Hacker News new | past | comments | ask | show | jobs | submit login
Windows Will Die: 90 Minutes to Do 5 Minutes of Work (gavinhoward.com)
25 points by gavinhoward on Sept 24, 2022 | hide | past | favorite | 115 comments



> I don’t boot into Windows often; it’s only to ensure my software works on it.

There's the issue. And it's the same issue I constantly hear from people who use Windows similar to how the author does. The problem? They use Windows so infrequently that every time they boot into, and out of, it it start downloading/installing/preparing updates + background tasks that have accrued over weeks/months.

Is this situation ideal for the author's use case? Absolutely not. But does it actually get in the way of office work, like the author references? Absolutely not, since those machines will be booting daily and get hiccups, not slogs.


Agree with what you're saying but the way Windows handles updates is probably my biggest bugbear with it so to me this more or less amounts to victim blaming. If I want to boot my machine for five minutes after a month of inactivity just to accomplish some quick task, why should I be forced into a thirty+ minute update process? Likewise those updates that happen in the background while you're working and when you need to reboot, surprise! You have to wait five-fifteen minutes while it does (something?) before you can proceed with your day.

I'm not totally down on Windows in general but this, coupled with other problems, has led me to switch to Fedora as my daily driver and so far I can't see myself ever going back.


Forcing updates on the majority of the population has absolutely been a net positive on the world. The lack of exploitable machines around easily out weights a slightly faster access in an odd use case.


That's a false dichotomy - there are other workflows available. And it's not an odd use case for me at least; I've cursed Windows Update more times than I can count in the last five years.


This has nothing to do with forcing the updates - just the way they're installed. If I boot a Linux distro after a few months it can be updated in the background in minutes. Then either you restart the services or they get restarted with the next reboot. This mostly does not affect your workflow at all.


> just the way they're installed

> they get restarted with the next reboot

And if after the next reboot your services or OS dies - it's only on you. But not if this is Windows OS - then it's Microsoft, all Redmond citizens and Gates personally is the culprit.

You can restart 99% of services in Windows just fine, it's just most of the software written by bazillion of 'software developers' always find a way to break if something suddenly disappears from the system. And no, they do NOT write a proper dependencies for their services, hell, they still install in 'C:\SOFTWARENAME' because they still can't in the long file names [including spaces] AND they are still shitting in their install folder (Hello, 'C:\Program Files\MariaDB 10.5 (x64)\data\ibdata1'!).

Add shitty drivers for a bazillions devices, written by other shitty developers - and now it is way, way, WAY easier to just reboot the machine.

And don't forget: your shitty webcam managed to crash the system while it updated? It's Gates fault, not the author of webcam's driver.

EDIT: Oh great, Baader Meinhof as it is, while I moved my workstation VM:

You know what is forcefully killed explorer.exe first (hey, it's my shell! I have opened some windows there and there!) and then rebooted the whole machine WITHOUT asking anything from me?

Fucking Nextcloud client.

And this is not the first time when this shit happens.


I agree. Windows updates and I are not on good terms ever since they decided that they knew better than me for which audio drivers to have installed and force uninstalled drivers a few years back in one of their bigger updates. I can't reinstall these drivers without windows automatically uninstalling them each update, so I've since had to use virtual audio cables and an equalizer to have audible voice input.


If its a problem you can disable updates.

The alternative is no updates by default, and we saw what a mess that was back in the win 95/2000 era.


Only with Enterprise or LTSB/LTSC but even that is changing as far as I know, and most people don't have access to those distributions anyway. Besides I don't want to turn off updates - that's a false dichotomy. I want to be allowed to set my own update schedule like someone who knows what they're doing with their own PC.


It works for windows 10 pro also - my laptop has been sitting pending a restart for about a week and a half now, and my shutdown menu has options to restart/shutdown without updating.


Can you? On my machine, you can only pause them for some time.


Yes, Shutup10 [1a][1b] and misc other powershell scripts people have written can disable some things with varying degrees of success.

[1a] - https://www.oo-software.com/en/shutup10 [freeware]

[1b] - https://www.virustotal.com/gui/file/86bf0b60f97f72bdbbf01e45... [scan results]


You need windows pro, but you basically set a bunch of registry keys/GPO to disable updates.


Honestly installing Windows updates is far less painful than installing MacOS updates - they seem to take an absolute AGE to finish, whereas Windows updates, even to relatively major update versions, are much faster.

Heck, even an Ubuntu dist-upgrade is not a particularly quick process.


dist-upgrade is a completely different thing to windows update.


I shouldn't need to run my laptop for days at 100% to please some Windows Store optimization service, just because I haven't booted it up for a month. I have a desktop, the laptop is for emergencies. But it's useless because Windows insists on emptying the battery when I need it.


You don't need to run your laptop for days at 100%. That's complete fiction.


Nope. It can do this for days if you leave the laptop off long enough. And then the CPU usage will suddenly drop to 0% when it's done. It's not even updating, it's Windows Store / universal app shenanigans.

You're assuming it's actually doing something important. I'm assuming there's a bug in Windows. Which is more likely?


I still find it hard to believe that Windows will take your CPU to 100% for literal days given any vacation from use.

That being said, I realized after replying that I removed most of the background processes like the store, so I was thinking mostly just update related stuff. I have no idea how bad some of the "optional" services are.


