My assumption is there's someone at Microsoft who gets a bonus as long as they keep presenting update graphs going up and to the right, and they don't care how much they mess up the Windows experience in the process.
My anecdote: I was about to have an important meeting and needed to print off some sheets from my Windows-based netbook. The power was low (<3%), so I plugged in the charger and resumed it from hibernation. It had been offline for a little while up until then because of traveling. Seeing the power was plugged in and I was shutting the machine down going into the meeting, Windows presumed that right then was a perfect moment to perform a Windows update. That night I wiped it and put Ubuntu on there.
FYI the Linux experience has been mostly great. There's been maybe a problem once a year, with some driver or package issue - but most of the time it runs great. I can safely run anything on it overnight knowing that it won't decide to do anything on it's own accord.
The web versions of MS Office tools are actually pretty well build and feel pretty close to their desktop counterparts.
If you have to write papers with a lot of math in them libreoffice's equation editor is way better anyway. It rarely crashes and uses it's own simple markup language instead of the finicky WYSIWYG editor in word.
1. Word - track changes
2. PowerPoint - equation editor
As for the formatting - yeah, MS doesn't seem to follow its own document formatting standards, even for the newer documents. That puts other suites in a hard spot: a) follow standards and present imperfect documents, or b) invest immense resources into bug-for-bug compatibility?
I don't know Excel but I know there are people running entire departments on Excel spreadsheets.
(Having seen many Excel spreadsheets, 80% of the work is done by rudimentary means: pivot tables are Advanced Magic in this regard)
You can edit the markdown in vim which IMO is mentally easier to write prose in because you're not distracted with all the formatting and spelling. You use aspell for spell check and pandoc/pdflatex for typsetting.
It should be able to handle most use cases
1 - https://www.softmaker.com/en/softmaker-office
Does that make sense at all or is this an unrealistic idea? EDIT: Maybe this is how updates already work, I'm not sure
This is what ChromeOS and Android do. At least on Android Pixel phones, there's A and B partitions for each of the boot, system, and vendor partitions. The bootloader tracks whether A or B should be booted. Userspace downloads the updates and writes to the opposite-than-booted partition. On reboot, the bootloader boots the latest partition. Downsize is these partitions are now double the size, so less space for the user partition.
So when Windows says you have an update available, the design is that it did all the work that is possible and all that's left is (in theory) the copy bit that's left.
My completely ill-informed guess is that there are exceptions to this rule by teams that aren't fully aware of how updates work under the hood and as a result upgrading takes longer than it really should. As time goes by those teams get hit with a stick and they fix things, but the upgrade team probably plays a fair bit of wack-a-mole.
The update stage all the files that need to be change and complete the release process on boot before any system processors are allowed to lock anything.
Programs running in memory just stay running, even if you update the files that are those programs. This is how Linux works. Restart them when you need the new version.
Its clean. Its easy. I kept watching netflix as firefox compiled (~30 mins), and got to close the browser after the next episode was done, and open the latest Firefox to monkey with the new settings.
Anyway, ChromeOS swaps the system and kernel between 4 partitions  - automatic update effectively installs a new system into the system partitions you're not booted from, then tells the boot loader to boot from the other ones next reboot.
You are nearly (and generally) correct about programs just stay running even when overwritten but of course programs generally consist of more than one file and might run more than once and if you update part of this and something reloads or whatever, you can end up in a world of hurt. That said, it is behaviour that I have relied on more often than I can count and so have you with your Firefox anecdote. I've done some remote Gentoo updates with systems in quite the broken state - it keeps IT interesting! I once had to get Puppet to install telnetd on some systems so I could repair sshd.
Hats off to people like you. You’re what we mere mortals aspire to, and it’s always encouraging to hear that someone else has attained a level of mastery. Makes it seem more achievable!
Yes, and that's a gross oversimplification. Android got the idea from ChromeOS.
I suspect it's possible to reload the OS in place with kexec, but haven't proposed it yet.
Google uses portage: http://www.chromium.org/chromium-os/packages/portage
It would make more sense for Google to be using Portage to make/install their binary packages. Emerge and quickpkg can both make binary packages: https://wiki.gentoo.org/wiki/Binary_package_guide#Creating_b...
File copy speed itself is rarely the problem. IOPS, where hundreds of thousands of small read and write operations in the registry and other small files dominate the process.
This is, unfortunately, also true of Sonos.
I would say that more than 10% of my attempted sonos usage is blocked by an update process.
