[To answer your question, the figures typically come from CDNs and major websites doing browser data analysis so, while there's quite a bit of wiggle room, they are going to at least be in the ballpark and definiltely not orders of magnitude out.] Personally, I think people are being too simplistic about the stats and likely outcomes and, arguably, focusing too much on entirely the wrong issue.

I totally agree on your main point; 7-8 years is a good run for a specific major release of an OS, or any other software application. You might not like the decision, but Microsoft announced the end of official support some years ago and that we're now approaching that deadline is just BAU and not something anyone should really have an issue with. They're not taking Windows 10 off you (yet?), but they are making it clear that if you continue to run it you're doing so at your own risk from next month. Pretty much everyone, including the FOSS community, does this with older versions at some point. It's been done countless times before, and will be done countless times in the future - in that light, singling out this one specific example isn't a particularly sound argument, is it?

Whatever the percentage of Win10 holdouts is (I've seen recent figures closer to 50%), it's highly unlikely to be entirely down to "lack of TPM". Windows 11 is a privacy raping UI/UX nightmare, so I suspect a large majority of those Windows 10 holdouts have hardware that actually could run Windows 11 just fine, but are actively choosing not to do so. When many of them inevitably get compromised (which they will), it's going to be interesting to see who gets the blame for that - and the fallout from whatever the resulting botnets are used for - in the media, but that's another topic for another day. There will also be another fraction who simply don't know or don't care; the OS is part of the hardware purchase, and if the hardware is working fine then there's no reason to change anything, and they'll only upgrade when things break (likely due to overworked fans packing up after the CPU has been running flat out for several months as part of some botnet or other). Given most users performance needs have hit a plateau, that could be quite a large fraction, and will naturally decline over time. Finally, you'll have the fraction that understand the issue and have legacy hardware, but can't / won't upgrade because of other user-specific reasons - e.g. they just can't afford it right now.

Key point: none of the people in those groups - probably the majority of that 40-50% - are going to be sending their old PCs to landfill any time soon, and certainly not all in one go on October 14th.

The real issue here is that Microsoft has arbitrarily decided - for the financial benefit of themselves and their hardware/advertising partners - to try and force an unnecessary hardware and OS/"telemetry" upgrade, rather than simply put a banner in the setup process starting with something along the lines of "This hardware lacks critical security functionality and your data may be at increased risk...". We know beyond a doubt that this is an entirely arbitrary hardware requirement decision because of all the workarounds posted online showing how to get Windows 11 running on hardware it supposedly doesn't support. That is pretty much textbook abuse of a monopolistic position in the market, and that's the tack PIRG (and the likes of the DoJ, FTC, EU, etc.) probably should be taking; force Microsoft to remove the arbitrary restriction but make it clear that if you don't have TPM 2.0, that's on you. If you understand what TPM actually does, then you probably also have at least a basic clue about PC/network security and will realise that is pretty much zero additional risk outside of some corporate environments.

Yes, there will still be holdouts, just as there still are on even older software releases and Windows version, but at that point it's entirely on them. They've either chosen the Windows 11 path, with all that entails, or they've chosen some other option (trying to secure a Windows version <11, Linux, Mac, whatever) with all that entails. As long as is not a monoculture with a common failure mode, we should be fine with that.

Indeed, No True Scotsman would use STL with C++.

The last job I had where I had to use C/C++, we automatically ran an expensive static analysis tool every time we checked in code. I'd estimate that it only found about half of the potential segfaults, and it made up for that by finding twice as many false positives.

ve never seen a software distribution mechanism as careless and sloppy as NPM. Bazillions of dependencies and no signing of packages. [ ... ]

Rust's cargo packaging system is almost exactly the same way. And the last time I looked, Go's packaging was very similar. And package signing won't help if the maintainer's key/cert has been exfiltrated and cracked.

This is what you get when you embrace DLL Hell -- the idea that you should pin your program to a single specific revision of a library, rather than, y'know, doing the engineering work to ensure that, as an app author, you're relying only on documented behavior; and, as a library author, to be responsible for creating backward compatibility for old apps linking to old entry points. Sticking to that principle lets you update shared system libraries with the latest enhancements and bug fixes, while remaining relatively sure none of the old clients will break.

"Sometimes you have to break backward compatibility." Agreed, but the interval between those breaks should be measured in years, not days.

Nope. If I had used st::list instead of std::vector, it would have been perfectly fine and officially supported. (Assuming I changed "i+10" to "i+11" in order to make the algorithm actually terminate, although that change wouldn't affect the vector crash.).

The problem is that there are dozens of different rules you have to remember to apply to the different types of lists and iterators. And that's only talking about that one topic. There are hundreds of other rules covering a multitude of language aspects that you have to mentally apply against every single line of code you write, many of which can potentially cause memory corruption.

