homegrown RSS-like technology. It was quite modern for its time.
The Weather app and maybe News were the exceptions where these actually were nice to have. Though sometimes they updated on a significantly delay, and clicking a news item didn't even bring you to that news story. Vista's Gadgets were nicer because you didn't have to open the Start Menu to see the weather.
However, the tiles largely were duplicative of... just having a desktop shortcut (if you don't want a desktop shortcut, just being in the start menu is good enough).
Add the fact that Microsoft never really fully backed off pinning things like Candy Crush to this even for domain account users, and I won't really miss it much.
Even when it was "just desktop shortcuts" you could never get a desktop as clean or orderly as the tiles (without 3rd party tools like Fences).
Live Tiles at their best made the Notification Center redundant. Have all your notifications in the "same place" every time and at-a-glance easy to read.
> Vista's Gadgets were nicer because you didn't have to open the Start Menu to see the weather.
Windows 8 had the right idea in making the Start Screen the default and making the Win32 Desktop "inside it" rather than the other way around. There just wasn't enough courage to let Windows continue to innovate after they lost the initial PR war. (Charms were a really good idea, the execution just needed cleanup. A tiling window manager by default was a really good idea, but they needed to sell the execution by better bringing Win32 applications into that world/window manager without the bulk of the rest of the desktop having to hobble along.)
How about they get rid of the remnants of Win 95 UI which pops up every time you dig into anything administrative or system settings related? Would that be possible?
The Settings app does get a little more serviceable with each Windows 10 release. I am not sure all of those legacy UIs will ever completely fully disappear, but they're certainly becoming less necessary over time.
>but they're certainly becoming less necessary over time.
It's a trick. The Microsoft I know never - NEVER, ever recreates an existing software completely before retiring it. And they'll retire the incomplete replacement with something less functional soon after.
Changing this stuff would be equivalent to burning an appreciable portion of the internet to the ground. Not even counting the work involved in updating the UI and ensuring that equivalent functionality was retained, the amount of documentation that would need to be updated gives me the willies.
Given that the old control panels are more feature complete and one can actually find the stuff one is looking for, I'd rather vote for removing all the new settings panels since they don't seem to improve much despite Win10 being out for 5 years now.
The way they've fiddled with all the network settings and how you navigate to the place where you can set them drives me nuts. I did it the same way for like 20 years and now it's buggered up.
Nope. You're just fine. The current MS is just nuts. Imagine the breaks, accelerator, wind shield, steering for cars change positions for each new model.
It wouldn't be your fault for feeling disoriented.
They're working on it. I'm surprised so many developers complain about Microsoft's approach here. They're slowly moving over settings from the control panel to the settings app a little more with each Win10 release.
It will be impossible (or at least wildly impractical) to attempt to do this all at once. Iterative releases is entirely the correct approach to a problem of this scale.
Personally, I still prefer the old control panel but I also know it's a little intimidating for the average user. Having a mobile-phone like settings experience is going to b be less surprising for the majority of users.
It's not just a restyle though: it's in most cases a complete rewrite. At the very least in almost all cases it is a very new UI framework compared to anything the control panel was previously written in (WinRT/XAML versus organically grown madness across mixtures of MFC/ATL/Win32-by-hand/other forms of macro soup). But if Old New Thing and other Windows blogs make anything clear, it's also a ton of new testing, because Microsoft has no idea ever what sort of changes might domino into some user's beloved fifteen year old application that no one realized instead of using the right API for its job was instead scraping the pixels of the control panel for the right shade of purple.
Windows seems to be designed for people who aren’t savvy enough to use Linux, and don’t want to buy a Mac. They get away with egregious design practices, because the OS is ubiquitous and the common user has no other choice.
Mac OS on the other hand, has kept a fairly consistent design, mostly due to the obvious barrier to entry for most people ($$$), so any mistakes that would impact usability could easily give someone a reason to not spend $1k on a laptop or even more on a desktop.
Microsoft needs more competition. What’s stopping them from re-skinning Windows 10 to resemble Windows 7, and add actual functional improvements? They’ve done that recently, but they didn’t need to redesign their start menu to add workspaces (just one example).
> Windows seems to be designed for people who aren’t savvy enough to use Linux, and don’t want to buy a Mac
25 years of IT, developed for the Linux kernel in 1994.
I tried to use Linux on the desktop for years and years and it is dreadful, marginally unusable.
Windows just works great with my hardware and monitors.
I would not use a Windows machine as a server (mostly because I do not know it enough, but also because it does not seem right) and a Linux for desktop.
The weather app was literally the only UWP application I was using of my own free will and it was because of its live tile. I liked being able to check the weather by opening my start menu.
I hate the other UWP apps. They're slow, they lack features, they have horrible UIs and they don't integrate well with the rest of the OS. The only reason why I'm using the new "Settings" app or the new calculator is because Microsoft forced me to it.
It's such a shame that Microsoft refuses to acknowledge that UWP was a mistake and keeps porting apps to it.
Yes and no. In the days when all apps leech your data, I want the sandbox that UWP provides. Do not know if it is feasible for MS to make it work properly with "legacy" APIs.
They would have to do some sort of virtualization since there's literally no restrictions on what APIs you can access. There's so much cruft and invalid paradigms in the old APIs it makes sense to roll new ones.