The time it runs at 100% seems to go up exponentially the longer you leave the system off.

It just seems like very bad programming to me.


Could it be all of your windows store apps updating? The longer you leave it, the more than will have updates?


yeah this is like someone complaining their petrol car keeps gumming up because they only start it every 2 years.


I have a box with openSUSE that hasn't been powered on in a couple years. If I plugged it in and powered it on today, I could use it right away with none of the crap that has been described in the article.

Of course I would want to do an upgrade in due time, but that's on me, and won't make the system unusable.


We can make a conclusion on this case being seen by Microsoft as non meaningful on average and from my understanding for those who care, they have options to be applied by your Corp IT team to enforce the needed type of policy.

Probably I'd do the same - professionalls have way to configure system, non-pros better be guided/enforced a bit


I have Linux installed on my laptop. It's always ready to go no matter what. I only update it when I want to. Usually every few months now, but I've left it alone for over a year and there were absolutely zero problems.


Corollary, I have an infrequently used Ubuntu install on my laptop which I recently booted up to run a serial terminal emulator on.

Went to install minicom and apt was locked due to package manager doing updates.


Windows chose to act behind users with it's OBSCENE update system because Windows designer consider users dumb and that instead of try educating them they need to act replacing people.

This is a business move because you can munge more money from dumb than from smart people but is definitively a bad choice for the society in general and for business in long terms.

Since first Xerox desktop concept implementation (desktop model of computing if far older) ALL major players try to cut out users control, initially IBM then Microsoft then the GAFAM, all try to sell locked-in tools able to do only few things in few way to sell more and the result is an explosion of crap to the point that modern systems fall on their own weight. And beware, a desktop these days it's just an endpoint or to use an old term a dumb terminal of remote services, typically on some giant's cloud. Your car crushing while driving at high speed will be a bit more lethal, and that's already happen with current "drive aids" like "speed limiter" who read road sign confusing for instance exit roads (sorry I do not the correct name in English, the small roads you take to going out of a highway and so on) signs to main roads one auto-breaking strongly and suddenly the vehicle or anti-collision systems who use a simple stereocam behind the internal mirror and see a flying leaf for a wall, just to cite something I've experienced personally.

Remember a thing: humans MUST ALWAYS have the last word on anything. Any tool must be in TOTAL control of the human it serve. Sometimes such total control will makes disasters that's for sure, BUT many other times avoid even bigger ones.


Agreed, but as a sidenote:

It's crazy that this is how it works. No other OS does this. I can not not into MacOS or Linux for years, and this still wouldn't happen.


I don't think it's crazy that windows tries to do all this work behind the scenes so the user doesn't have to worry about it. I do find it crazy how poorly Windows seems to be at running forced system background tasks without absolutely throttling a system, without notice, though.


The problem is it isn't behind the scenes if the machine is totally unusable. This bad behavior during updates and Windows' tendency to get progressively slower to boot even without adding software is why I walked away.


> No other OS does this.

Really? I get the "System restart required" advisory from Ubuntu Server pretty much weekly unless I disable the auto-updater.


I get this on Ubuntu servers too. It just won't force it on you but I assume Windows server doesn't either.


That's because something like the kernel is ready to be upgraded, and this is something Windows will also need a rebooted for.

Also, _weekly?_ That sounds off.


Because the point is not what happens when the user/administrator should reboot, the point is what happens if the user/administrator ignores that advice for a week or two.


You can turn all the auto updates off if you have windows pro and above.

I cant say ubuntu is immune for example, I've booted into a rarely used ubuntu installation on my laptop before and wanted to test something, and apt was locked due to the update manager running in the background.


To be fair, even Windows 7 will not do this (assuming you've configured it to be), it's really Microsoft's insistence on Windows 10 that we have this mess.


Except then you're probably stuck on a version that has exploitable holes UNTIL you upgrade. At least Microsoft is now ensuring people are somewhat protected.


> There's the issue.

No. There's the catalyst for the situation - maybe. But systems do not have to do updates that way. The issue is how Windows does updates and software installation. Other systems prove it can be better, faster, and doesn't have to take over what you're doing.


> But does it actually get in the way of office work, like the author references? Absolutely not, since those machines will be booting daily and get hiccups, not slogs.

It absolutely does get in the way. Less than a month ago the Windows machine at work once again decided that updating itself was more important than allowing me to do my job. I timed it. Windows wasted over 20 minutes of my time. Do I get to send Microsoft a bill?


I usually have a VM for Win stuff so I don't have to reboot my computer. Especially for the use case of 15 minutes of building with MSVC.


Office workstations — agree, not an issue. But what about all those single purpose notebooks in the office that are used to control specific hardware?


Without specifics, it's hard to say. You can disable windows updates and only update when you please. That's probably a better solution for a company anyway. Bring the power back to yourself to decide when to update your device.


> office that are used to control specific hardware?

So they [configuration] are valuable? Then why they are allowed to connect to office WLAN and talk to the whole Internet?


You can mark your internet connections as metered and windows will never use those to download updates.


Presumably this happened because of how rarely the OP boots Windows. It was likely feverishly applying months of security hotpatches to itself in the background, trying to outpace the user clicking on an email attachment.

People who use Windows primarily, and leave it running and idle when they’re not using it (even just asleep is fine, for a laptop), don’t have this problem; Windows does this stuff incrementally, and mostly when you’re not using it.