If you're just turning on the news for 2-3 minutes while you pack up your things, that's a 100% outage for you.
Granted, I've also never had Windows 10 unexpectedly reboot on me or force an update at an inconvenient time, either. If you're proactive about running updates at times that are convenient for you, and don't just ignore every popup and warning that Windows gives you when it's waiting for a reboot, you'll never see it reboot when you don't want it to.
This'll work for 99% of cases. The remaining 1% is millions and millions of people.
I wonder if the system could give you a few weeks grace time to update, and if you keep ignoring it the system starts slowing down (while explaining why and asking to update). Some people will try to ignore forever and there needs to be some way to motivate them to choose a time (instead of taking control away and risking choosing a very bad time).
I'm used to using a Mac, and the update notifications can definitely be annoying, but Windows' default update behaviour is another level of excruciatingly bad UX. I assume — perhaps wrongly — that you can stop windows from installing updates on shutdown/restart. However, the other night (on Windows), forcing an update on shutdown was the only obvious option available to me. I can't fathom what the rationale is to force this kind of experience on users.
When I want to shut my computer down, it's because I want it to turn off, not start a new (oftentimes looping) operation to install software that I won't immediately see (or need to see) the benefit of.
Throughout Windows’ history, even before I had ever used Macs, it has felt as if someone at Microsoft keeps having the wrong idea about what people adore about Macs, and tries to imitate that in a “we can do it better” way.
Overall my experience on Microsoft store has been sub-par, eg comparing to steam.
Pleasantly surprised so far with Ubuntu, hardly any hacking to start coding and running databases etc.
0. install whichever Windows 10 has the group policy editor - I've got Windows 10 Pro, I think. (I'm at home right now and my work laptop is at the office)
1. run gpedit.msc
2. Go to Local Computer Policy\Computer Configuration\Administrative Templates\Windows Components\Windows Update
3. Double-click Configure Automatic Updates
4. Set up as follows:
- Overall group policy setting = Enabled
- Configure automatic updating = 3 - Auto download and notify for install
- Install during automatic maintenance = unticked
- Scheduled install day = 0 - every day
- Scheduled install time = 03:00
- "If you have blah blah" = Every week
- Install updates for other Microsoft products = unticked
Three things to note:
- this works for me. Ask yourself: are you me? But even if you aren't, don't let that put you off, because you might still get lucky
- there's a popup that pops up to say something like "updates are ready, click to install" - and this actually means literally precisely what it says, nothing more and nothing less, so for god's sake don't do what I did at first and click it thinking that it means you'll get the option of saying no afterwards and then wonder why all the updates are installing. Go to the updates section of the control panel by hand to see what updates are ready before approving them
- the above settings are just what I have set on my work laptop, copied out from the dialog box, based on a screen grab. Some of the settings may be the defaults. I don't remember which
It’s one of the most anti-user policies Microsoft has pulled, and they have a long list..
MS know how their bread is buttered. Who there would dare shut this down? The program managers know that if they break this stuff, there'll be a legion of angry IT guys clubbing together to pay to have them killed.
I must be getting old and / or wealthier because now I just use a
In a few more years, I have a feeling I'll be using neither Apple nor Microsoft.
I do think the market will provide something, just depends on who figures out there's a lot of us out there who have money to spend.
On my Win10 Pro machine, I've switched to the semi-annual channel and disabled automatic installations via GPE.
Does windows not give you the 250 warnings that it used to to postpone the restart? I turn off my computer every night, so I haven't even noticed that updates occur for ~1.5 years.
As an industry we need to get better. A software update should mean things like "improved performance" and "better security", not "totally different UI" and "20 features nobody wants". If users really want the new shiny, it should be optional.
...or reduce performance on otherwise perfectly good hardware, possibly due to the those changes or the "20 features nobody wants".
I'm a computer professional, and even I have this anti-update attitude for exactly that reason. At least I have the competence to mitigate the security risks through other means, but I can't well recommend anyone else do the same. To quote Tom Lehrer, this makes me feel "like a Christian Scientist with appendicitis".
They haven't done that for 'decades.' It started when they realized they literally couldn't give Windows 10 away.
All that is needed to make the system usable is the ability to turn off all automatic restarts; tell me to do it, even nag after a while, just don't ever do it automatically.