I think the approach they're taking to refactoring the Control Panel into the Settings App is smart. They're dog-fooding the Frameworks they're expecting others to use, slowly migrating functionality that's relevant to a modern Windows platform, and keeping existing functionality around until they've replicated it.
I also like that the apps are responsive. I use the snap feature a lot and it's nice that the Apps will adapt to the dimensions I relegate them to.
I'm a .NET developer and I can't tell you how many times I've searched for documentation and come across something that's targeting UWP which is different than WPF and entirely not the same as WinForms and it takes way too much energy trying to identify why exactly the documentation isn't what you want. I thought Microsoft was more or less sunsetting WinForms until they also announced WinForms for .NET Core a couple months ago. https://docs.microsoft.com/en-us/dotnet/core/porting/winform...
I literally have no idea what Microsoft is doing anymore. Sometimes it feels like Linux was 10 years ago when I would accidentally install Konquerer on Gnome and all of a sudden I had two desktop systems installed and my root drive had doubled it's consumed harddrive space. It basically worked, but it was a terrible experience.
The Windows 10X preview actually does this by running Win32 apps inside their own container, with their own registry, kernel, drivers, and isolated from the rest of the rest of the system.
I don't have a problem with UWP and I really like the new Calculator. Being able to pin the calculator on-top and resize it so that it's touch friendly really boosts productivity for me on a laptop without a 10key. If you're not using a touchscreen then I can see the improvements not being that big of a deal but I'm not aware of any regressions in functionality.
I'm sorry, but I find the Calculator app woefully unimpressive.
I can't even type a simple expression with parenthesis. There's no RPN option. There isn't a "financial" setting (only scientific and programmer, for some reason). The list goes on; I'm sure I'm not alone on this one.
It is so, so, so basic it's essentially useless, I end up being forced to open Excel, run Python in the terminal, go on Wolfram Alpha or, worse yet (!!) grab pen and paper to perform basic math.
I know this is unrelated to UWP, but I'd rather they focus less on improving an UI that is inevitably trivial (little boxes with numbers arranged in a grid within a bigger box) and instead added some much needed features.
The fact that I am using a computer must be lost on whoever decided to green light this abomination as the default calculator in any OS ever, let alone one released in the 21st century.
Contributions doesn't mean just patches. Also, technically you didn't pay for Calculator, you paid for Windows and got a calculator. The fact that Calculator today is open source adds additional insight that it was always probably a "bonus" accessory, as the most interesting stuff happening to it today are from community contributions.
calc.exe never had RPN support. Some of y'all seem to be remembering a very different calculator than the one built in to Windows. "UWP" Calculator has feature parity with the calc.exe of Vista/7, and has features that calc.exe never had. I wish more of the source control history had been migrated to the git repository, but it's pretty clear from the header comments that one was built out of the codebase of the other (the core of the calculation engine itself goes way back).
"Yeah, well technically you bought a car, you didn't pay for an engine, so the engine being crap isn't something you can complain about."
I don't think this sort of argument works.
The reason it's open source is most likely because it's got nothing of value to competitors but lets them point and say "see, we've changed, look we make things open source now"; aka marketing.
I find that analogy weak because Windows works fine without a calc.exe or Calculator app, but a car doesn't run without an engine. Calculator is more a cup holder. I'm not trying to stop anyone for complaining about their cup holder, that's a great day-to-day productivity tool for a lot of car owners. Just pointing out that fit/finish/polish on a cup holder comes after the things like the engine.
Fwiw, it seems to go far beyond marketing that Calculator is open source, as community contributions have made it into releases and Calculator has seemed to benefit from more (and more diverse) development resources than it had in XP/Vista/7.
So become an unpaid Microsoft employee to re-add features that used to exist and that do exist in nearly every alternative? Don't they make enough money off of os sales, subscriptions, forced ads and telemetry? They need free labour to fix their broken software too?
Why not put those efforts towards an actual open source calculator project that could really use the volunteer efforts and doesn't already have an army of paid employees?
Is that a rhetorical question, or do I really get to argue the benefits of contributing to one of the most installed calculators on the planet? Contributing to small projects is good, but it's not for everyone and you're definitely kidding yourself if an MIT licensed software on Github doesn't qualify as "actual open source".
Yeah I tend to consider moral and ethical implications of my decisions also, I forget not everyone does.
I feel time spent working on projects for a company that hasn't spent the better part of 3 decades trying to dominate and control the entirety of the computer market using strategies that destroy competition and supress innovation, that tracks and monitors its users, that has such a small amount of respect that it will inject ads into a product most customers are forced to purchase should they wish to own a computer, is more valuable in that it makes the world just a slightly better place, rather than a slightly shittier one.
Though I didn't just mean patches from "contribute", it could also mean voice/ideas/discussion: instead of pseudonymously complain on HN, there's Feedback Hub and GitHub Issues. As an option that's available.
Yea, the new calculator is kinda bad.
But you can search the web for "windows old calculator" and replace the new one with this.
It is exactly the calculator as you are used to from Win7.
The regressions are in that it is much slower and wastes screen real estate. 90% of the window size is for faux buttons that have zero purpose for anyone who doesn't use a touch screen on their Windows computer (ie. at least 99% of the Windows users) and the 10% that remains uses oversized elements with wasteful padding.
and not only is much faster to open than Windows Calculator (ironic, considering it loads an entire browser engine) but also more compact and doesn't waste any screen space (100% of its window is dedicated to its actual functionality instead of being an obsolete skeuomorphism that was originally made to help people associate desktop calculators with the virtual desktop metaphor used on the Macintosh back in 1984 when actual physical desktop calculators could be found everywhere).