I think he's mostly ranting, and his conclusions on the death of Windows are conclusions on that ranting -- i.e. nonsense -- however! Saying the reason Windows does this is X, doesn't excuse anything. It's insane that this is how it works.


Not really, the old way was that updating was an event and a task within itself, and your mom doesn't want to sit there waiting for updates to complete to write a letter or browse facebook. Silent, backgrounded incremental updates are the better fit for 99% of users.

Full disclosure: I use linux as my daily driver and don't have a windows installation.


But it's not the old way, at least not for for MacOS (and last time I checked, Linux, which you use?).

Granted, the MacOS version upgrades are indeed _a thing_, and pretty annoying, but that problem still exists even with the incremental upgrades Windows does. There's just no way around certain types of upgrade that touch core components.


> even just asleep is fine, for a laptop

Except that Windows never-ever-ever reliably sleeps. There is always something, some program, some driver, whatever, waking it up from sleep.


Linux isn't better in this regard.


When my devices - mostly ThinkPads of different vintage running Debian plus a 27" iMac running the same - sleep they don't wake up until I want them to so yes, "Linux" is better in this regard. If your device wakes up when it should sleep I'd check if Wake-on-Lan is enabled.


It is now though. I agree that 5 years ago it wasn't ideal, but now the hardware support linux way better than it used to do, and on my dual-boot, i have 1/3rd more battery life with linux when i'm using it.

For the "never really idle" part, i've found a script on github that deactivate and uninstall most of the bloatware windows come with, and it's now way better, so i cannot judge.


I find hibernation to be the best approach. Zero battery drain, my laptop boots in 5 seconds (or negligible time to be concerned about). Just set it up to hibernate when the lid is closed.


The comments in this thread are mostly valid, but are actually skipping over several more fundamental mistakes that the author of the article is making. From the sounds of it, he's multi-booting an entire physical machine into Windows, which is the wrong workflow to use in 2022 as a developer! (He mentions being forced to copy the files out to a USB stick.)

In case it's not obvious (it wasn't obvious for the article author) my recommendations for scenarios like this are:

- Use virtual machines. The benefits here are that you don't have to do horrible things to your disk partitions, you can use snapshots and roll back to them, and you can have many OS versions running in parallel for testing. More importantly, even if catching up with patches took 90 minutes, you can let a virtual machine do that in the background instead of locking up your entire PC as in traditional multi-booting.

- Use cloud-hosted virtual machines. Same benefits as the above, but without taking up disk space locally. Especially useful if you have multiple computers and you like to shift between them. There are generous free tiers available in both AWS and Azure. Both provide up-to-date base images for VMs that take about 2 minutes to deploy instead of 90 minutes. Deploying an image and installing Visual Studio from scratch would take less than 90 minutes!

- Use Azure DevOps Pipelines or GitHub Actions. Both have free tiers and let you run automated builds using Windows-based runner images maintained by Microsoft. They're always up-to-date with patches, Visual Studio updates, and various related tools. This is the superior option for building packages of things where you don't need a fast edit-build-test cycle.


People until WinXP: Windows is constantly vulnerable, M$ should do something!

MS: *invents Windows Update*

People: *disables Windows Update*

People: My Windows has died for another MS Blaster variant! M$ should do something!

MS after Win8: *makes intrusive, hard to disable, forceful Windows Update which... forces the Windows to be updated, duh*

People: My Windows is unusable! M$ should do something!

...

Sure, quality of the how the process was implemented desires better (I'm too in the victim list of the forced reboot after update), but, BUT! the overall quality gone better. With Win10 not from a fresh install it's just a couple of monthly updates from the last month (Quality and Security Update rollups?), not all updates for the last N months you didn't booted it up.

> Linux is my daily driver, but I have Windows on a separate hard drive in my machine.

And if that Windows is really on a hard drive...

For all "BUT MUH LINUX!1111" people out there - try to use a recent Ubuntu/RHEL with a desktop environment on the hard drive. Chances you forgot how slow HDDs are.


Well, uh, quite fast? The last Lubuntu boot faster than my windows, while on an HDD partition, and the windows is installed on the SSD partition. I think that my linux mint was on an HDD too, and while not the "latest" ubuntu, booted faster than my windows too, on a worse computer.

I don't really care about that tbh, its booting fast enough. The only thing that bother me on windows is the telemetry eating my internet when i'm in a train (fuck microsoft, i'm earning a lot so i can afford unlimited data, but when i was an intern i couldn't have afforded this). And on my work laptop, the company security bloatware (Symantec endpoint protection anyone?). Also HyperV. I don't understand why the company forbid WSL and then ask us to run container on our workstation, HyperV is probably the worst virtualizer. I think that's it.


Mine pretty much respect Metered Connection settings and even OneDrive disables sync automatically. Not tried to trace telemetry though.


I have a couple openSUSE/KDE installs on spinners. They're quite fine. Slightly longer to load up and after that it's mostly not noticeable depending on what I'm doing.


6 hours later:

It's quite telling what instead of Ubuntu/RHEL we got Lubuntu, openSUSE on KDE (from which year?) and OpenBSD.

Guys, no. Ubuntu/RHEL, not your favourite lightweight distro.