And worse: I'm all for auto updates...unless they break the system! (See article.) No wonder people are holding off upgrades. (Not Windows-specific, for sure - but Win10 seems to be the worst offender)
I'm still on Windows 7 and my laptop is Windows 8, where I can safely disable updates until I'm good and ready.
My understanding of Windows 10 was that there is a way to control when updates happen if using the Pro version? Is that not the case?
The forced updates, even for advanced users, is the single reason I refuse to update to Windows 10.
I do agree though that there should be a simple option, "do not restart for updates until I click on a prompt". I'm on the same page, but I'm not convinced the situation is as dire as everyone makes it out to be.
I'm aware it's a security risk, but I also keep nothing of value on that machine nowadays.
Without that, computer might go to sleep, which would happen on any other OS too under default settings.
If I ever wanted to be sure that no updates took place I'd use the freeze.
Switching OS might be a solution, but it is worth having a look at these first. The upside is never being behind on patches, and there are actually some bloody useful fixes coming down the pipeline especially as far as security is concerned.
See, turning on a computer is an explicit declaration that you want to use it. Windows updates ignores this declaration by taking that time to itself. It doesn't matter if I can defer the update by how many days if one day it will stop/delay me from using my computer when I want to. Almost no one outside of IT Administrators turns on computers solely for updates and maintenance.
Contrast that with Linux (Ubuntu distro in mind) where updates happen when the computer is shutting down. In contrast with turning your computer on, shutting down is a declaration that you no longer need your computer. That's the OS' cue to chime in and perform some routine maintenance.
I wonder how much of this is an architectural problem. The first time I remember being frustrated with this Windows behavior was in Vista. Someone correct me if I'm wrong but I seem to recall that 95, 98, and XP used to do updates on shut down. I know Vista isn't fondly remembered but, gee, did they deliberately make it so that it updates when you start your computer? Why couldn't a company who takes pride in backwards compatibility keep the old update behavior?
To wit, people make a big deal about how immediately can you use your computer (i.e., start-up times) but all your fancy-schmancy SSD + hyperthreading set-up is useless if Windows decides to update just before that one presentation you've spent weeks on. Updating on shut down is the user-friendly way.
iOS made them pretty seamless and second-nature, and I can’t recall macOS ever ruining my day because of forced updates.
Maybe even 98 didn't have automatic updates, but I won't be surprised if 98SE already had this capability in some form.
For whatever reason it hasn't impacted me, and that's also having a laptop that only gets irregular use as well as my regular desktop, but I'll be more observant about this going forward.
Updates may have an "on startup" component to them. If you shut down, currently you are forced to install updates on the shutdown - this happens silently, without permission. But: It will just do the shutdown side - so that the next time you start, you're stuck there looking at the "startup part" where updates continue to install.
MS ought to fix this, either by changing how updates are installed in two phases, or to incorporate an automatic "shutdown/install"=>"start windows, run updates"=>"shut down" option.
Also they should let you just shut down the PC without having to install updates, especially if you don't have time to do so.
Currently, as a user, the only really safe thing to do would be to "update and restart" followed by a shutdown.
With a little creativity, one can redecorate routine maintenance into a parade of glorious futurism. And of course that parade needs to be the first you experience on the computers return to conciousness.
Something similar happens on Surface updates.
On the projects I've worked on, most people never open the preferences or settings menu.
The worst for updates, by far is Windows 10 and server 2016. I'd better qualify that. Gentoo can take a really long time to crunch through say 1GB of source code if not updated for a few months. Firefox can take an hour or so on its own and LibreOffice is a whopper as well. Older Windows versions, say 2008 R2 can take a good hour some months. NetWare updates need a skilled surgeon (or someone who can read a manual). pfSense (FreeBSD for me) upgrades can be a bit fraught but generally not. However all of those will get you there eventually.
W10 and 2016 are different beasts altogether. They do a pretty decent job of hiding the nitty gritty away from you and I actually like the auto "just do it already" approach and can set "active hours" to tailor the experience. However, it takes bloody ages to run and often needs more than one reboot - sometimes three (ie update, reboot, something doesn't work, reboot, still doesn't work, reboot, now it works).
The error messages you get are awful and generally need forensic analysis elsewhere - a naiive Google will get you to some MS forum and a suggestion to reinstall Windows.
Switching OS is not really a solution but whining about the update experience within potential hearing (HN) of those that might be able to do something about it is a good idea 8)
My other pet hate on Windows that's been out of control since 7, and getting worse with each generation, is the insane hidden wastage under /Windows that simply constantly grows.