I bet a lot more than 1% of Windows users have touch screens at this point, since they've become more-or-less standard on high-end laptops.
Mind, it still could easily be < 10%, given how many people have old/cheap PCs.
I don't quite know what the solution is though. The last thing Windows 10 needs in addition to its two separate control panels, command lines, and web browsers are two separate calculator apps.
At least they only have one kernel running both Windows and the Linux subsystem. Because that is somehow better than just running a Linux VM with a real Linux kernel.
Does Windows not have something like the Mac's built-in Spotlight search? On the Mac, whereever I am, I can hit command-space, and in the resulting universal search pop-up I can type in math problems, unit/currency conversations, etc.
Even on iOS I use the search field for math & unit conversions rather than the Calc app, since it's easier to refine/fix your formula
It does - just press the Windows key to open the start menu and type your expression - but... the math is done by Bing. Which means that 1. it's unnecessarily slow and 2. without internet, your machine can't even show you the result of 2+2.
It's a travesty that you're even using Windows. With Arch Linux which is the best operating system, it's really appalling to hear that people use Windows these days. Come to arch my nigger. Fucking windows man listen to it. Arch Linux is the only sane operating-system my nigger.
My calculator is bc in a bash so I don't really care about either.
However, I really like the live weather tile, and I use it every day, sometimes several times a day by just pressing the start button, looking at the forecast, and pressing it again to go back to my app.
It was not intrusive and helpful. I also used the live calendar and mail that way (I hate notifications) -- and the photos too to see if new pics had been synchronized.
I'm afraid Windows 10 is going way from Windows 8, and closer to Windows 7. I don't think this is progress.
MS shot themselves in the foot with the store on PC. I don't get why I can't use the store from my browser. I want to sign in to the bloody store from any browser of my choice, click the download button and have store app do the downloads in the background.
One other problem the store has is that it handles poor network horribly. Or rather, it just can't work with poor internet.
Play store on the other hand is a master piece. It will open even offline (I tested it right now). It will show a cached page from last time.
Even settings work offline. And finally, the play store will alert you when you're back online.
MS store on the other hand will throw an incomprehensible ERROR Code at you.
They have obstacles to overcome, and then there is just the glaring fact that they had so much momentum with Win32.
My biggest gripe is that startup times are just too long. So many times I've gone to launch something like calculator only to have a beefy machine hang for almost a second, if not a full second. It's likely an engineering obstacle that can be overcome but it just makes UWP look like a bad move perf wise.
There was a brief period in Windows 8 where (proto) UWP apps launched smooth and fast and the Win32 desktop took several additional minutes to launch. With so many stacks living side-by-side on Windows it seems an interesting trade-off between past apps, present apps, and future wishes.
I hope future wishes don't involve adding yet another stack though. They need to fix the stuff they have (without breaking them), not introduce new stuff.
That's part of my sadness in seeing so many posts here talk bad about the UWP. The UWP barely had half a chance to live, and with Microsoft having to spend so much time in the Chromium codebase and working on Android code, with Electron and React Native seemingly becoming the most common development platforms for them, the future for Microsoft already doesn't seem to be UWP.
Cocoa was an improvement though, while UWP is a regression and a massive PITA to work with. Win32 isn't a great API either, and some corners are downright ugly, but when a successor needs more lines of code to get a window on screen but only delivers a small fraction of the features of the old system, something went seriously wrong.
Major Win32-only applications like Microsoft Office, SAP and countless in-house applications written over the last 25 years are the reason that many companies use Windows.
Microsoft understands that the Windows API "is so deeply embedded in the source code of many Windows apps that there is a huge switching cost to using a different operating system instead".[1] If they force companies to incur that cost by dropping Win32 support, that gives companies the chance to make their application cross-platform at a much smaller additional cost, and shed their reliance on Windows.
Windows licensing fees are still quite a large chunk of Microsoft's revenues, so I think that they will not make this move any time soon.
I can't answer your question exactly. I came around only after Cocoa was established.
This does make me wonder if the transition from Carbon to Cocoa is analogous though. Win32 comes off as more of a "functional" approach to programming (C) and UWP is looking more object oriented (as to C++). It makes me wonder what kinds of conceptual advantages are being brought on board by moving from win32 to UWP. I think that conceptually I am in favor of the transition, and I'd wager that all the issues that arise are common to a functional to object-oriented transition.
But that real questions: What are the benefits? Is it conceptually easier to grasp thus promoting more developer interaction? Are we expecting speed benefits (are the performance losses expected and in range?) Is it a false dichotomy to only look at Functional vs OO? Maybe a context-oriented data approach is more appropriate?
Tbh outside of maybe a sandboxing aspect, UWP made zero sense to me when Windows Phone fully died. Before that, the ability to have one codebase target both desktop and mobile seemed interesting at least, but now, it would seem that there are only a tiny of handful of applications that could be useful as universal between the remaining UWP targets (desktop, xbox, IoT, whatever else)
UWP is what .NET should have been since the beginning, an improved COM, and an alternative design solution to Longhorn ideas.