OpenBSD user here, with an HDD under an Atom netbook, 1GB of RAM. CWM+Links/lynx, Luakit otherwise. Windows 10 would crawl here.


I've been using Windows on a development machine that has seen the last 5 releases of VS installed and then upgraded every time. It was a Windows 10, migrated in place from Windows 8 computer, with a mechanical 5400rpm drive and 16GB RAM.

I complained as much as I could to my manager about its boot times (10 minutes to even have an operational OS), but once it booted, it was ok. I could work all week just by hibernating it and shutting it down only when leaving on Friday.

It ran SSMS along VS and other software like Meld, Teams, Outlook, two browsers, and Sublime text.

It was fine.

I have been using MS OSes since DOS4, and while Linux is my daily driver at home, I think modern Windows is ok. It can be operated instantly by anyone with low computer knowledge. It is kind of reliable.

Posts like this just follow a stupid sentiment. All of it reads like "OMG, Windows is so bad, amirite" and boiling it down it ends up just being virtue signaling to show the world the author uses Linux.


I have a Windows dual boot which I occasionally use. The last time was to stream some sports which requires silver light so no Linux.

I booted in 5 minutes before the start of the game but it was unwatchable due to the fact that Windows updates were running at 100% for the first half an hour


I have this issue on my older laptops with Windows 10 that I rarely use. The culprit is some Windows Store "optimization" process (wsappx). It'll run for days at 100% if you don't boot Windows for long enough. I hear the exact same from friends. It's clearly a blatant bug that's been going on for years now, and Microsoft doesn't give a shit.


Just tell Windows not to install the updates.


I really doubt that Windows will globally die because you experienced boot issues. If you were a non-lazy dev/engineer, you would have parsed your event log to understand the reason it took so much time to boot and you would have fixed the issue. But you rather invested your time in writing a rant on internet. A clue: when you boot your Windows system on rare occasions, at the moment it boots it will update or finish the installation of updates from the previous boot. Which can explain some important boot duration. My take: complain less, investigate more and configure your Windows system for your specific usage.


> If you were a non-lazy dev/engineer, you would have parsed your event log to understand the reason it took so much time to boot and you would have fixed the issue.

Might as well switch to Linux at this point. Why put in that much effort into a proprietary system that you paid for just to make it work reasonably?


If I understand well the guy's need, he only boots this Windows partition for some compatibility software testing... Which means he has to use this Windows. Moreover, being a daily Debian user, let me tell you that bad suprises after patching happen too when using a non-proprietary system... But at one point, whatever the system that you use, investigating the root cause of the issue should be the first taken by a professional, rather than complaining on internet.


Funny how you assume that my Windows system is not configured to my usage.

I've turned off Windows Defender. I've turned off background tasks that I don't want. I've uninstalled every piece of software I don't want. I've turned privacy up as high as I can.

And it's still a slog.


Well, now you get that these were not the adequate steps to fix your issue, you might want to try to configure / use differently the system updates, which in my opinion could be the root cause of your issue ? And why, more importantly, haven't you checked your event log to understand the issue rather than wasting time to complain on Hacker News ?


Seriously?

Anyway, the process that was taking so much CPU was Windows Explorer. I'm pretty sure that is a Microsoft problem, especially since the computer just was booted up.


In my office I do traditional office work by the computer. Windows and it's Office applications is such a pain for your visual and muscle memory. That is one reason for why it being really slow to work with.

The second reason is the sluggishness in each and every application Microsoft ships.

The third reason is the one the author experienced. Really slow and intrusive updating which gets worse the longer you go without rebooting or booting into the os.


Windows won't die because the people making the decisions to buy it aren't doing any productive work and don't know how bad it is.

They just see an established, unbelievably wealthy corp and think "that will do."

There's no way out of this. Whining will just get you labelled as a weirdo by your boss.


> people making the decisions to buy it

Most people just buy a laptop and expect to get Windows on it because that's what they know, unless they know they want macOS and go with a MacBook. Otherwise, if you want a Linux machine, you'd have to hope the manufacturer gives you the option of shipping without Windows, or end up buying it and then formatting. As long as that's the default, Windows won't die, and I would think Microsoft will like to do anything to keep it the default.

When I buy a computer, I want Windows to be opt-in, not opt-out. Ideally, computers should come with a network installer like Raspberry Pi started doing [0], giving you the option at first boot to choose what operating system you want.

[0]: https://www.raspberrypi.com/news/network-install-beta-test-y...


I don't think that is good idea. It limits the options too much. It should only come with absolutely empty unformatted disk. Everything else is too much intrusion. Best case would be even without any UEFI or similar key material in place. So user would have to choose their own.


I logged into Windows 11 this week and suddenly it installed some update automatically (i tried to disable automatic updates), which, rebooted and started a terribly slow firmware update !?

I just wanted to do 15 minutes of work, ended up wasting nearly an hour (meanwhile i had given up and grabbed my Macbook...)

Windows Update is horrible, in my opinion...


Author here.

I love how most of the comments here are saying something to the effect of "he's holding it wrong!" or "this is PEBKAC."

And this especially when people have to make a lot of assumptions about my setup, including if Windows ran an update or not! Because I never said it did in the post.

When I shut down, it did, but the first boot, it did not. The 15 minutes of hang there was not because of an update. (I had happened to boot into it recently enough.) That 15 movie hang is pure bad programming.

