As a tech worker, I don't want to live in a world where tech workers are fired because they make a small, fixable mistake.
> Did Microsoft offer a way to undo an unwanted install of Windows 10?
Yes. You can downgrade Windows 10 to whatever you installed it on top of. I think this has been possible since the Vista -> 7 upgrade.
It's been possible with every version of Windows, I believe. Windows 95 would rename your old Windows directory. And earlier versions didn't try to control the hard drive, as they weren't your OS - though 95 didn't try to own it fully, it lacked a My Documents directory and expected you to make your own root directory folders to store file!
Wasn't it possible to store user files in C:/Windows/Profiles/{username}/ , too ?
OSR2 did indeed do that, but it was a later addition. And Windows NT 4 never got My Documents.
> Wasn't it possible to store user files in C:/Windows/Profiles/{username}/ , too ?
I think so, but I don't know if the OS really especially encouraged you to store things there.
This doesn't work in a huge number of situations. It just reports failure.
What's a "huge number"? I'd say only a tiny minority of Windows users even try to downgrade. I tried to Google this ("windows downgrade failed") and couldn't find many examples.
Unless you work at Microsoft and have real data from users, I don't think you (or I) have the data to claim that few or many of these downgrades fail. There's just no way to really know.
The quality of work coming out of the Windows Update group has been seriously problematic lately. It's well past time to fire some people.
On Debian or CentOS, updates are simple - if any updates are available, you install them. End of story.
On Windows, as I have learnt the hard way, simply installing all available updates is pretty much begging for trouble. During my training at Deutsche Telekom, when I heard that they had a whole team whose only task was testing Windows updates, I thought they were being wussies. Now, I envy them. The reality for me is more like checking online discussion boards to see if anyone was dumb enough to just install all the updates, yet smart enough to figure out which ones caused them trouble, then only approving those updates nobody has complained about. It is kind of like playing Russian roulette.
As as Unix-person looking at the Windows world from an outsider's perspective (kind of), I am kind of shocked that this is the OS powering approximately 90% of corporate desktops. Imagine if, following the end of the Cold War, Western car companies had been driven out of business by the Trabbi. That is pretty much what has happened in the PC world.
And, yes, I get that Windows is tortured by more combinations of hardware and software than anyone wants to imagine, and QA for their updates must be a living hell. But IIRC, Microsoft also made something like 20 billion dollars in profit last year, so they could easily spend one billion on making Windows update not suck - or at least suck slightly less - and still be disgustingly profitable.
The worst thing is that people who have known no other OS but Windows all their lives fail to see how utterly wrong this is. Either they just don't know any better, or it's a massive case of Stockholm syndrome...
(Sorry, but I've had that on my mind for a while, and since it was patchday this week, I hope you can understand my need to vent a little...)
I've done this since Windows 95. It hasn't caused me a problem since Windows XP SP2. I actually love Windows update and missed it when I was on Linux. Ubuntu's kernel upgrades would randomly break sleep, audio, special keys on my keyboard, or other weird stuff.
There is a difference between administrating a single machine at home, or maybe a handful, or taking care of a company network of (in my case) about 80 machines (and that is still a tiny network, I suppose). Problems that a typical user only sees once in a couple of years become disturbingly common as the number of machines grows, and I guess that is true regardless of the operating system being used.
Having said that, I have encountered some rather strange and frustrating problems with Windows updates. One time, a Windows update would somehow not successfully register itself as being installed, but it did require a reboot. So after installing the update, the user rebooted her system, finding yet another update (the very same, in fact) available, install it, reboot, find another update... For some reason, only two machines in our network were affected by this problem, and I have no clue why it was these specific machines. One time, an update to Office 2013 overwrote some VB script that Autodesk Inventor happened to depend on, causing it to crash while displaying the splash screen; reinstalling Inventor hepled with that, but now Outlook would crash when trying to write an email. Another time, an update caused a few machines to hang while booting. There was an update to Windows 8.1 that was not called a Service Pack, but for all intents and purposes, it was, and Microsoft said that without this update installed, 8.1 machines would no longer receive updates; however, this update caused 8.1 in combination with a specific version of WSUS to no longer talk to the WSUS server - I think this is what people call a Catch 22.
And the list goes on and on. I have only a few years of experience taking care of a relatively tiny number of systems. Like I said, I understand now why large companies run labs where they test Windows updates - and updates by other software vendors, too, I suppose - against each and every combination of hardware and software being used on the corporate network.
