Live data from Hacker News

German federal office publishes Windows 10 telemetry analysis

ghacks.net

211–220 of 239 posts

Re: German federal office publishes Windows 10 telemetry analysis

#211
post #113

Earlier quoted context omitted.

Care to elaborate? Afaik and tested w10 does NOT bypass its own firewall when set to 'drop all' traffic.

From what I've seen "drop all" works at the moment but directly blocking the telemetry programs keeps getting undone whenever the file changes and it intentionally doesn't use the hosts file.

Dont know what you mean by undone, but running firewall in drop all + whitelisting works fine and block all telemetry (one caveat : dnscache needs to be disabled to remove possibility of dns tunneling)

Re: German federal office publishes Windows 10 telemetry analysis

#212

Earlier quoted context omitted.

> Very much disagree. If you pay attention at all to security you'll always hear stories of that one client who hasn't done security updates in 3 months (or 3 days) and gets taken out by the inevitable "exploit Wednesday" shenanigans. And if you read past the headline you find out that they had numerous other problems with their security practices that were actually responsible for them being compromised, because if…

I agree with the sentiment that "everything is bad". Computers are still in their birthing stage and full of problems. But Windows 10 feels like an improvement over any previous version of Windows at least. Which is really all we can ask for. A little too transitional/work in progress for some, I can see that (2 control panels? madness!), but otherwise pretty smooth. But maybe it's also because I left computer suppor…

Lucky you. Meanwhile Microsoft has delayed the October update for the third time because of all the crap they keep breaking whenever they release it. Yesterday my boss and I had a conversation about just turning Windows Update off forever for our org because it's caused more trouble than pretty much anything we could think of as a consequence of not updating. This was brought up (again) because of several of our developers having their network cards randomly and inexplicably stop working and requiring a reboot to fix after 1809.

I will say that since Vista there has been a ton of improvement on the driver model, display system, network stack, audio stack (as you mentioned), and a few other relatively low-level pieces, but pretty much everything above that has been a continual shitshow of user-hostile bullshit and overly complex interlocking crapware.

Start Menu used to just be a menuized view of a folder structure, now it's some horrid amalgamation of junk that breaks whenever the Windows Apptore database gets corrupt, which is surprisingly often considering no one here even uses Windows Appstore apps.

Older Windows versions had their bugs, sure, but on the whole they cared a lot more about the user, about not breaking their shit and making sure their updates were actually worth applying. Microsoft has made it very clear since Windows 10 that they hate the user.

Re: German federal office publishes Windows 10 telemetry analysis

#213
post #211

Earlier quoted context omitted.

From what I've seen "drop all" works at the moment but directly blocking the telemetry programs keeps getting undone whenever the file changes and it intentionally doesn't use the hosts file.

Dont know what you mean by undone, but running firewall in drop all + whitelisting works fine and block all telemetry (one caveat : dnscache needs to be disabled to remove possibility of dns tunneling)

From what I remember, when you say "block all connections by this program" on the relevant files it doesn't just use the filename, it tracks the exact version of the file, and as soon as windows update touches it your firewall block stops working.

So a whitelist method works, good, but I think it's fair to said that specifically blocking telemetry is extremely difficult with the way blacklisting works.

Re: German federal office publishes Windows 10 telemetry analysis

#214

Earlier quoted context omitted.

> You are contradicting yourself here. No I am not. I was just saying that the only thing I've hadn't have work out of the box was some dodgy PCI card from China. It is literally the only thing in the last ten years that wasn't a video card that didn't work out of the box for me. > Besides, linux hardware support is incredible nowadays, and most of the time if you have problems with linux, you would have them with wi…

You confuse your subjective experience with the overall picture. The list of devices which kernel does support is not only incredible, but higher than that of any other OS safe Windows maybe. And even in windows you will have pretty the same hardware problems, just with the different set of hardware, which you were lucky to avoid. Shit like this [1] [2] is pretty common in windows world as well. >The touchpad and tou…

> You confuse your subjective experience with the overall picture.

No it is a common complaint that has been happening for years (over a decade) with all manner of consumer laptops.

> The list of devices which kernel does support is not only incredible, but higher than that of any other OS safe Windows maybe.

Yes and I would wager quite a lot of these devices are for ancient hardware, embedded devices, servers, micro-controller etc and other stuff THAT IS NOT ANYTHING TO DO WITH BUSINESS LAPTOPS and a reasonably priced consumer usb headsets from well known manufacturers.

It is a fallacy that just because there is a large number of devices it also means:

1) They are supported well.

2) They are my devices.

3) That there are other parts of the distro (Pulse Audio, ALSA or whatever the nonsense they are using for an audio stack these days) will interfere with how the device works.

The situation will never change. It will never change because

1) Device manufacturers don't care about Linux. They will care about MacOS, Windows, Android and iOS.

2) None of the large corps that basically contribute to the kernel really care about Linux on the desktop. They don't make a lot of money if any from it. Redhat kinda bother, but they've been bought by IBM now so that won't last much longer.