That is a dangerous misconception. You don't need to use any pointers to get memory errors in C++:

#include <stdio.h>
#include <vector>
 
int main() {
    std::vector<int> v = {1, 2, 3, 4, 5, 6, 7, 8, 9};
    for (auto i : v) {
        if (i % 2 == 0) {
            v.push_back(i + 10);
        }
        printf("%d\n", i);
    }
 
    return 0;
}
 
$ g++ -Wall -pedantic t.cpp
$ echo $?
0
$ ./a.out
 
1
2
-947527061
1600570778
5
6
7
8
9

Are you seriously trying to suggest that never allocating memory is not also a "straight jacket"?

You seem to be saying that a currently existing bowdlerized version C++ is safe for close-world problems. Possibly so, but that still leaves C++ unsuitable for open-world problems. That makes C++ only suitable for niche applications. Why learn it?

If you just use Rust or any other memory safe language, you won't have to worry about what kind of "world" you're writing for, or about choosing from a range of increasingly dangerous "profiles".

We have W+ and Prime and these are probably the main 2 sites I cross-shop these days. They don't overlap entirely; Amazon sells a lot more variety under its own return policies etc, while Walmart is better for things like motor oil that they have locally (if you wanted to go into a Superstore). But there's enough overlap that it's worth cross-shopping, for now.

I dunno. China is a "market socialist" system -- which is a contradiction in terms. If China is socialist, then for practical purposes Norway and Sweden have to be even *more* socialist because they have a comprehensive public welfare system which China lacks. And those Nordic countries are rated quite high on global measures of political and personal freedom, and very low on corruption. In general they outperform the US on most of those measures, although the US is better on measures of business deregulation.

It makes no sense to claim Chinese courts have a lot of power, although it may seem that way â" itâ(TM)s supposed to seem that way. One of the foundational principles of Chinese jurisprudence is party supremacy. Every judge is supervised by a PLC â" party legal committee â" which oversees budgets, discipline and assignments in the judiciary. They consult with the judges in sensitive trials to ensure a politically acceptable outcome.

So it would be more accurate to characterize the courts as an instrument of party power rather than an independent power center.

From time to time Chinese court decisions become politically inconvenient, either through the supervisors in the PLC missing something or through changing circumstances. In those cases there is no formal process for the party to make the courts revisit the decision. Instead the normal procedure is for the inconvenient decision to quietly disappear from the legal databases, as if it never happened. When there is party supremacy, the party can simply rewrite judicial history to its current needs.

An independent judiciary seems like such a minor point; and frankly it is often an impediment to common sense. But without an independent judiciary you canâ(TM)t have rule of law, just rule by law.

Hereâ(TM)s the problem with that scenario: court rulings donâ(TM)t mean much in a state ruled by one party. China has plenty of progressive looking laws that donâ(TM)t get enforced if it is inconvenient to the party. There are emission standards for trucks and cars that should help with their pollution problems, but there are no enforcement mechanisms and officials have no interest in creating any if it would interfere with their economic targets or their private interests.

China is a country of strict rules and lax enforcement, which suits authoritarian rulers very well. It means laws are flouted routinely by virtually everyone, which gives the party leverage. Displease the party, and they have plenty of material to punish you, under color of enforcing laws. It sounds so benign, at least theyâ(TM)re enforcing the law part of the time, right? Wrong. Laws selectively enforced donâ(TM)t serve any public purpose; theyâ(TM)re just instruments of personal power.

Americans often donâ(TM)t seem to understand the difference between rule of law and rule *by* law. Itâ(TM)s ironic because the American Revolution and constitution were historically important in establishing the practicality of rule of law, in which political leaders were not only expected to obey the laws themselves, but had a duty to enforce the law impartially regardless of their personal opinions or interests.

Rule *by* law isnâ(TM)t a Chinese innovation, it was the operating principle for every government before 1789. A government that rules *by* law is only as good as the men wielding power, and since power corrupts, itâ(TM)s never very good for long.

Electric is safe when it's not being used. Gas is never safe: some gas stoves have pilot lights which are inherently polluting 24/7; all gas stoves can fail and fill the house with explosive and unbreathable gas.

This is the spec sheet for the Pixel 10. Please point where it says it has Thunderbolt.

This is the spec sheet for the Samsung Galaxy S25. Please point where it says it has Thunderbolt.

One does wonder what they are thinking - why would anybody want or tolerate this?

Most ads we are stuck with because we want the media or service that the ads support. Oh, you want to watch two teams of 53 millionaires play football? OK, but 30% of your time will be watching ads.

On a fridge what is the payoff?

Please name one. There is a difference between a USB-C connection and a Thunderbolt connection.

How about the hardware does not support it? To use a Thunderbolt connection, the motherboard must have a Thunderbolt controller chip. I do not know of any phone that has one currently. Due to cost and space limitations, phone manufacturers do not include them. Laptops and desktops can have them.