Hacker News Reader

Hacker News

Comment by rkagerer | Hacker News Reader

Parent Full thread

rkagerer

They ought to consider why users don't want (and are actively circumventing) the updates.

View on HN

> ought to consider why

It's a stunning indictment of the entire concept of telemetry. They know what people are doing but have no idea why.

You can root out and eliminate bugs this way, but improving the experience is a crap shoot.

next they'll hook the camera and microphone into the telemetry service

(sounds far fetched, right? if you told me 10 years ago that the 2020 version of Windows Pro had adware, spyware and reboots whenever it wants I would have laughed at you)

I may be a cynical, and lets face it angry, old man, but I am quite confident they do know why and just don't care because their priorities do not lie with making the user experience good.

Counterpoint: A critical security bug in Windows led to the power grid going down due to a worm. Maybe security updates shouldn't be dismissible for weeks and weeks?

Counter-Counterpoint: if Microsoft made updates painless, users would notice them less, and defer them less, and be more up to date/secure.

Updates are highly disruptive. An update happening at the wrong time forever taints a user's view of updates: an unexpected reboot and hour long install the night before a term paper is due, a reboot to fix a crashing program unexpectedly installs some lengthy update, update progress bars are bigger liars than politicians, an update fails so Windows has to un-apply the update which takes $ages for some reason the user doesn't care about. The way updates happen on Windows is anti-user so it's not surprising so many users go out of their way to avoid them.

To extend the counter-counterpoint: there are at least two types of updates: feature updates, and security updates. I don't think people would mind the latter all that much, but they often mind the former. As long as they're not put on parallel tracks, many people will refrain from updates entirely.

They are to some extent, but so many security updates are coming and breaking apart that it makes no difference.

> Counter-Counterpoint: if Microsoft made updates painless, users would notice them less, and defer them less, and be more up to date/secure.

But it is painless, at least for me. Just install it a few days after patch tuesday before you go to bed and you should be fine. Sure, you can probably make the experience even better (eg. hotpatching), but the current process doesn't seem hard or demanding.

Windows 10 will literally auto reboot while you are playing game, ignore the fact it's a game and think of it as high resource usage. People probably don't care if their machine reboots in the middle of the night, they do care when it decides to reboot right in the middle of usage.

Or a simulation/long running task you kick off knowing it'll take most of the night, then wake up to a rebooted computer.

I don't even defer updates if I can help it, but have still been bitten by this multiple times.

>Windows 10 will literally auto reboot while you are playing game, ignore the fact it's a game and think of it as high resource usage.

I feel like we're missing some context here. Is this happening the week that the patch was released? Or was the system unpatched for weeks/months, and the user clicked "delay" multiple times?

>People probably don't care if their machine reboots in the middle of the night, they do care when it decides to reboot right in the middle of usage.

The grandparent comment was complaining exactly about this.

> I feel like we're missing some context here. Is this happening the week that the patch was released? Or was the system unpatched for weeks/months, and the user clicked "delay" multiple times?

It doesn't matter. No means no.

Not by design generally speaking; if this happens, it's a bug, and you should file one.

Eh, it was just one of the multitude of user hostile actions that drove me off Windows. Intentional or not, it isn't my problem any more.

Tell that to millions who were surprised to find their printers broken just this month.

https://news.ycombinator.com/item?id=23516897

Hotpatching actually used to exist in Windows versions prior to Windows 8, but it was removed because few updates were made to support it. In the 12 years of support for Windows Server 2003, only 10 updates were released that supported hotpatching. Additionally, malicious software has been reported to have taken advantage of the feature. [1]

[1]: https://www.computerworld.com/article/3061998/group-uses-win...

Your parent comment, and all of their ancestor comments, are talking about updates in general, and not about security updates specifically.

Here over on the Linux side of the world, we have enjoyed the separation of security updates from feature updates for such a long time. We even voluntarily turn on automatic updates on our servers because of how much confidence we have in the stability of the system and that the updates shall not compromise that.

In a corporate network full of critical systems, absolutely! On a home PC used to check email or play games... no way should updates be forced upon an end user who specifically opts out. Note this update applies to Win10 "Pro" versions, not just enterprise or volume licensing, so developers who need to test/support older versions will also have problems.

I think you have it backwards. It's the home users that you have to force updates on. Corporate users will have their own patching schedule that's usually on a 1-2 week delay from patch Tuesday.

Like it's wild to me that people want to leave their boxes unpatched for a year because it's annoying.

Getting your pc compromised as part of a botnet or malicious code keylogging or uploading your data online through an exploit is something a lot of power users feel they are immune to.

The corporate users I know are rather 1-2 years delay...

A lot of system I see are left unpatched because no one has time to fight bugs in windows update stuck at some random update. If updates didn't take 1-10 hours to install and still fail 1/4 of the time uptake would be better.

That should be up to the organization, because the other side of that coin is the power grid going down because a critical process was disturbed by changes brought in the update.

Counter-Counterpoint: The features allowed you to specify different delays for feature and security updates.

Why not just take out the security one and leave the feature update, especially since 2004 has been, once again, a complete trainwreck that left many users without a functioning system until they manually blocked people from updating. I've been lucky enough that by the time my PC found out about 2004 it was already on the blocklist, but friends of mine haven't been so lucky.

In order to clarify for those unaware: '2004' here is not referencing a year, it's a version/release number for Windows 10 (with a build number of 19041, just because Microsoft likes to confuse people).

Similar to Ubuntu releases, it's last two digits of year followed by two-digit month (though the actual release isn't always in the nominal month, IIRC, but it's usually close.)

Ubuntu puts a dot between year and month so it's not confusing.

Yeah. I see both sides of this argument, and don't really know where to stand. Probably some middle ground. But I think that this isn't as straightforward of an issue as some people are making it out to be, and pretty much any decision is going to piss off someone. But I also see the points of separating security updates and feature updates.

I think that's a failure in the organization that didn't have a security update policy, not a failure in Windows.

Counterpoint: this was 17 years ago, and that type of worm would not work today for many other reasons.

Reply on news.ycombinator.com