After it's a year old, between winsxs and the various installer and update folders you can easily burn 50-100GB of space under ./Windows for an OS that took say 3GB to install. You can clean up only some of it.
No other OS gets anywhere near this idiocy.
Regarding switching to Mac, it never ceases to amaze me all of my colleagues who opted for a company Mac when all the productivity tools they need run natively on Windows, and sometimes not on Mac at all. So they set up a Windows partition which is never big enough and they always have problems with. Their choice was purely out of fan-boyism, and they all regret it.
 https://www.codeweavers.com/products/crossover-mac (don’t mind the old style website.)
That being said, I personally haven't had problems with Windows 10 except for the very first months. I'm quite happy how stable it is and how at the same time they frequently push out new, useful features (the new console, ssh, nicer gui, immersive search, package management, ...). Microsoft, please don't mess it up and force me to go LTSB...
No Store, no Cortana, no Edge, Windows 7 calculator, no Candy Crush, no bullshit. Don't miss out. ;-)
Problem is... 75% of Microsoft’s customers will be satisfied with Office 365 - the other 25% are utterly allergic to the notion of losing sovereignty over their data, but knowing Microsoft they’ll optimise for that 75% and eventually stop servicing the remaining 25% who desire perpetual licenses and on-prem storage, presumably they’ll be happy with OpenOffice thus giving MS’ Office org an excuse to ditch that market segment entirely.
...and that would trigger Microsoft’s aping of Apple’s similar departure from their customer base, which means they lose their halo effect.
Get the ISO from here (it's a magnet link, load it using a BitTorrent client):
slmgr /ipk M7XTQ-FN8P6-TTKYV-9D4CC-J462D
slmgr /skms kms.digiboy.ir
I don't want Microsoft siphoning off data from my PC. I don't want Cortana. I don't want ANY FUCKING THING except for a system that is minimal and it works.
Why doesn't Microsoft make a developer friendly version NOT targeted for consumers (aka without all the bloat)!?
Edit from the other post:
Don’t forget the registry system, poor DPI scaling, UI piece patch going back to Windows XP (mouse pointer speed dialog box), BSOD if a third party hardware segfaults, fuck you Windows Explorer, everything freezes until network access time out has expired, control panel nightmare, Metro UI with giant lists(ever tried changing default program for say .chm file extension!?), why can’t I force quit without having to go to Task Manager?, etc.
Windows is a poorly made piece of software rotten from the core. If you look deeper, you discover the horror of how things work and fragments from Steve Ballmer days.
You can try to make an argument that it’s not rotten to the core, but no way in hell anyone would say “Windows is beautiful” in the way one says Unix is beautiful.
There are always problems. Bluetooth breaks often, the MFC device on upgrades, enterprise wifi is iffy (and now it's only enterprise wifi but up until very recently, it was 5GHz wifi) ...
I am now on Windows 10 w/ Linux subsystem for Windows. The big upgrades are put off with Group Policy, the small updates run overnight because I leave the laptop docked and powered when I am sleeping. But reboots are rare. The big upgrades I wait 4-5 months with installing. This is a solved problem. O&O Shutup10 deals with the privacy issues, I think (also I think this is overblown, the privacy fight is lost already).
So why suffer needlessly with the Linux kernel?
Linux used to be terrible on laptops, but I recently bought a laptop, installed Linux (Fedora with KDE) on it, and everything just worked out of the box.
You're probabaly waiting for the "except..." - but there is no "except". Everything works the same or better. I get hours more battery life than windows after installing powerTOP, and about the same before doing that.
I ran Arch before I tried Fedora, and from that experience I found that the only way to get Enterprise WiFi is by using network-manager and nm-applet (both of which come with KDE on Fedora so I had no issues when I switched).
I'd rather get a cheap piece of shit than an expensive piece of shit.
Most engineering softwares are only and exclusively available on Windows. I’m a Mechanical Engineer.
Linux, I use for writing code with open source tools (gcc, clang).
Mac, I use for general purpose computing and media consumption.
From there you can turn off or on any feature or function.
I specifically use it to disable all windows defender applications and firewalls (which also automatically kills windows store and updates unless turned back on)
With something called execTI. It lets you modify the registry as a TrustedInstaller so you can turn off any services running aka the 5 or 6 different windows anti malware, defender, firewall, and security services.
If you want more instructions I will help because it is extremely annoying and every update does turn everything back on for the most part. So I do this cat and mouse game about every 2 or 3 months.
