A few days ago, for the first time in years, I tried installing Windows. Booting Windows for the first time was shocking, and I didn't realize how bad things had become. I experienced the same banner warning trying to dissuade me from installing Firefox among many other annoyances (focus stealing, missing drivers, Windows update taking far longer than it should, adverts, just to name a few).
People give Linux a bad rap, but holy shit, using Windows for the first time in years was a truly miserable experience. Not once did I feel like it was a tool there to help me, it felt hostile.
And that's after some rather serious debloating...
The only reason for the upgrade is that some software doesn't support win7.
As expected the only thing which works better/faster on the new PC is games.
It's a downgrade in UX/UI too, it looks way worse than win95 classic theme (no point comparing it to aero)
1. VirtualAlloc:
1.1: separation of reserving memory from committing it, thus making overcommit unnecessary, when you want to have memory arenas and reserve a large contiguous chunk of virtual address space, but commit it (so tell the system that you actually want this virtual memory mapped to some physical memory), when you need it. So you can grow your arenas as you need more memory, without worrying that you may not have enough virtual space due to fragmentation, and without relying on invisible things like lazily committing memory on page-faults as long as your program doesn't have any bugs which make it touch that memory in advance and thus make the bugs go unnoticed silently and increase memory use
1.2: ability to implement a circular buffer[1] with a contiguous view instead of head + tail parts; thus making it possible to pass your ring buffer to functions which expect a normal array, and also avoid branches and complexity in code operating on the buffer.
Generally, VirtualAlloc > mmap IMO.
2. Behaviour in OOM conditions. Linux can become completely unresponsive for minutes. Windows is usually responsive enough to kill the offending applications.
3. GUI apps have reasonable baseline for performance of GUI. On Windows it's expected that watching YouTube/Netflix you won't see screen tearing. On Linux screen tearing with VOD in browsers is completely unsurprising.
4. Powershell is nice for a whole variety of reasons.
5. CreateProcess > fork [2]
6. MSVC, unlike GCC and Clang, doesn't come with guns for your children when you unknowingly dare trigger UB.
[1]: <https://docs.microsoft.com/en-us/windows/win32/api/memoryapi...>
[2]: <https://www.microsoft.com/en-us/research/uploads/prod/2019/0...>
(Posted from Linux btw.)
You likely have “Allow Apps from Anywhere” selected whereas the person in the posted tweet has the “Warn Me…” option enabled.
What is your point? That all those popups were photoshopped and don't actually exist?
Or what kind of headline about this topic would you like to see on HN? "Windows sometimes does not show its nag screens when you install Firefox"?
Your comment feels like gaslighting, considering you already know the answer to the question since OP is the one who suggested banning it.
There is really something weird with our virtual tools: we tolerate things that we would consider bonkers IRL. Like a newspaper shop asking for your phone number before you can read an article, or a library reporting the books you buy to the local candy shop.
A/B testing is one of those things: you have product you paid for, and the people who sold you the product change it without telling you, then revert abruptly the change, like it was a bad dream. But only sometimes, and not for your friends, who think you are crazy to believe it happened.
Anyway, there is an easy test to know if what you are doing is the right thing: ask the users if they want to opt-in. Most of them won't say yes, because, who would want that? Like tracking and advertising, only a tiny minority of users are actually ok with it if they have a choice. If you had a ublock origin for A/B testing, wouldn't you use it?
The only thing this test doesn't work with are duties and chores, like payment. But A/B testing is neither of those.
I think backend/performance-related testing, opt-in testing, and internal employee A/B rollouts are fine though.
An A/B test in such an instance is a HUGE diservice. I imagine there are others, but that that is what I can immediately relate to.
In such a case, I would feel like whomever implemented the A/B test directly caused needless pain and suffering: both of our frustrations and whatever inadequacies/helplessness my mother might feel.
So yes, I'm in the same boat: either make it opt-in, or make it illegal.
Nobody would think this is alright.
Besides, I'm advocating for _non explicit_ A/B testing. Let the people know you are fucking with them at least.
Perhaps you should give your recording another watch. It clearly shows the "There's no need to download a new web browser" message box above bing's results.
I originally thought the word "download" in "firefox download" could be the contextual trigger for that behavior, regardless of browser. But your recording clearly shows that you searched only for "firefox", with no other context, and still had bing giving you the message, which confirms that they're singling out Firefox as a threat. Given the passive suggestion in their message that Firefox is slow, insecure, or costly in time or money, I'd say the outrage is justified.
You haven't experienced the rest of the issues because in your recording you haven't actually tried to install and use Firefox.
Unless you're saying that person that experienced this is lying or misunderstood what happened?
Mozilla was caught "studying" (read, spying) on a tiny portion of its user-base who have explicitly opted out of telemetry. The fact that most weren't spied upon doesn't mean that it didn't happen.
This is what I don't like about people who make misinformed statements on HN.
This is what I don't like about social media sites like HN either where only the rage will reach the front page
Edit: screen recording https://streamable.com/wpjv2h