3) As demonstrated in this very discussion on this topic. Most Linux users will trot out the same tired old excuses why shite doesn't work. They will blame it on Microsoft, the User, the hardware anything other than the accepting the fact that because everyone has their own idea what a distro looks like, the whole community is fragmented. Fragmentation causes problems, instability and compatibility issues.

I've heard it for 15 years now. If I have to use Linux (I do unfortunately), it is whatever the latest LTS of Ubuntu is because it mostly works (it is still shite though).

> And even in windows you will have pretty the same hardware problems, just with the different set of hardware, which you were lucky to avoid. Shit like this [1] [2] is pretty common in windows world as well.

Cherry picking nonsense. The only driver I've needed to install in the last 5 years is my video card driver (I am rocking a 1080Ti, which is rather nice) and a wireless driver in my laptop which took all of 5 minutes to install.

Windows will download the drivers from the internet if it can find them.

> Most of the time you just have an OEM preinstalled for you or even an OS preinstalled on a very particular hardware (macos).

No I installed Windows myself. I always wipe and do a clean install. I've done my own install of MacOS in the past as well (not much point though as they don't fill the OS full of shit).

I've heard all of these arguments before. They are all deflecting blame away from what is the Desktop Linux community. I'd heard these arguments back in 2003. Nothing has changed much in 15 years.

Maybe in 15 years time when Desktop Linux still doesn't work correctly you might get wise to the myriad of reasons why it will never work.

Hopefully I will have retired to somewhere like Cambodia by then.

Re: German federal office publishes Windows 10 telemetry analysis

#216

Earlier quoted context omitted.

No one forced you to use Windows 10. You could have used ReactOS.

I was pretty happy with Windows 7.

I am pretty happy with it (well, relative to the newer ones)

Re: German federal office publishes Windows 10 telemetry analysis

#217

Earlier quoted context omitted.

A domain like 'microsft.com' is used to trick users into thinking the communication is legitimate and to Microsoft. But why would the actual Microsoft need to engage in this?

Disclosure: Works for MS but not on anything related to the topic of discussion here. These are my personal opinions from experiences at multiple companies. Any user-facing names obviously get reviewed by the whole departments that are dedicated to ensuring it fits with the brand, translates well into all languages, not offensive to anyone, etc. For identifiers that are not expected to be user facing, they're likely…

So, basically, laziness

Re: German federal office publishes Windows 10 telemetry analysis

#218

Earlier quoted context omitted.

Disclosure: Works for MS but not on anything related to the topic of discussion here. These are my personal opinions from experiences at multiple companies. Any user-facing names obviously get reviewed by the whole departments that are dedicated to ensuring it fits with the brand, translates well into all languages, not offensive to anyone, etc. For identifiers that are not expected to be user facing, they're likely…

So, basically, laziness

So, basically, efficiency

Re: German federal office publishes Windows 10 telemetry analysis

#219

Earlier quoted context omitted.

Okay. What about hardware that works better on Linux? Why do you assume that this issue strictly favors Windows?

There is very little standard (consumer) hardware that works better or exclusively on Linux (1). Due to market pressure, Windows drivers are widely available, but Linux is not necessarily (fully) supported. As another example: printers. Quite a few (even SO/HO) printers don't have proper Linux drivers. I can't reliably print double-sided on our office brother printer. Getting the networked scanner to work is a lesson…

> Due to market pressure, Windows drivers are widely available, but Linux is not necessarily (fully) supported

That sounds like something Microsoft would say back in the 90s.

> I can't reliably print double-sided on our office brother printer. Getting the networked scanner to work is a lesson in debugging. On MacOS and Windows: Trivial task. Just works.

I've seen many people over the years saying the same thing in the opposite direction: their stuff works better on Linux than on Windows. You're just assuming that Windows is always better based on your personal experience. In reality, vendors often push out poorly written Windows drivers in a hurry and move on. If you're lucky, you might be able to dig out an update from 7-pages-deep on a vendor's web site, and maybe it won't introduce a new bug that keeps you on the old one.

Neither side of that binary is useful. Instead, look at the hardware you need, see how well it's supported on the platform you use by looking at actual user reports, and don't make any assumptions.

However, one assumption you can generally make is that, if a driver is in the Linux kernel, and people are using it, it will continue working, because Linus doesn't tolerate regressions in code that's actually used.

Re: German federal office publishes Windows 10 telemetry analysis

#220

Earlier quoted context omitted.

> I don't think it is unreasonable to expect to be able to download a package from the internet that is marked as "Ubuntu 18.04", use the package manager to install it and for it to work properly. That's preposterous. If you downloaded a random MSI marked "Windows [your version]", installed it on Windows [your version], and it crashed, you wouldn't blame Microsoft. > There is also about 3 or 4 different methods (that…

> That's preposterous. If you downloaded a random MSI marked "Windows [your version]", installed it on Windows [your version], and it crashed, you wouldn't blame Microsoft. No. This is the fundamental misunderstanding of the point I was making. So I am going to make it crystal clear. The problems I've had with installing slack, I've had with Spotify, Steam, Dropbox and there are quite a few others I am sure I've forg…

> No. This is the fundamental misunderstanding of the point I was making. So I am going to make it crystal clear. The problems I've had with installing slack, I've had with Spotify, Steam, Dropbox and there are quite a few others I am sure I've forgotten.