My laptop usage on 16GB of ram goes from 25% idle to just under 10% idle. which is my entire point of doing it so I can save battery life that I don't really have.
Out of curiosity, does it run under Wine or it has some DRM that prevents it?
I used Linux on the desktop for over a decade, usually with a highly customized FVWM2 config. I found windows 8 to be okay, and windows 10 to be a pretty good minimal desktop.
All my dev work is done remote on PuTTY anyways, so beyond a browser and terminal, other features are really the deciding factor.
Yeah, that's a big one. There is no other C/C++ IDE out there that can measure up to Visual Studio. Either you use vim or emacs with a hundred cobbled together plugins to sort-of approximate a productive development environment (and you still don't have a usable debugger); or install one of KDevelop or Code::Blocks, which are fine for toying around with toy/academic projects of less than 100kloc; or you shell out lots of money for one of Qt Creator or JetBrains' CLion and get a product that is almost as good as Visual Studio 5. I'd say if you want to actually write software, just install Windows in a VM and do a full restore every so often. It's a bit of a pain, but you get the absolute best C/C++ IDE ever created and can enjoy unrivaled productivity and unmatched performance.
Have you had any luck running it with Wine?
Edit: I realize you may have been talking about HotS, not OW, but according to Google that should also work if you configure esync: https://github.com/lutris/lutris/wiki/How-to:-Esync
If you have a laptop with switchable graphics, you will need to use nvidia-xrun to enable Vulkan passthrough to the dGPU, but aside from that I just followed these instructions: https://github.com/lutris/lutris/wiki/Game:-Overwatch
Note that the Blizzard app will fail to launch with DXVK enabled, so you have to disable DXVK, launch Battle.net, enable DXVK, and then click the 'Play' button to start Overwatch. I made a few scripts for this, so I just do `./disableDxvk ... ./startBlizzard ... ./enableDxvk ... click play` to run the game.
WINEPREFIX=~/.overwatch_wine/ setup_dxvk64 # enable DXVK
WINEPREFIX=~/.overwatch_wine/ winecfg # disable DXVK: delete the overrides from Libraries tab
exo-open Battle.net.desktop # launch the Blizzard app, the .desktop should be generated on install
Yep, in the front.
Stuff works. You don't need 32GB of memory to run anything. The system doesn't get locked down all the time while the OS uses the entire IO capacity. The computer does not misbehave all the time and instead does what you order. You get software for almost everything right from the package manager, and it works for you, instead of shoving ads or licensing constraints into your face.
I see Windows using the entire IO capacity of my work computer every working day, 1/3 to 1/2 the time.
Windows update destroying your configuration isn't even news anymore. It restarting when people don't want is a widely known fact, and more than 1/4 of the times I have a half-hour meeting at work the computer stays there updating for 1 hour or more.
Windows software installation is a worst in class experience on nearly any dimension. It's not something that happens once in a while.
I don't know what you call "cherry-picking". Those "ultimate worst experiences" happen every day.
Out of curiosity: how does Microsoft stand in the way of developing with GCC and Clang?
Recently I heard about Windows 10 Pro for Workstations, which allegedly doesn't include all the bloated crap and lets you disable most annoyances. But I haven't been able to definitively confirm this is the case. I read somewhere you can upgrade from Windows 10 Pro to Windows 10 Pro for Workstation, but it's quite ridiculous to pay them even more money just to remove / disable some features.
I've run a few programs that let you disable most of the things I don't use or care for... But you have to re-run them each major upgrade cycle, because Microsoft usually re-enables a bunch of crap after upgrades.
At this point I've pretty much accepted that Windows 10 is diametrically opposed to my interests, so I just try to avoid it as much as possible. I dual-boot Linux and Windows 10, and I only use it to play games which I can't easily run on Linux.
To anyone that feels like they have to comment because they're happy with Windows 10 and all those features: great for you! You can stick with something that you like and works for you. I'm not trying to force my preference down other people's throat.
What [well-known] OS doesn't do that? (Except for specific sandboxes for particular driver classes that every OS now has, like GPU drivers, or drivers for non-DMA USB peripherals. Windows has the most of these sandboxed driver classes, AFAIK.)
Of course, if your standard of comparison is something less well-known like https://genode.org, then few OSes are going to meet your expectations. ;)
> Why doesn't Microsoft make a developer friendly version NOT targeted for consumers (aka without all the bloat)!?