From that point of view it makes perfect sense.
What it did not make sense was creating an incompatible .NET runtime for it, or the way it was introduced with device specific APIs (Win 8, Win 8.1), using shared code folders.
Why don’t they just have a toggle, either global or per application. Live tiles, yes/no. Then the user has a choice of high/low resource usage and “busy-ness”
I’ve definitely enabled/disabled the tiles on an individual basis. I just don’t recall whether it’s universally available to all programs, or a particular subset.
Yes. Yes they do. Quite a lot. They take up space reminding me how the alphabet works. This "C" is to indicate that the following programs begin with the letter C. This "D" shows that the next items have names where D is the common starting character. And so on, and so forth.
Actually, I just looked at this .. and they have a function (who knew??). If you click then it shows the alphabet as a block in a dialog and you can jump to a different letter!
I'd still get rid of them (or push/float the app links to the right) and save the vertical space.
The behaviour you found out is an inheritage from the good old Windows Phone 7.
I use it quite frequently, because when I'm going through the list it's usually because I couldn't reach to the program I wanted by just typing the name. So if what I need it's under "W", then it's a way to be quicker.
Imagine a world where Microsoft cares and does not force people to use something like this at the first place. It would be trivial to A/B test if tiles were a good idea, if they are unsure.
The sad thing is that with things like WSL, VSCode, chocolatey/scoop and .NET Core, MSFT has done a good job of fostering a decent platform on which to do modern software development but then completely squandered a lot of that goodwill by turning Windows 10 into an ad delivery platform.
Nice. Now kill UWP[0] and bring some actual improvements to the base Win32 C API that any application written in any programming language and available from any source (direct downloads form author pages, repositories like GitHub, eshops such as Steam, etc) can use.
[0] of course for backwards compatibility reasons this cannot be totally removed, though considering that 99.999% of the applications made for it were obtainable through the online Microsoft Store there is an assumption of internet availability so it can be installed on-demand like older versions of .NET framework.
I'm not sure if this is what you have in mind (because sandbox/container can mean several things, including things i wouldn't like to see like limiting what applications can do to each other) but one thing that i'd like to see is something like a "desktop chroot": being able to right click an EXE (or zip/installer/msi/whatever) and select "Run this in fresh Windows 10 VM" where the application runs in a VM (either inside its own window or seamlessly, could be an option) as if it was the only application installed in a fresh Windows 10 installation, optionally with a way to make this VM permanent as a desktop shortcut, until you decide to delete it and remove the entire thing as if it never existed. This would make it easy to experiment with applications you do not fully trust or simply trying out applications in an environment that wont affect your real working one.
It is kinda/sorta possible to do that now using something like VirtualBox, but you need to create a full Windows 10 installation (which takes several GBs), it takes time to install (you can save some time by exporting a fresh installation to an appliance - at expense of more space - though you'd also need to disable networking otherwise after a while the moment it notices internet will start upgrading) and it can't use the computer resources properly (it runs much slower, you have no access to the GPU, etc). A properly integrated Windows 10 solution would be to map existing files to VM so that it reuses 99% of the stuff that would be the same and, if needed, do a copy-on-write in cases where these files are altered inside the VM (it could also help to see what exactly the programs inside the VM touch, sort of a system-wide diff... though that can also have false positives if the changes are by the OS itself).
Kinda. But it should be available to everyone, not just Pro, and it should be permanent until you decide to delete it instead of getting deleted after you close the application (as you may want to take your time using the application for a while before you decide if it is worth installing on your base system). This also implies that it should be possible to move files between the VM and the base system, preferably in a manner as simple as drag-and-dropping the files to/from the VM.
... and it shouldn't break hybrid sleep, either (by the way, is that still a fundamental limitation of using Hyper-V, or just a peculiarity of my specific system configuration?).
Would sandboxing make my old 32bit midi apps stop talking to my new 64bit midi apps? If yes, that would be a shame, since the ability to still use ancient midi software and hardware is one of the things that’s nice about Windows 10.
> Now kill UWP[0] and bring some actual improvements to the base Win32 C API that any application written in any programming language and available from any source (direct downloads form author pages, repositories like GitHub, eshops such as Steam, etc) can use.
People said the same thing about Carbon on macOS—"why do I have to use this new Cocoa thing?"—but now nobody misses it. The Win32 API needs to go away (at least as the main way to develop apps) sooner or later. (It already is gone to a major extent: when was the last time a new Win32 app, written in Win32 with C++, made the news?)
I don't know what people said about Carbon, but i do know that many people do miss it - like the Lazarus developers that had to throw away years of work.
> It already is gone to a major extent: when was the last time a new Win32 app, written in Win32 with C++, made the news?
Every single PC game made since the mid90s?
Also why the restriction to C++? A big benefit of Win32 is that being in C it can be used by many other languages since pretty much every language has some FFI that speaks the platform's C ABI.
When was the last time a new UWP app made the news (except that development on it was stopped). Nearly all apps are written for Electron now, and that sits on top of Win32 ;)
Not quite; it's more like a cross between a widget and a shortcut. The idea being that instead of having a Weather Widget and also having a shortcut to the weather app from the app drawer, the icon for the weather app would just be the widget. It feels different.