It's not a misunderstanding at all--you're explicitly blaming the platform, and your subsequent comment does just that:

> Even with attempts such as Flatpack, SnapStore etc. The same situation is present. This is because of the nature of any GNU/Linux distro is that they are all fragmented in terms of underlying libraries, kernel versions, UI versions, package managers etc.

Meanwhile, millions of Linux users have no problems with Steam or Dropbox. But you have unspecified problems, and therefore Linux is bad.

> Nothing remotely equivalent exists for Linux.

That's simply false. If you are being honest, then you don't know what you're talking about.

> Have the wrong distro ... there probably isn't a package.

So then use a popular, supported one. Packages for RHEL/CentOS, Fedora, Debian, Ubuntu, and even SuSe are nearly ubiquitous. If you freely choose to use obscure, forked-3-levels-deep Bob's Linux 2018, don't complain that there's no bespoke package for arbitrary software you want. That's silly.

> Have a slightly older distro ... There probably isn't a package for you.

Also false. Most vendor packages are for years-old LTS releases. If you've actually tried this, you should know this.

> Even with attempts such as Flatpack, SnapStore etc. The same situation is present. This is because of the nature of any GNU/Linux distro is that they are all fragmented in terms of underlying libraries, kernel versions, UI versions, package managers etc.

Bizarre. Flatpack and Snap are specifically designed to solve that problem, but you claim that they have that problem. Why are you saying these things that are patently untrue?

> It is a mess. It will always be a mess as long as people don't recognise it as a real problem.

It's just weird that you say "as long as people don't recognize it as a real problem" immediately after mentioning attempts to explicitly solve the problem. Are you writing these words seriously?

> I doubt it will get solved in our lifetimes.

Meanwhile, millions of people happily continue using their Linux desktops. And millions of people continue using Windows, suffering all of its problems, such as not actually being in control of their own machines anymore (it's amazing how far Microsoft has fallen just since Windows 7).

> If you want to make smart ass remarks about it while ignoring the real issue so be it.

Your words agree with neither my comments nor your own. Bizarre.

> Thanks for taking me out of context. I was specifically talking about user environments which are supplied by those who control the distro.

Not out of context at all. You're blaming the platform instead of the app vendor.

> Please don't be disingenuous.

Please don't project. You say A in one sentence and not-A in the next. You don't even seem to agree with yourself.

> People become familiar with certain user interfaces. Saying that everyone should just understand how to use a new interface even if it mimicks (btw poorly) another OS which is ubiquitous really shows that their lack of touch with the majority of users.

You're saying that an aerospace engineer can't, in a few seconds, click what appears to be a Start Menu-like menu in the Start Menu place, type "browser" into the box labeled "search", and find "Firefox Web Browser"? I don't believe you. If he can figure out how to make an aircraft fly, and use complicated fluid dynamics modeling software, he can figure out how to launch Firefox. Going from Windows to Plasma is no more complicated than going from Windows 7 to Windows 8 and 10--in fact, probably much simpler, considering the enormous UI changes Microsoft made. And it's claims like this that suggest it's you being disingenuous.

> Lots of users can have real problems moving between versions of one application. That why user interface guidelines, best practices, user testing and user interface design are its whole own field in this industry.

Then you ought to be heaping criticism on Microsoft for Windows 8 and 10. How many times have I seen someone move from Windows 7 to 8/10 and say, "I can't use this UI, it's awful!" Instead you act as if Linux UIs are uniquely problematic.

> Pretending because something looks superficially the same means that someone should be able to use it is ridiculous.

"A imitates B, but you can't expect people who are accustomed to B to be able to actually use A." Bizarre.

How many times I've heard someone say, "I installed Ubuntu on my 80-year-old grandma's computer and haven't heard her ask for help since. She barely even noticed a difference. She uses the Web and everything." But your hypothetical aerospace engineer can't launch Firefox. Ah, anecdotes.

> KDE plasma UI is what happens when a teenager that has just learnt how to use Photoshop effects tool has gone mad with gradients and bloom effects, the whole thing is a mess visually.

Now this is where I agree with you. I don't like the default Plasma themes. That's been a problem since KDE 4. But that's superficial. You can change all that. With Windows and Mac, you can't--you can take it or leave it. "But that's complicated, users won't want to do that"--beside the point: they can. Windows has an ugly mess of a UI, and you can't fix it. That's where you ought to be complaining.

> Each time there is a update to Windows 10 they have changed the location of some control panel option for the umpteeth time. So I already know how to use powershell and install the management tools, it is normally easier FOR ME to use that rather than try to navigate the labyrinth of control panel options.

Understandable. A few scripts can make life a lot easier on Windows.

> In response to powershell comment. It is normally easier to just learn a shell like bash on a *nix system and just use that then try using the absolutely awful UIs they normally provide you with.

I'm not a fan of the default UIs, but for an average user, they're certainly more usable and discoverable than looking up shell commands. This is another example of the inconsistency of your complaints. You seem to be complaining about 5 different problems at the same time, mixing them all together.

Post reply on HN