It also shouldn't matter if I use spinning rust. Sure it might be slower, but the actual time to login screen wasn't bad, so it is possible to still be usable. I'm not sure the 15 minute hang after boot was caused only by that. HDD may be slower than SSD, but it's not that much slower.

Also, if there are updates that install within 3 minutes, and Windows hangs on "Finishing Updates" for 45, there's still bad programming.

In addition, I have disabled Windows Defender permanently by the Windows Registry, so there should be far less thrashing of my HDD. For those people saying that I don't know what I'm doing in Windows, maybe I don't, but I know enough to do that.

(I don't visit any website in Windows that I don't trust. I usually only visit my self-hosted Gitea instance. I feel safe enough, especially since I could just reinstall and don't keep anything important on there.)

On top of that, none of you have explained why my typing speed is 50% of maximum, at least in Visual Studio.

As for why I don't use a VM: I don't fully trust emulation of Windows from experience, I need to do actual testing, and I used to have Steam on Windows for Kerbal Space Program. (I used to; I uninstalled it before the first boot talked about in the post.)

You're welcome to ask me more questions about my setup and criticize me without assumptions, but please don't assume things about my setup.


>HDD may be slower than SSD, but it's not that much slower.

It's at least something around 100~200x slower to do tons of random reads and writes on an HDD than an SSD. Updates and other background tasks that have been delayed and now going all at once are going to hammer the system drive in a random manner, so this difference in speed will have a massive effect.


This doesn't explain why the slowdown is recent, though.


Your description sounds like every Windows box I've ever fixed whose disk was going bad.


The disk is about one year old and uses less than 5% of the space available.


Unlike SSDs, HDDs don't particularly benefit from leaving significant free space open. The fact you seemingly don't know this speaks volumes.

The core issue is that Windows is thrashing your system drive, which is an HDD, with tons of random reads and writes stemming from built up updates and other tasks Windows wants to do. This thrashing pulls Windows and any other executing programs from that HDD to a screeching halt because the HDD simply can't keep up.

One way to solve this is installing Windows onto an SSD, and in fact Windows nowadays is written with the assumption the system drive is an SSD. This will alleviate coming to a screeching halt when tons of random reads and writes are demanded. Even a simple SATA SSD is fine for this.

Another is to disable Windows Update and other background maintenance tasks you don't find valuable. Your system drive won't get thrashed if there's nothing to thrash the system drive.

If you really purport to be a programmer, please at least understand the environment you are working in and tune it to your needs and desires before going on a pointless rant that might make you look like a fool.


I use HDD's on purpose. If Windows cannot use HDD's, that is bad programming. That's just how it is. Microsoft should not assume that HDD's don't exist.

I use an HDD on Windows for two reasons: one, I am cheap, and two, I am writing a VCS. The Git experience on Windows is already terrible, so I'm setting myself up to have a worse experience on Windows than any of my users. Why? Because if it's painful enough, I'll fix it. This is the same reason I use HDD on Linux: to ensure that my code is as fast as possible for users.

I have disabled Windows Update and other background maintenance tasks. Don't assume I haven't.

If you really purport yourself as good enough to give advice to a programmer, please at least don't assume stuff about that programmer and his environment before going on a pointless tirade that did make you seem like a jerk.

By the way, I scanned my Windows drive with various tools. No errors, and the lifetime number of hours, which I was hinting at with my glib reference to its age and how much space is taken, is only a fraction (1/5) of the average lifetime of drives like this.


>I use HDD's on purpose.

>I use an HDD on Windows for two reasons: one, I am cheap, and two, I am writing a VCS. The Git experience on Windows is already terrible, so I'm setting myself up to have a worse experience on Windows than any of my users. Why? Because if it's painful enough, I'll fix it. This is the same reason I use HDD on Linux: to ensure that my code is as fast as possible for users.

So you're a "programmer" who willingly and knowingly uses hardware with terrible random access performance. Why are you complaining about bad performance stemming from terrible random access performance, then? You get what you pay for. You made your bed, you may now lay in it.

>If Windows cannot use HDD's, that is bad programming. That's just how it is. Microsoft should not assume that HDD's don't exist.

Unjustifiable as it may be for Microsoft to optimize Windows for SSD system drives, a 1TB SSD can be had for ~$100 and a 500GB SSD can be had for ~$50. The age of exorbitantly expensive SSDs is long over, they have become accessible to the commons at large; so it makes some amount of sense for Microsoft to optimize for hardware resources that are widely available.

>I have disabled Windows Update and other background maintenance tasks. Don't assume I haven't.

You clearly haven't, because updates and other background tasks that built up during downtime got in your way. Even Windows 10/11 will still let you disable Windows Update if you have Professional or above.

If you are using Home (because you're "cheap", as you put it yourself), why are you complaining again? You get what you pay for. You made your bed, you may now lay in it.

>If you really purport yourself as good enough to give advice to a programmer, please at least don't assume stuff about that programmer and his environment before going on a pointless tirade that did make you seem like a jerk.

To put it bluntly, you don't come across as someone worth respecting beyond that of common courtesy. You don't know what you're talking about, and you complain about situations that you willingly put yourself in. If you want to earn respect, please act the part.