The details are a matter of taste of course, but Android widgets do kind of provide this as well: the home screen has a mix of shortcuts and widgets (corresponding to pinned apps in your app drawer, I guess?) and clicking the right spot in the widget usally brings up the app, so usually I use the calendar, weather and notes widgets as (or in preference to) shortcuts to open the apps.
It takes 5 minutes to whip up a VisualElementsManifest.xml and a custom PNG to make your own static tiles for applications, and we can't even get companies to adopt that. Steam should have had a full win10 static tile ages ago.
So if they can't even adopt static tiles, no wonder the live tiles have received nearly zero traction.
Now we only need option to install windows WITHOUT ANY APP PINNED IN THE START MENU - and while we are at it, option to not actually install anything but kernel + most basic stuff that can't be removed - explorer, shell & friends - no candy crash, maps, movie maker and other useless garbage that one could install via Store anytime anyway
The first thing I do is to remove them all with powershell script.
Is there a Chrome/Firefox equivalent of tiles for the homepage? I didn't appreciate tiles when they screwed up the Windows 8 Start Menu but I think a nice use case would be a browser homepage. (And something with more flexibility than just "here's your most recently opened windows").
189 comments
[ 3.4 ms ] story [ 316 ms ] threadTo me, the "advancement" of the human-computer interface came to a screeching halt at that time. People grok menus. Memorizing sequences of Mnemonic accelerators is natural when overlaid on a haptic path created by mouse movement/menu navigation.
Then you take everything and throw out the baby with the bath water by trying to put every past bit of functionality on screen at the same time when what one needs is to be able to set up a work bench quickly, and just do your work, and if necessary save said work bench state to return to later.
I've seen more and "better" advancements in UI design in exactly 1 place, in all the different software suites I use, and that is in IDEs. Ironically, the one place that programmers have to make due day in, day out.
It makes me wonder whether anyone who presides over these programs is actually proficient in the non-computing version of said task. A Typesetter/document composer for instance. Nothing about the modern interface leads me to believe that any actually invested individual is involved, and that the design was pushed by the need for someone at Microsoft to be seen as needed in order to change something.
Then again... Can"t judge 'em for that. We all want to be wanted.
What I hate is the save / options / preference / file menu of office 2016 upwards. It's a nightmare - just too many steps to save a simple file to disk. Even after using it for years, I still need to pay full attention in this menu rather than muscle memory.
Just because they are pushing OneDrive.
Even after years I prefer menus and toolbars over the ribbon. With them you can scan in one direction whereas with the ribbon you have to up and down, left and right. I find it really difficult to find something there.
No, what happened is Microsoft placed a bet on us moving to a multi-platform future where there are many fewer PCs and more tablets, hybrid devices, whatever, all with the need to use touch, pen as well as keyboard for input. The existing start menu and taskbar would be hideous on mobile devices.
The problem is they were wrong. But I'm not going to blame "designer types". If you ever used Metro UI on a Windows Phone device you'd see that it was a truly remarkable system... that was bent to fit a task it didn't fit.
For choice, they finally relented with windows 8.1 and allowed people to boot to desktop directly rather than the start screen. Windows 10 took off the start screen completely.
Choice matters. I liked the start screen but the horizontal scrolling killed it. Like pouring salt in Tea.
Too many changes.
Facebook's learned this lesson well. They no longer do large overhauls. Small changes at a time and in a few months, it's a whole different interface. Only infrequent users would notice.
That's what they are doing with Windows 10 - many small interface changes. The little changes are mainly minor inconveniences.
One last thing.
When they released windows 8 - hardware support was almost non existent. Touch screen PC from that era really sucked. Now with detachable screens, Win 8 style pc interactions might be cool.
> The problem is they were wrong.
They weren't far off either. While PC sales haven't crashed as many feared, the majority of growth in larger form factor computing has been iPads & Chromebook. That plus the fact that a big percentage of desktop software has moved to Javascript/ Electron.
I admit being sad that Metro was wasted though because it really was very nice on phones.
Unfortunately that’s not what Windows does any more. It lags. It does weird stuff. It keeps moving things around.
This isn’t just one installation. I’ve installed W10 a few times over the past few years and the lag, weirdness and cheese-moving continue in whatever installation I try to get along with.
There are plenty of useful things that can be done below the surface that would make Windows better over time. Breaking the interface due to incessant fiddling isn’t adding value: it’s removing it.
And yes, these are the main reasons I’m currently sticking with MacOS everywhere. I just need the OS to work and let me run my stuff in peace.
Besides a shit UI, the creepy phoning home and ads created many a defector away from MS.
Sure, everything loaded slow off disk, but once you got your basics into RAM it felt so much faster by comparison to how Win10 responds today. Explorer.exe in particular. What a gem that implementation was. I could play folder tetris at lightspeed with that thing, because it actually felt faster than I could ever go. Today, I am sad to report that despite my reflexes not being what they used to, I now wait upon file explorer to do its thing on a regular basis.
To see what I'm talking about, just run something extremely lightweight like mspaint, then right click it in your task bar. Be sure to have a stopwatch handy. It might take a little while for the context menu to appear and you don't want to lose track of your seconds. It's the perfect amount of delay right between complete disengagement (e.g. walking off to get coffee) and proper instantaneous feedback. I sometimes wonder if psychologists working at Microsoft have specifically tuned this delay to be as misanthropic as possible. I used XP not too long ago and I recall a visceral emotional experience due to the realization that we will never have a UI that feels that snappy again branded by Microsoft.
I realize I can probably get my low-latency UI fix with various Linux installs... Anyone have any recommendations? Who has the fastest window manager these days? With .NET Core working on Linux and any code targeting it being largely portable... I could maybe get used to a new normal on Linux for purposes of software development, which is also when I'm looking for a snappy UI. Visual studio on windows being the bloated nightmare it is... Some days I spend a few minutes writing specifications for a new IDE just as a means of catharsis. There is zero excuse for the kinds of delays I regularly see between keypress and UI feedback.
LXDE (the original, not LxQT) is still well-supported on Debian Stable, and that's probably the closest thing to a 'XP-like or better' experience on Linux these days. Xfce is not far behind, though. Or you can experiment with more niche stuff such as WindowMaker, Fvwm95, IceWM and whatnot, but you're more likely to run into unimplemented stuff and these are simple window managers, so utility programs must be installed separately.
Of course, you should be mindful of RAM requirements. A modern Linux install will need at least 1GB to feel really snappy, which is far above the typical XP machine.
Mint + XFCE is really a fine combo.
Not picking on KDE specifically -- I'm mentioning it because I use it on a few machines and it's the one that I'm most familiar with.
Compositing will inevitably give you some lag, but there's cap on how much, depending on the Vblank time. Extra lag is entirely self-inflicted :-).
I usually disable it on my work machine, too. I don't use flat themes anywhere, so I don't really need shadows, and I'm more than happy about the reduced latency.
>I'm looking for a snappy UI.
Problems like these I think stems from poor design decisions made at Microsoft as a whole.
Take ARC compared to Garbage Collection as an example. Not sure why Microsoft chose a Garbage Collector design when implementing .NET, but given how Garbage Collection works, there's always risk of it causing UI stutter when it's doing its thing. Apple in comparison chose an ARC design when implementing Swift to avoid potential responsive UI issues... even if ARC is overall less performant.
Then you also have the decisions to support backwards compatibility and you have a mess on your hands.
Frankly, I hate it somedays when I click on my windows calculator app and it opens a blank window and just sits there. I have to kill it and restart it 2-3 more times before it actually opens up. Stuff like this should never happen!
Same UI issues can happen in reference counting with deeply nested data structures.
Sorry you're right, I meant tracing garbage collection.
>Same UI issues can happen in reference counting with deeply nested data structures.
Only if done poorly. The difference is this is deterministic and therefore an issue you can control and avoid with ARC (albeit at a cost to overall performance). The point was that ARC doesn't have the potential for a uncontrollable random process kicking in, stopping your app and causing UI stutter as it cleans up memory, whereas tracing GC does.
A good comparison: http://developer-toolbox.blogspot.com/2017/12/garbage-collec...
Invasion of privacy and leeching of attention.
Not as bad as the boot times and the overall performance I suppose, but still ...
God, I remember the near instantaneous boot of 7... I was giddy when I had that way back when. Now? I'm just angry at all the garbage.
Oh gosh the visual noise! I couldn't stand it the moment I saw it.
What is all this stuff, changing every second? It completely disabled my ability to focus or find anything.
I wouldn't mind seeing the Desktop Gadgets from Vista and Windows 7, which were killed off in favor of Live Tiles, make a comeback.
Then it all vanished for some reason.
That's a little generous, but the rise of smartphones (and their data connections) brought widgets and gadgets to the handheld, with-you-everywhere device and a little advertising on top didn't hurt. The motivated persons who developed the gadgets for free or donation moved to apps because it worked better for them _and_ they were able to get some extra money on top.
Then you have MS playing catch up and trying to merge mobile and desktop in all the wrong ways with 8 and the userbase shrunk. What was left was OS X and I'm not sure what happened there.
However, the tiles largely were duplicative of... just having a desktop shortcut (if you don't want a desktop shortcut, just being in the start menu is good enough).
Add the fact that Microsoft never really fully backed off pinning things like Candy Crush to this even for domain account users, and I won't really miss it much.
Live Tiles at their best made the Notification Center redundant. Have all your notifications in the "same place" every time and at-a-glance easy to read.
> Vista's Gadgets were nicer because you didn't have to open the Start Menu to see the weather.
Windows 8 had the right idea in making the Start Screen the default and making the Win32 Desktop "inside it" rather than the other way around. There just wasn't enough courage to let Windows continue to innovate after they lost the initial PR war. (Charms were a really good idea, the execution just needed cleanup. A tiling window manager by default was a really good idea, but they needed to sell the execution by better bringing Win32 applications into that world/window manager without the bulk of the rest of the desktop having to hobble along.)
It's a trick. The Microsoft I know never - NEVER, ever recreates an existing software completely before retiring it. And they'll retire the incomplete replacement with something less functional soon after.
Changing this stuff would be equivalent to burning an appreciable portion of the internet to the ground. Not even counting the work involved in updating the UI and ensuring that equivalent functionality was retained, the amount of documentation that would need to be updated gives me the willies.
Maybe this is a sign im finally getting old.
Nope. You're just fine. The current MS is just nuts. Imagine the breaks, accelerator, wind shield, steering for cars change positions for each new model.
It wouldn't be your fault for feeling disoriented.