Maybe you have a different interpretation of "developer" than I do here (maybe POSIX development?), but to develop for Windows APIs on Windows, and test your code, those APIs have to actually be around to call, and have to actually do something. You can't update some old codebase that uses legacy Windows features to use newer Windows features, without your development box supporting both the legacy and the new Windows features.
> why can’t I force quit without having to go to Task Manager?
Explorer allows arbitrary shell extensions to embed themselves into its memory space, so it can't be trusted to run as root. An Explorer-launched "Force Quit" tool, that doesn't actually force most things to quit (because they're running as a different user), would be pretty useless/annoying/surprising.
I don't trust torrent sites or magnet links sorry not sorry
for all M$ needs and untouched or ISO directly from Media Creation Tool
(there is a modified version of MCT that pulls an enterprise ISO if needed)
I'm on the same boat yet somehow 8.1 still felt better...
I rarely get upset at computers these days, but this is one of the things that easily boils my blood. It happens with Windows and macOS machines at the most inopportune times. I'll be typing and hit enter, seeing a mysterious dialog appear and vanish in a flash as it stealthily hijacked my "Enter" stroke. "Great, who knows what I just accepted or declined."
It seems like a very basic UI consideration to discard any keystroke if the dialog spawned less than 1 to 3 seconds ago. Make it a global rule (Or better yet, a user setting) for all UI dialogs in the OS that is impossible for developers to override.
Linux is not a window manager.
I just don’t understand; I cannot fathom what implementation could be so complex that they can’t achieve full test coverage on something like whether or not a damned window appears when required.
It does seem that complicated from my observations. Notice that it's not a regular window, but part of the overriding/overlayed renderer (DWM? something else?) that draws stuff on top with grayscale smoothing (Direct2D? DirectWrite? something else?)... which probably involves lots of communication between several processes. It feels like an incredibly heavyweight UI component, just judging from the lag between when the icon is clicked and when the list of networks pops up.
Disabling cortana, xbox, candy whatever, news, telemetry, etc., setting windows not to maximize when dragged... is no fun
Disk IO is a huge bottleneck in the setup process.
The only feedback in the entire update process is a little spinning circle in a list of updates, which spins for hours on end without apparently doing anything until Windows suddenly out of the blue is “ready” to install and reboots. It will appear to get somewhere, reboot again, and somehow return to the desktop as if everything is OK, yet the update list claims that the update failed. Then it spins again.
I mean, even if it’s “good” that it doesn’t auto-install broken updates, it sure seems to be spending the absolute maximum amount of resources in terms of downloading data, consuming energy, wasting my time, etc. I suspect the size of the update is the factor they should most easily be able to control; patch something smaller and maybe it will actually finish. I don’t know, just seems like an utter mess to me.
Things to check:
Are you UEFI booting an MBR disk? It works, but Windows Setup will fail in future upgrades.
Do you have all the correct and required partitions? If you don't have the right partitions (OEM, EFI) of the right size, Windows Setup will again fail with cryptic errors, none of which say "your partitions are wrong".
Worth pointing out you don't have to be sharing the drive with another OS to get this situation. I had a system which I upgraded from 7 -> 8.1 -> 10. My linux install and bootloader are on a totally separate drive. This left the drive with:
* 100mb system reserved partition (unused, created by win 7 installer)
* 250gb C: partition
* 400mb system reserved partition (created by 8.1 or 10 update, actually used)
Then I used the Samsung tool to copy to my new ssd and ended up with a 100mb system reserved partition and a 1tb c: drive. But the 1709 update tried to unpack the >100mb update into the too small system reserved partition and failed in a loop
Gradually, with much trial and error, I managed to work through each problem, until I wound up with a mostly correct GPT disk and a fully functioning system. It's now successfully applied several updates correctly, where previously it would get to 99%, fail, then roll back, only to try again the next day. As a side effect, I now know more about Windows 10 booting, the BCD, and GPT than I ever wanted to.
The only "problem" I have now is an unused old boot partition that I'm frankly terrified to try to remove, since it'll renumber all of the partitions and cause some problems. I'm relatively sure I know how to fix it, but it's too risky to contemplate right now.
Client of mine had an old copy of the Novell client installer (it was never installed on this machine) sitting in a folder named C:\Old_Computer\ . Last year's fall update kept aborting with an unclear error dialog. The install log only gave a vague suggestion that the issues was unsupported software and DID NOT reveal the full path to the offending file, just the title of the software.