>By the way, I scanned my Windows drive with various tools. No errors, and the lifetime number of hours, which I was hinting at with my glib reference to its age and how much space is taken, is only a fraction (1/5) of the average lifetime of drives like this.

Anyone who knows about HDDs would quote the drive's SMART data in addition to any tests and other diagnostic data they might have. They would also know SMART and other diagnostic data are only broadly useful guesstimates about a drive's health and whether it might be on its way out.

You don't appear to even know any of that, despite being a "programmer". And you have the gall to say Windows "will die" because you just don't know what you're doing? Give me a break.


> So you're a "programmer" who willingly and knowingly uses hardware with terrible random access performance. Why are you complaining about bad performance stemming from terrible random access performance, then? You get what you pay for. You made your bed, you may now lay in it.

It's still a bed laid by Microsoft. I remember a time when Windows was quick and snappy on worse hardware. I wrote a fiction book in Corel WordPerfect on Windows XP using a machine from before 2000. I never realized computers could be slow during that experience.

And let's do some math, shall we?

The ISO of the current edition of Windows 10, as downloaded by me just now, is 5.5 GB. Let's assume that during an update, Windows replaces all of that 5.5 GB. Let's make this double pessimistic and say that the previous data must also be overwritten with zeroes, so that Windows has to write 11 GB of data to disk over the course of an update.

Let's be triple pessimistic and assume also that my 7200 RPM disk uses 512-byte sectors. (It doesn't; it uses 4096-byte sectors.)

Let's be quadruple pessimistic and assume that it's not just random access, but fully pessimistic access: the disk has to spin one full rotation for every access.

Let's be quintuple pessimistic and assume that there are not contiguous sectors written; every single sector written requires a full rotation.

In this, let's assume that a sector is written at SATA's usual 6.0 Gb/s.

How long does it take to write the 11 GB to disk in those terrible conditions?

The disk takes 8.333 milliseconds to spin to a new spot, and 6.0 Gb/s divided by 4096 (512 times 8 bits) seconds, which is 682.666 nanoseconds to write a sector. There are 21.48 million sectors in 11 GB. 21.48 million times the time it takes to spin to a new sector and write it is 179051 seconds or about 50 hours.

Yikes.

That's the exact worse case. Let's see how long it would take at full tilt (no extra spinning): 14.666 seconds.

What you're saying is that it's not Microsoft's job to get an update as close to 15 seconds as possible; it's my fault that it takes longer than that.

Really? Should I just be grateful that the update does not take 50 hours? I paid for this software!

> Unjustifiable as it may be for Microsoft to optimize Windows for SSD system drives, a 1TB SSD can be had for ~$100 and a 500GB SSD can be had for ~$50. The age of exorbitantly expensive SSDs is long over, they have become accessible to the commons at large; so it makes some amount of sense for Microsoft to optimize for hardware resources that are widely available.

Are HDD's not widely available anymore? That's news to me because I thought they were widely available. Which by your argument means that it makes sense for Microsoft to optimize for it.

> You clearly haven't, because updates and other background tasks that built up during downtime got in your way. Even Windows 10/11 will still let you disable Windows Update if you have Professional or above.

I clicked "Update and Shutdown" when I shut down because I'm not a complete idiot. Don't assume.

> If you are using Home (because you're "cheap", as you put it yourself), why are you complaining again? You get what you pay for. You made your bed, you may now lay in it.

I'm cheap with hardware that I don't stress. I'm not cheap with privacy. I have Professional. Don't assume.

> To put it bluntly, you don't come across as someone worth respecting beyond that of common courtesy. You don't know what you're talking about, and you complain about situations that you willingly put yourself in. If you want to earn respect, please act the part.

Ah, I see.

Well, if I don't deserve respect, can you paste your CV here for me? Otherwise, you don't deserve that respect either because I don't know if you really know what you're talking about with Windows and hard drives. I mean, you're assuming a lot, and you're talking as though you know, but how do I know?

> Anyone who knows about HDDs would quote the drive's SMART data in addition to any tests and other diagnostic data they might have.

I'm not going to quote my disk's SMART data to you! That would be just like putting the full `uname` of my Linux install in my blog post, which I didn't do on purpose!

> They would also know SMART and other diagnostic data are only broadly useful guesstimates about a drive's health and whether it might be on its way out.

I also ran `ntfsfix`. Don't assume.

> You don't appear to even know any of that, despite being a "programmer". And you have the gall to say Windows "will die" because you just don't know what you're doing? Give me a break.

This appearance is based entirely off of things you assume.

I'm okay with appearing like a stupid programmer to you; I can't fight assumptions. I'll spend my time "earning" respect from people who don't make assumptions.


>It's still a bed laid by Microsoft. I remember a time when Windows was quick and snappy on worse hardware. I wrote a fiction book in Corel WordPerfect on Windows XP using a machine from before 2000. I never realized computers could be slow during that experience.

The fact you think Windows XP is "quick and snappy" shows you truly don't know what you're talking about. Windows XP was and still is infamous for being slow and bloated compared to its predecessors.

The only versions of Windows that received universal acclaim were Windows 95, 98, and 2000. All other versions of Windows were either super janky (Windows 1 through 3.11) because of their era, or increasingly slow and bloated (ME, XP, Vista, 7, 8, 8.1, 10, 11, etc.).