It will be impossible (or at least wildly impractical) to attempt to do this all at once. Iterative releases is entirely the correct approach to a problem of this scale.
Personally, I still prefer the old control panel but I also know it's a little intimidating for the average user. Having a mobile-phone like settings experience is going to b be less surprising for the majority of users.
[1] https://web.archive.org/web/20200225190518/https://www.lapto...
Mac OS on the other hand, has kept a fairly consistent design, mostly due to the obvious barrier to entry for most people ($$$), so any mistakes that would impact usability could easily give someone a reason to not spend $1k on a laptop or even more on a desktop.
Microsoft needs more competition. What’s stopping them from re-skinning Windows 10 to resemble Windows 7, and add actual functional improvements? They’ve done that recently, but they didn’t need to redesign their start menu to add workspaces (just one example).
25 years of IT, developed for the Linux kernel in 1994.
I tried to use Linux on the desktop for years and years and it is dreadful, marginally unusable.
Windows just works great with my hardware and monitors.
I would not use a Windows machine as a server (mostly because I do not know it enough, but also because it does not seem right) and a Linux for desktop.
I hate the other UWP apps. They're slow, they lack features, they have horrible UIs and they don't integrate well with the rest of the OS. The only reason why I'm using the new "Settings" app or the new calculator is because Microsoft forced me to it.
It's such a shame that Microsoft refuses to acknowledge that UWP was a mistake and keeps porting apps to it.
I think the approach they're taking to refactoring the Control Panel into the Settings App is smart. They're dog-fooding the Frameworks they're expecting others to use, slowly migrating functionality that's relevant to a modern Windows platform, and keeping existing functionality around until they've replicated it.
I also like that the apps are responsive. I use the snap feature a lot and it's nice that the Apps will adapt to the dimensions I relegate them to.
I literally have no idea what Microsoft is doing anymore. Sometimes it feels like Linux was 10 years ago when I would accidentally install Konquerer on Gnome and all of a sudden I had two desktop systems installed and my root drive had doubled it's consumed harddrive space. It basically worked, but it was a terrible experience.
Also, drivers? o-O
I can't even type a simple expression with parenthesis. There's no RPN option. There isn't a "financial" setting (only scientific and programmer, for some reason). The list goes on; I'm sure I'm not alone on this one.
It is so, so, so basic it's essentially useless, I end up being forced to open Excel, run Python in the terminal, go on Wolfram Alpha or, worse yet (!!) grab pen and paper to perform basic math.
I know this is unrelated to UWP, but I'd rather they focus less on improving an UI that is inevitably trivial (little boxes with numbers arranged in a grid within a bigger box) and instead added some much needed features.
The fact that I am using a computer must be lost on whoever decided to green light this abomination as the default calculator in any OS ever, let alone one released in the 21st century.
Anyway, for instance, you could see that RPN is on the radar: https://github.com/microsoft/calculator/issues/128
I don't think this sort of argument works.
The reason it's open source is most likely because it's got nothing of value to competitors but lets them point and say "see, we've changed, look we make things open source now"; aka marketing.
Fwiw, it seems to go far beyond marketing that Calculator is open source, as community contributions have made it into releases and Calculator has seemed to benefit from more (and more diverse) development resources than it had in XP/Vista/7.
https://qalculate.github.io/ https://bitbucket.org/heldercorreia/speedcrunch/
I feel time spent working on projects for a company that hasn't spent the better part of 3 decades trying to dominate and control the entirety of the computer market using strategies that destroy competition and supress innovation, that tracks and monitors its users, that has such a small amount of respect that it will inject ads into a product most customers are forced to purchase should they wish to own a computer, is more valuable in that it makes the world just a slightly better place, rather than a slightly shittier one.
Though I didn't just mean patches from "contribute", it could also mean voice/ideas/discussion: instead of pseudonymously complain on HN, there's Feedback Hub and GitHub Issues. As an option that's available.
Use Scientific mode and parentheses work fine
I've not used a physical for some years, but even then the Casio ones had a backspace, and had apparently done so for a good few generations.
Here is what i use as a calculator on Windows: http://runtimeterror.com/tools/calc/calc.hta.txt
It looks like this: https://i.imgur.com/eWZfcXB.png
and not only is much faster to open than Windows Calculator (ironic, considering it loads an entire browser engine) but also more compact and doesn't waste any screen space (100% of its window is dedicated to its actual functionality instead of being an obsolete skeuomorphism that was originally made to help people associate desktop calculators with the virtual desktop metaphor used on the Macintosh back in 1984 when actual physical desktop calculators could be found everywhere).
I bet a lot more than 1% of Windows users have touch screens at this point, since they've become more-or-less standard on high-end laptops.
Mind, it still could easily be < 10%, given how many people have old/cheap PCs.
I don't quite know what the solution is though. The last thing Windows 10 needs in addition to its two separate control panels, command lines, and web browsers are two separate calculator apps.
Even on iOS I use the search field for math & unit conversions rather than the Calc app, since it's easier to refine/fix your formula
However, I really like the live weather tile, and I use it every day, sometimes several times a day by just pressing the start button, looking at the forecast, and pressing it again to go back to my app.
It was not intrusive and helpful. I also used the live calendar and mail that way (I hate notifications) -- and the photos too to see if new pics had been synchronized.
I'm afraid Windows 10 is going way from Windows 8, and closer to Windows 7. I don't think this is progress.
