They conclude:
> When it comes to your privacy, we strive to make choices easy to understand while also providing clear visibility and control over your data.
This is entirely the wrong approach. Just don't harvest the data in the first place. Have opt-in measures if you must. That's it. That's what a serious commitment to privacy looks like. Anything else is a serious commitment to get away with harvesting as much data as possible while minimizing PR backlash.
That's right-- there is no supported way to opt-out of telemetry on Windows 10 Home or Pro SKUs. You need to use a third-party program to do it, and these settings are lost with every major upgrade, which come twice per year.
> Just don't harvest the data in the first place.
for better or worse, I just don't think this is feasible anymore. Telemetry is, by now, a fundamental part of the engineering process. Products that don't incorporate it are going to be clobbered by products that do. Microsoft didn't start this paradigm, but I think they had to incorporate it in order to stay competitive.
If you can't avoid telemetry, I think the next best thing is to be transparent about it (with respect to what's being collected, how it's being used) and allow the user some control over it (by letting them 'clear' data that's been collected, and opting in/out as needed). In that respect, I don't think the steps taken today are sufficient, but they are headed in the correct direction. I also think they are being more open/transparent than many other tech companies (who do similar things without receiving nearly as much negative press as MS).
Earlier I said "Products that don't incorporate [telemetry] are going to be clobbered by products that do". High QA costs are the biggest part of this.
As it should be when you're developing an OS for 90% of the market across a huge variety of hardware. There's really no way around that, and jettisoning QA teams as a cost-saving response is incomprehensible to me.
Trusting telemetry alone to ensure the robustness of your product will result in a "good enough to ship" focus on mainstream work-flows and leave corner-cases unaddressed since they're rare. And corner-cases are where the bad, system-breaking things happen.
Telemetry should help QA people write mundane automated test-cases to cover the 99% and then enable them to go and spend the constructive part of their time focusing on breaking the system. But it's absolutely not a replacement for QA.
From where I sit, both roles are combined. The output is a far smoother engineering process and a higher quality OS. You mentioned 'corner cases [being] unaddressed' but I've witnessed the exact opposite: most of the criticism I hear about Win10 has nothing to do with quality.
I understand the value of telemetry-- it truly does help improve software. And I'm fine with telemetry being active by default too. If it wasn't active by default its value would plummet as very few people would /choose/ to turn it on.
Windows telemetry should be set to "Basic" by default on every Windows install, and the user can choose to turn it off entirely or set it to "Full".
As a more sophisticated user who values his privacy, all I want is a supported method to opt-out. That's it.
I don't want to be forced to use a third-party program to do it. I don't want Windows updates to frequently reset my clear preferences, so I need to keep running that program on a regular basis. I just want a switch to turn it OFF.
Microsoft acknowledges that full documents along with stack traces (that can contain sensitive information) can end up on Microsoft servers when error log reporting is enabled. Sending a call stack consisting of function names and pointer offsets and register values to aid debugging is one thing. Sending entire globs of application memory along with the crash log is much less defensible.
My objection about linking user data to real world people still stands. There is no technical reason why _personalized_ logging is necessary. Telemetrics can be pretty innocent (e.g. track which percentage of Windows users have multiple monitors), but as they become more personal it's again much harder to justify.
The absence of a single switch (during install or first login) to get rid of all logging/monitoring/telemetrics and other privacy-invading nonsense is indefensible.
I'm sorry, but it's not economically realistic to demand that companies trying to minimize cost stop all data collection. They're not selling it: they're using it to make their developers work on real problems and improve real features. That's the same as nearly every mobile app. Watch how much your phone sings to mixpanel, kahuna, or google analytics to get a sense of what I mean.
You can raise the price substantially, or you can find ways to make things cheaper. The market is really clear they want the cheaper and are fine with benign telemetry collection. But what's really absurd is that you then decide that doing what a small minority of users are asking for is a PR stunt because it doesn't go far enough in one step.
Are you aware that Windows 10 is ad supported? If the data isn't being sold today, it will as soon as quarterly profits dip low enough.
Which brings me to your second point, that customers are fine with it. The underlying assumption is that consumers are Homo Economicus: rational agents making intelligent, informed decisions. This is not the case, as evidenced by the overwhelming difference between opt-in, opt-out, and must-choose behavior. Customers don't care about tracking, as you say, but not because they've thought about the issue and concluded the trade-off is in their benefit -- in reality they've never thought about the issue at all. Regular people are indifferent to iPhone privacy until they're asked if they ever have conversions or ever take photos they don't want to become public. And then it all clicks, and then they do care about their privacy. People don't care, until they do.
When you call telemetry "benign" you're begging the question: the definition of benign implies you can't reasonably be opposed to it. So that's fallacious.
I never said what Microsoft did was a PR stunt. My claim is that Microsoft isn't genuinely committed to the privacy of their customers, because these privacy measures are (a) insufficient (b) an afterthought (c) opt-out only. I see this as an effort to appease disgruntled customers, not as PR stunt. And as somebody else pointed out already, it still isn't possible to opt-out entirely.
Yes I suppose if you draw an unrealistic line between data collection and Facebook suddenly sharing photos you could scare up any sort of response you want.
Sharing content vs app usage metadata are quite different in point of fact though.
> People don't care, until they do.
Yes. I suppose this is true in the sense that it's tautological. I only bring that up because of your meticulous dedication to discarding social context in:
> When you call telemetry "benign" you're begging the question: the definition of benign implies you can't reasonably be opposed to it. So that's fallacious.
(P.S., this is not "begging the question". "Begging the question" is not "Obviously we should ask." If you're going to use a technical term, perhaps look up its technical definition? I do not think this assertion is unusually cyclical.)
> I see this as an effort to appease disgruntled customers, not as PR stunt.
I sure do wish people would hold desktop to the same uncompromising standard they hold Windows to...
> And as somebody else pointed out already, it still isn't possible to opt-out entirely.
Actually, they were referring to the current state of affairs for specific editions of windows. We are not entirely sure how this will change with the new edition.
2. I am using the technical (dictionary) definition of "begging the question" and I'm using it correctly. We're not in agreement whether the data collection is fine. You say it is, I say it isn't. Begging the question is the fallacy where you assume the statement under examination to be true, which do you when you label collection "benign". After all, if I agreed it's benign we wouldn't be having this discussion.
3. Non-sequitur.
4. No evidence has been provided that opting out entirely will be possible, and it's not possible today, so it stands to reason it won't be possible in the future either (except perhaps with registry hacks).
2. No. You didn't. And no, it isn't quite that.
3. Industry security as a holistic vision does indeed seem to be immaterial to this conversation. I talk back against this point in this community because I'd prefer not to see the slashdot-esque "M$" mindset take over the discourse in an industry that sorely needs competition in this space.
4. Since no evidence has been provided either way it doesn't seem safe to assume either way. I'm not sure why the lack of clear policy here is license for you to say "It is not changing."