>Let's make this double pessimistic and say that the previous data must also be overwritten with zeroes,

Because you're using an HDD, this does not apply unless you go out of your way and tell Windows to perform a full format (aka zero out the drive), and that isn't part of most update procedures.

If you're using an SSD, it still doesn't apply in any practical sense because of how fast drive accesses are.

>In this, let's assume that a sector is written at SATA's usual 6.0 Gb/s.

The fact you would actually propose an HDD will saturate SATA3's bandwidth, even hypothetically, shows you have no idea what you're talking about.

>What you're saying is that it's not Microsoft's job to get an update as close to 15 seconds as possible; it's my fault that it takes longer than that. Really? Should I just be grateful that the update does not take 50 hours? I paid for this software!

Everyone (besides you, I guess) knows Windows is a slow and bloated piece of crap. Therefore, anyone (besides you, I guess) who cares about their Windows computer being fast uses hardware that is completely overpowered for the purpose. Bloated software? Throw more hardware at it. It's a dirty solution, but it works.

Anyone who "cheaps out" on their hardware and expects Windows to be happy, especially in this day and age, is going to have a bad time.

>Are HDD's not widely available anymore? That's news to me because I thought they were widely available. Which by your argument means that it makes sense for Microsoft to optimize for it.

Both HDDs and SSDs are widely available, and there is no reason to not use an SSD for your system drive. Microsoft has an argument when they optimize for hardware that is objectively better and widely available.

If you really want to cheap out but still not deal with HDD's terrible random access performance, grab a 250GB SSD for ~$20 and make that your system drive and use a big HDD for your bulk storage.

>I clicked "Update and Shutdown" when I shut down because I'm not a complete idiot. Don't assume.

Windows Update frequently continues to do things during and after the subsequent bootup following a shutdown to install updates. Anyone who has seen Windows Update do its thing should know this.

>Well, if I don't deserve respect, can you paste your CV here for me? Otherwise, you don't deserve that respect either because I don't know if you really know what you're talking about with Windows and hard drives. I mean, you're assuming a lot, and you're talking as though you know, but how do I know?

I've been using Windows for about 30 years and have messed with my fair share of hardware and software. I'm not going to bother giving you a resume since I'm not employed or looking to be employed by you.

>I'm not going to quote my disk's SMART data to you! That would be just like putting the full `uname` of my Linux install in my blog post, which I didn't do on purpose!

A drive's SMART data contains no personally identifiable information aside from maybe the drive's serial number, and that's easy to redact if it is a legitimate concern. That you don't seem to understand this demonstrates your ignorance.

>I also ran `ntfsfix`. Don't assume.

File system problems are not directly related to physical problems with the HDD. A healthy HDD could have faulty file systems on it, and an HDD on its death bed could have a healthy file system.

A reduction in drive access speed is usually a symptom of impending physical failures in the HDD. Fragmentation in file systems does have an effect on data access times, but it's usually so minor that it's not even a footnote. Windows regularly performs defrags on HDDs by default anyway to help alleviate it further.

>I'm okay with appearing like a stupid programmer to you; I can't fight assumptions. I'll spend my time "earning" respect from people who don't make assumptions.

Whatever floats your boat, but an angry rant with no substance isn't going to reflect positively on you.


I diagnosed the issue with the shutdown.

It was because I kept a USB thumb drive plugged in when I tried to shut down.


Neither fact establishes that the disk is working correctly now, and you have evidence to the contrary.


I have scanned the device. No errors. So now I have evidence contrary to the evidence you assume I have.

Windows is still bad. It should work on HDD's.


I had a Surface book die due to the battery swelling which warped the glued down display. This happened because stupid me left the machine on all the time because I got tired of all the windows updates. Little did I know that Windows had the habit of maxing out the CPU to 100% for some reason. The little fans in the Surface book couldn’t handle the heat. The only reason why people like the PC as far as I can tell is because of Steam.


Sad that so many comments here blame the user for what's obviously really bad programming in Windows Update and Windows Store.


A dev/engineer who spends time to whine on internet about a product because of a specific issue while he has not spent a second to investigate the root cause of the issue needs to take a lesson. What would you think about someone who wrote an article stating that "Ford is gonna die soon because it took 5 minutes to start my car this morning and I have strictly no idea why" ? I guess you would think this person is somehow not very smart...


I'm 3 hours in and my laptop is still spinning 100% on Windows maintenance services. What root causes? It's simply bad OS programming.

Don't gaslight users just because it works on your machine and in your use case.

A dev/engineer who believes 3 hours of self-maintenance on bootup is normal needs to re-evaluate their priorities.


"What root causes? It's simply bad OS programming"

How would you know ? What investigations have you made ? When my Debian has issue when booting after updating my kernel, is it because Debian releases s*** ? No. It is because my setup has a problem I need to find and fix. Stop whining and fix your issue...


What the hell, man. This happens on every Windows 10 device that's left off for long periods of time. Stop blaming users for what's shitty programming/design.


Just booted up my laptop. CompatTelRunner is spinning two processes, one UpdateSoftwareInventory and one DoScheduledTelemetryRun. Meanwhile Edge and OneDrive have updated. I'm 10 minutes in. Spinning 100% CPU. Wuauserv and BITS are peeking in the background while telemetry is hogging everything.


15 minutes in. Wuauserv is stealing the show. Spinning 100%. Meanwhile, Google and Adobe took their turns in updating as well.