Windows desperately needs a shibboleet option, so power users can stop fighting it.
I was able to get the upgrade to finally work after I uninstalled the DisplayLink driver and a Splashtop app (for using an Android device as an additional monitor) and then it finally worked. I don't know which of the two (or both) were the exact culprit, but it was driving me bonkers trying to get that update to work.
By the way, later that month I learned that I had been lucky as some people ended up being stuck within the update.
I don't doubt you at all. This laptop took up took so long to boot with Windows that I'd take the time to make coffee.
When I was _forced_ into the windows 10 update, it went through the process and appeared to finish but didn't put my desktop back. No problem I figured, they put it somewhere.
So I did a file search, found the desktop in a folder, moved it back to the desktop.
A day later it self-restarted and _completed the update_, replacing the desktop with the now empty desktop folder.
I went to the Microsoft store to get them to do a file recovery and they had the _gall_ to tell me it would be $250 plus 7 days and they wouldn't guarantee recovery.
I moved to Mac this year after being exclusively on Windows since '98.
I hope Microsoft dies andisgraceful death.
Regardless of OS it's always a good idea to have backups of your important data.
From the articles you linked:
"The images get corrupted on copies to the USB attached external drive. "
"However, as Bombich notes, ordinary APFS volumes like SSD startup disks are not affected by the problem described above, so the vast majority of users won't be affected by it – the flaw is most applicable when making backups to network volumes. "
I am not sure you have read the original article and the two articles you linked, otherwise it would have been obvious that these are not the same by volume and severity.
If users don't have backups, any unintended data loss caused by an operating system bug is bad because it can be difficult or impossible to recover the affected data. Would you not agree?
You may not have been affected by the two bugs I described above, but some people would have been (more so with the image corruption bug).
> the described bug has not happened as a result of an upgrade, it was not forced on any user and definitely was not affecting the whole desktop.
If you go back further in time, there actually was a bug that resulted in data loss during an OS X upgrade.
Oxford Semiconductor has issued a statement with regard to the emerging Panther and FireWire data-loss debacle.
The company says: "Oxford Semiconductor has been investigating reports that some FireWire 800 drives have lost data after an upgrade to the Mac OS X 10.3 Panther operating system is installed (released late October).
"Currently we believe this issue relates to a change in the way Panther uses FireWire that affected version 1.02 of the OXUF922 driver software. A new version, 1.05 was issued by Oxford Semiconductor to the manufacturers of external drive products in early September."
As Macworld UK first reported yesterday, users installing Panther while having an external FireWire drive connected to their Mac have seen data loss; similarly, users with FireWire drives connected to their systems have seen data loss once they reboot Panther. At this stage, it appears that the problem is confined to FireWire 800 drives.
In 2001 there was also a bug in iTunes 2 that caused an entire hard drive partition to be deleted if the volume label was prefixed with a space.
Some Macintosh users who rushed to download the latest version of iTunes – Apple's popular digital-music player – were singing a song of woe on Friday. A bug in the installation procedure caused the application to completely delete their computers' hard drives.
The bug seems to have affected computers with a very specific configuration: people running Mac OS X who had "partitioned" big hard drives into several smaller ones, and who'd typed a space at the beginning of the drive name.
For example, if a Mac had a drive named " music" instead of "music," it might have been deleted by iTunes.
Tom Fisher, a computer repair technician who lost about 100 gigabytes of information during the installation, said that people often include a space in the drive name to ensure it shows up at the top of the list when they examine their drives.
According to Mac experts who examined the code of the buggy iTunes installer, the problem arose from a very tiny programming mistake – a forgotten quote mark.
Sometimes it's recoverable from blindly pasting god-knows-what from stackoverflow. Most of the time I just have to do a complete reinstall.
(I guess it's not a complete brick on update because I have never suffered data loss, but it's similarly infuriating)
It's also nowhere near as bad as actually loosing data, as you mentioned.
I have also been using linux for over a decade and I still can't recommend it to any non-developers friends unless their hardware build is extremely common.
I am currently running two nvidia gtx 1080 SLI and multiple 4k monitors. Even for just installing the OS, I need to physically remove both graphic cards and use a smaller monitor first. This physically laborious workaround just to reinstall the damn thing is what amplified my rage when it decided to break on autoupdate.
Funny, I use a Macbook pro at work, and windows at home. I'm not the kind of person that has strong preferences on tech, I use stuff until it stops working well and then I switch.