I dont see any advantage of win calculator over it.
> They're slow, they lack features, they have horrible UIs and they don't integrate well with the rest of the OS.
This is the case with almost all android apps too.
[1]: https://answers.microsoft.com/en-us/windows/forum/windows8_1...
One other problem the store has is that it handles poor network horribly. Or rather, it just can't work with poor internet.
Play store on the other hand is a master piece. It will open even offline (I tested it right now). It will show a cached page from last time.
Even settings work offline. And finally, the play store will alert you when you're back online.
MS store on the other hand will throw an incomprehensible ERROR Code at you.
My biggest gripe is that startup times are just too long. So many times I've gone to launch something like calculator only to have a beefy machine hang for almost a second, if not a full second. It's likely an engineering obstacle that can be overcome but it just makes UWP look like a bad move perf wise.
Do you feel the same about the macOS Carbon-Cocoa transition? Nobody misses Carbon anymore, and Win32 is as old as the classic Mac Toolbox.
Eventually they have to shed the old API.
The comparison between Apple and Microsoft vis-à-vis backward compatibility is illuminating, but not in the way you seem to believe it is.
Microsoft understands that the Windows API "is so deeply embedded in the source code of many Windows apps that there is a huge switching cost to using a different operating system instead".[1] If they force companies to incur that cost by dropping Win32 support, that gives companies the chance to make their application cross-platform at a much smaller additional cost, and shed their reliance on Windows.
Windows licensing fees are still quite a large chunk of Microsoft's revenues, so I think that they will not make this move any time soon.
1. https://www.zdnet.com/article/microsoft-wed-have-been-dead-a...
This does make me wonder if the transition from Carbon to Cocoa is analogous though. Win32 comes off as more of a "functional" approach to programming (C) and UWP is looking more object oriented (as to C++). It makes me wonder what kinds of conceptual advantages are being brought on board by moving from win32 to UWP. I think that conceptually I am in favor of the transition, and I'd wager that all the issues that arise are common to a functional to object-oriented transition.
But that real questions: What are the benefits? Is it conceptually easier to grasp thus promoting more developer interaction? Are we expecting speed benefits (are the performance losses expected and in range?) Is it a false dichotomy to only look at Functional vs OO? Maybe a context-oriented data approach is more appropriate?
Dunno.
From that point of view it makes perfect sense.
What it did not make sense was creating an incompatible .NET runtime for it, or the way it was introduced with device specific APIs (Win 8, Win 8.1), using shared code folders.
I’ve definitely enabled/disabled the tiles on an individual basis. I just don’t recall whether it’s universally available to all programs, or a particular subset.
I don't really understand how having a handful of huge icons helps anyone, especially when you can pin icons of apps you actually use to your taskbar.
I'd still get rid of them (or push/float the app links to the right) and save the vertical space.
I use it quite frequently, because when I'm going through the list it's usually because I couldn't reach to the program I wanted by just typing the name. So if what I need it's under "W", then it's a way to be quicker.
Only a minor issue though compared to the search being so utterly broken after so many years.
[0] of course for backwards compatibility reasons this cannot be totally removed, though considering that 99.999% of the applications made for it were obtainable through the online Microsoft Store there is an assumption of internet availability so it can be installed on-demand like older versions of .NET framework.
Win32 ("WinAPI" nowadays) is the API everyone will keep using (plus .NET for higher level).
Make a sandbox/container for Win32 apps and that's it.
It is kinda/sorta possible to do that now using something like VirtualBox, but you need to create a full Windows 10 installation (which takes several GBs), it takes time to install (you can save some time by exporting a fresh installation to an appliance - at expense of more space - though you'd also need to disable networking otherwise after a while the moment it notices internet will start upgrading) and it can't use the computer resources properly (it runs much slower, you have no access to the GPU, etc). A properly integrated Windows 10 solution would be to map existing files to VM so that it reuses 99% of the stuff that would be the same and, if needed, do a copy-on-write in cases where these files are altered inside the VM (it could also help to see what exactly the programs inside the VM touch, sort of a system-wide diff... though that can also have false positives if the changes are by the OS itself).
People said the same thing about Carbon on macOS—"why do I have to use this new Cocoa thing?"—but now nobody misses it. The Win32 API needs to go away (at least as the main way to develop apps) sooner or later. (It already is gone to a major extent: when was the last time a new Win32 app, written in Win32 with C++, made the news?)
Windows 10, for example. And the compiler stacks for Windows. I can go on forever.
All major new APIs since Vista are based on COM, with UWP being COM vNext.
WinUI is becoming the basis of React Native for Windows, and is the only Windows UI toolkit with a roadmap beyond bug fixing.
> It already is gone to a major extent: when was the last time a new Win32 app, written in Win32 with C++, made the news?
Every single PC game made since the mid90s?
Also why the restriction to C++? A big benefit of Win32 is that being in C it can be used by many other languages since pretty much every language has some FFI that speaks the platform's C ABI.
So if they can't even adopt static tiles, no wonder the live tiles have received nearly zero traction.
The first thing I do is to remove them all with powershell script.
Link: https://github.com/Open-Shell/Open-Shell-Menu
Screenshot: https://i.imgur.com/ujspeiD.png
Ignore the shit font rendering, I have every option toggled for max performance over RDP.