Just because you did not run into problems does not mean there are none, just as my experience with updates on Linux and FreeBSD is not representative of what it is like to run dozens or hundreds of machines with these operating systems. However, currently the operating being used on the majority of company networks is Windows, and the company making billions of that is Microsoft.
Although, to be fair, in many cases where an update to Windows breaks a third party application, I suspect the third party is at least as much to blame. In my case, I have grown very weary of Siemens automation software (Simatic, WinCC, TIA), which is very sensitive with regards to updates, to say the least. So, to a degree, Microsoft has merely become a victim of its own success.
I'm not sure that firing people is the right approach but this is not acceptable.
The end result of all this stupidity is to incentivize users to turn off automatic updates. Everybody with an Internet connection will suffer for that.
Hanlon's Razor. Which do you consider more likely: someone screwed up (as Microsoft's updates often do) or that Microsoft deliberately started upgrading Windows machines in secret hoping that either nobody would care or that they would be okay with just upgrading a few then stopping and having to make up an excuse to cover it up? It's just absurd.
Some mistakes should surely be forgiven in the interests of learning and doing better next time. We might call them "honest mistakes", perhaps actions taken with the best of intentions but which proved to have unexpected and undesirable side effects that could not reasonably have been anticipated.
Other mistakes are either not really mistakes at all or are actions taken by someone who might not have intended the final outcome but should clearly have known better. Consider installing a defeat device in a car to actively circumvent regulations, or a government agent or corporate officer abusing their authority for personal gain, or a police officer who has a track record of being more violent than was necessary when dealing with suspects. At some point, the mistake is no longer relatively harmless or something to be tolerated in the interests of overall progress, it's a demonstration of bad judgement that isn't acceptable for someone in that position.
If you're looking for mistakes that can bring material harm to others, it might be easier to examine industries where errors in judgment lead to physical and/or financial harm. Transportation industries (aviation, trucking, etc), construction, and a whole slew of others offer a variety of examples to choose from that don't necessarily require intentional misuse of resources to cause harm.
Think of how many aircraft accidents have occurred over the decades as a direct consequence of mistakes made by the air crews--or controllers, or maintenance staff--and were, frankly, errors in judgment. Sure, there have been cases where maintenance shortcuts were taken to either save money or time, lead to the deaths of passengers and flight crews, and have been successfully tried in a court of law as negligence (or worse). But that clearly crosses what should hopefully be a distinctive line into unlawful actions, and even still, there have been circumstances where the maintenance staff wasn't held accountable. Again, it's a matter of intent.
So I think context is important. In the case of Microsoft, I don't think that the mistake in question should result in wanton firing. Perhaps if someone could prove that they were materially harmed by the upgrade, then it might be pertinent to seek damages from the company. But oftentimes the developers on staff either did something wrong (and fixed it) or were specifically told to implement that feature by management.
Everything else just seems like splitting hairs.
From my observations, people get fired over mistakes for two reasons - because somebody is stupid/overreactive/wants blood, or because the company wants to signal strong action taken (firing someone whose actions resulted in PR problems).
There's also a class of mistake, which I suppose the one discussed here belonged to, that's a combination of 'honest mistake' of a person along with a failure of the system to capture it in time. If it is known that a mistake in e.g. an update process can affect millions of users, a system should be set up to capture those before they have a chance to do real-world damage. Maybe a second person in another team should sign off update changes, etc. Lack of such a system may be considered not-so-honest mistake of senior management. Sadly, some companies design systems preventing (knowledge of) mistakes from getting out instead.
Cries for blood and similar hyperbole are unlikely to be helpful in any case, but questioning the position of anyone who thought this was acceptable behaviour or who was responsible and signed off on the update seems fair, and so does expecting Microsoft to provide fair compensation if anyone really does get significantly messed around because of this action.
Both true, and yet still it doesn't mean you have to fire the poor guy if the mistake was honest, regardless of damage done. Hell, he'll probably be beating himself over it harder than you imagine if the consequences were big.
Also, this: http://i.stack.imgur.com/eJy3b.jpg.
I disagree. Would an average user, the kind of person who is going to trust updates from Microsoft and leave them on by default and therefore wind up installing in this case in the first place, necessarily realise the significance of one extra confirmation dialog? People are notoriously blind to such messages, so I'd want to see exactly what that extra human intervention involved before making any judgement here.
We set up processes not so that we can have an excuse for firing people who made errors, but so that those errors have limited consequences and could be safely used as learning opportunities.