My several (over time) work macbooks have been really bad. OSX has so many bugs, many of them seem really serious potential security issues (graphical corruption across processes, login screen flickering to the desktop upon waking from sleep, ...). Lots of other bugs are just annoying and make things janky to use. It also does weird things that make me fear it is a fire hazard (e.g. battery draining within 24 hours while the lid is closed in my laptop bag).
On the other hand, a windows small form factor connected to my TV, and a surface pro have been basically hassle free. Obviously windows has bugs too, but none of the ones I've seen make me question security the way I do on OSX.
OSX has its share of dirty bugs but nothing out of the norm. You could entrust me your surface pro, and I’ll make it crash randomly with only legit pro software.
Next thing I knew I was on throttled internet for 3 weeks. Setting all the machines so that they don't update turned out to not be too hard, but was FAR harder than it needed to be. I'll decide when my equipment updates.
Delivery Optimization: https://www.howtogeek.com/224981/how-to-stop-windows-10-from...
And there's ways to set the Ethernet connection as metered if you search a little bit online.
One study says nearly 7% of all internet users in the US use a satellite connection. These are slow, and almost always metered. That is a lot of people getting surprise hammering of their connection.
You can cache iOS device updates with this mechanism as well. I’ve never bothered as I’m fortunate not to worry about metered bandwidth, but it seems like a nice feature if it works as advertised.
idea that no sane user could possibly have anything more critical in their life than installing whatever code comes out
The problem is when you examine user behavior, they inexplicably seem to have something more critical than installing updates, 24/7/365. Which is how bot farms begin.
I'm not sure users have been finding the alternative that has been playing out any more acceptable. Does it look that way to you?
Now, do you have the right to own and drive this autonomous car around on public roads, if you’ve modified the car to be an “open server” where anyone can anonymously connect to it from anywhere on the Internet and drive it around?
And, if not, then what’s the difference between that modification, and knowingly driving the car when it has an unpatched vulnerability allowing people to do the same?
And if you find that there is no difference, then what’s the difference between a vulnerable car that can DDoS physical infrastructure, and a vulnerable PC that can DDoS virtual infrastructure?
MS could deliver security updates separately to feature changes but chooses not to.
The Tragedy of the Commons is that well-publicised incidents like this (and the trend of updates to consumer software, supposedly under the guise of enhancing security, to bring about significant changes in appearance and behaviour) make people less, not more, inclined to defer updates to all software with the result that developers feel the urge to strong-arm users into updating.
The only justifiable reason for updates to be forced in the example with the vehicle is the physical danger that could otherwise result, and that simply doesn't exist in the example with the home computer. To my mind, the line of thinking you are engaging in here is a perfect example of the rampant authoritarianism that seems to be so rife in the computer security community these days.
How about if there not only was no FDA approval process for the vaccines, but the pharma company itself didn't bother testing them?
Because that's Windows Update in a nutshell. Every couple of months somebody breaks into my house in the middle of the night, even though I locked my doors and windows and posted a no-trespassing sign, and pokes me with a needle... and I'm supposed to just sit there and take it in the name of "security."
Imagine what the FDA already does, and then imagine that they were verifying a drug that would be given to every person in the country. There'd be a crazy strenuous verification system for that.
And that's a big part of the problem. Not only is there no 'FDA' to test these patches -- nor should there be -- but the manufacturer evidently doesn't test them either. They fired their QA personnel a few years ago, so that's now our (unpaid) job as users.
Even worse, there are some indications that this particular bug was discovered and reported by insider program members and actively ignored by the company.
Windows Update is apparently the team where Microsoft employs their B- and C-level players and managers. That's not OK. If you're going to insert yourself forcefully into everyone's critical path, you'd better know what you're doing.
Besides, if you look at what sorts of vulnerabilities they're actually patching, the majority of them require local access anyway; remotely-exploitable-by-default ones (fortunately) tend to be few and far between.
They mover force an update but make you aware you're on an older version.
I'm much happier in a world where only those that put in the effort to research how to block updates and then go through the steps can do so. They are much more likely to encounter info on how they should really think about what they are doing, and whether there are alternatives or partial solutions that achieve most their needs without being as extreme.
Please don't ruin the little bit of herd immunity we've built up.
At any rate, iCloud is not “backup” in the traditional sense, it’s really “cloud sync” for user data. In that context it’s not really surprising that connecting a new device to iCloud and restoring pulls the newer OS version, given you never really backed up the OS data at any point.