50 minutes in. Still going 100%.

90 minutes in. TiWorker -Embedding (servicingstack) is now spinning.

Meanwhile I believe Windows 95 already finished installing from floppy disks on my old Pentium laptop.


3 hours in.

Rebooted for update.

Now mscorsvw's turn to spin 100%.


Sure the sub$500 PCs and laptops will struggle to run Windows fast enough to matter, they run Linux faster anyway. But the $1000 and up Mega Max hardcore PCs will run Windows just fine only problem is that they are gaming machines and corporate doesn't want to spend the money on gaming machines.

Just upgrade to better hardware or run Ubuntu.


From the article:

> And my machine is beefy with 32 GB of RAM and a 16-logical-core CPU.


Also from the article:

>I have Windows on a separate hard drive in my machine.

>First, once Windows booted, which took the expected two minutes,

Emphasis "hard drive" and "two minutes" boot time.

Windows 10/11 is absolutely going to shit the bed if it's asked to run off an HDD because Windows nowadays runs with the assumption it lives on an SSD.

You know what happens when tons of random reads and writes hit a drive? An SSD just shrugs and goes about the requests. An HDD literally screeches to a halt as it frantically thrashes across the platters.

So the only thing that's dying here is his overworked HDD and maybe a bit of his wallet to buy an SSD to replace the HDD with.


For the tldr crowd the root cause is Windows' insistence on automated self maintenance and updates etc. If you keep a mostly stale out of date windows VM around for when you absolutely have to use it and linux is your daily driver, things will constantly break and require extensive self maintenance on the windows front. If the situation was reversed the most you'll get from linux is notification that there are a very large amount of updates available that you can install when you like.

For this reason I don't think the core argument is really sound, because in almost all corporate environments windows is the daily driver and the regular maintenance doesn't come in massive temporally divided chunks of work that make up the large cited overhead in this instance but instead trickles in on a mostly daily basis.

If windows is going to die it's probably just because Microsoft continues to abuse it and try to turn it into a spyware and marketing platform rather than a workhorse and the equilibrium between a hands off linux install and windows slowly tips to favour linux.


> If windows is going to die it's probably just because Microsoft continues to abuse it and try to turn it into a spyware and marketing platform rather than a workhorse and the equilibrium between a hands off linux install and windows slowly tips to favour linux.

They have been doing this for years and intensify their efforts (e.g. by hiding the local user option during install). And I don't see a massive movement towards desktop Linux - Microsoft understands the power of inertia well and they will abuse it as much as they can get away with it.


Absolutely, thus if.

I am no longer under the illusion that any significant fraction of the global population of which the market consists will eventually figure this out and migrate away. There is every possibility that they won't and will remain a captive audience providing a steady revenue stream regardless of how they are exploited.

They do this for every other industry in the world, why not software? Humans simply aren't, at scale, what the mythology would prefer we believe. They probably really are closer to livestock.

One can accept this and work with reality or lie to oneself and deny it and continue to insist otherwise in the face of endless evidence to the contrary. After decades, I have given up and go with the former. Especially in the face of the most egregious examples of livestock like behaviour from the vast majority of humanity in the immediate past.


For this use case would it then be faster to simply download a new VM image each time you needed Windows? That is to say if Microsoft creates images with all the required patched applied?


Microsoft does. Unfortunately the VHDs are fairly big to download, and the ISOs are infrequently updated.

For local development, I can install Windows in under 2 minutes on a PCIe 4.0 NVMe SSD in a virtual machine from an ISO image. Patching it takes longer, about 5 minutes, but I generally do other things while that occurs in the background.

These days I find the most convenient method to get the "latest" fully-patched version of Windows in a VM is to spin one up in a public cloud. Both Azure and AWS have sufficient free tiers to make this perfectly viable for most developers doing testing or whatever. I just tested Azure's deployment time, and it's about 2 minutes to create an up-to-date VM from scratch.


Author here with an update.

The hang on shutdown turned out to be because I kept a USB thumb drive plugged in.


In the author's place I would seriously consider cross-compiling to Windows and test it under wine.


If you have to suffer Windows, disable updates via GPO's and get WSUS Offline Update.


Why doesn't he use a VM on his beefy 32GB ram, 16 logical core cpu.


And how exactly would it solve his problem?


He could boot it an hour before he needs to use it and let it update while he works on the Linux side.


Sounds a lot like pebkac…


I wanted very much to say the above. Doesn't sound like the author put very much effort into identifying what was going on, fails to understand that Windows is doing what it was intended to do for the non-technical users (i.e. updating), fails to understand that there are good reasons for wanting Windows to do that for non-technical users (even if they might not agree with those reasons), and doesn't bother to take any of the commonly known steps to avoid the issue (i.e. either let it update regularly or break the the update mechanism (not recommended)).

So what was the point of the article other than to demonstrate they don't have a clue as to how to properly operate Windows and can't be bothered to invest the time to do so? And if they couldn't be bothered to do so, what are the chances they actually bothered to learn how to operate Linux or any other OS at anything other than a surface level?


What makes a user “non technical”?


A "non technical" user will gift away a brand new with stickers on still Microsoft Windows laptop stuck on wake from sleep and the technical Linux/BSD user will throw it in the trash because life is too short.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: