> Society can hardly intervene in every individual's self destruction. But it cannot allow drugs which are not useful for anything BUT self destruction.
Why not? Should I not be permitted to destroy myself if I so choose?
Because society needs you and who are you thinking you can take decisions for yourself not asking the government for permission? Get in line, you dangerous anarchist hippy, and be a good cog. And if you try to harm yourself again, we will put you in the cage with most violent and sociopathic people we could find and keep there until you understand we're doing this for your own good. That provided we don't kill you while trying to take from you dangerous chemicals that aren't good for you.
In a phrase: "does this just go into my iTunes then?". A large majority of Apple's customers presumably doesn't care about anything you mention, given by the simplicity of the iTunes model: you get music from the same place you listen to it and you put it on your iPod from there too.
The strategy of "few things, well executed" seems to be working fairly well too.
IIRC most of these bugs arise because of things that should be available while the device is unlocked: the dialler and camera for example. Camera is supposed to restrict gallery access and the dialler is supposed to only permit emergency calls. I'd expect that every app trusted with running while the device is locked will have these bugs as Apple goes forward too.
The bugs seem to a bit more nuanced than just testing for a locked device; the attacks seem to rely on performing actions simultaneously to exploit race conditions much like weird glitches in games. This class of bugs is really hard to test for due to the large search space. Model checking might offer a solution, but it's not a magic bullet by any means.
Speaking of games, that's exactly what the humble games tester spends a good chunk of time doing: uncovering bugs by coming up with weird things to try, like spamming input at unexpected moments.
Scientific publishing isn't just about putting stuff somewhere for people to read it. Publishers, through their parasitic relationship with academia, have kept themselves relevant through the critical mass of people continuing to use them.
The purpose served by publishers isn't really the hosting of articles, it's the names of journals the publishers own which carries an implication that top-tier journals get better reviewers and therefore higher quality articles. As long as academia values impact factor and journal rankings, the parasite can subsist.
Most academics tend to self-host preprints, or stick them on arXiv so the publisher as a hosting service isn't really the "value-add" offered.
On a similar thread, Julian Bangert and Sergey Bratus have an implementation of the game of life that performs its computation on the MMU using traps[1] that technically uses 0 x86 instructions to run the computation. There's a video of it running available too [2].
> Words cannot express how much of an impact this OS would have had if the amount of time and resources dedicated to Linux, which is yet another UNIX clone in an age where 4-year CS majors do UNIX clones as assignments[1], were dedicated to ReactOS. Where might Microsoft be now if that had happened?
MS does offer some of the NT kernel source under an academic license precisely for this purpose [1]. I'm not sure if it's a response to ReactOS (does anyone know?).
The other companion to this work (recovering COS disk images) is worth a read too [1]. Someone picked up where Chris left off and managed to get the images booting in an emulator [2].
Isn't the point of dynamically loading content to reduce the perceived waiting time though? I wonder if focus on cleanly showing the user the page is rendering might be better placed on getting content to load faster instead.
At the UI level, perceived time is usually more important than real time. Half a second with nothing happening feels long. A whole second with feedback in some form (a loading bar, a spinner, etc.) feels fast.
For longer wait times, this becomes critical. 5 seconds with no feedback feels incredibly, ridiculously long. 5 seconds with feedback feels like completely reasonable. I sometimes use almost entirely fake progress bars which asymptotically approach full when I have no way to measure how long something will take. It's bizarre, but even knowing that the progress bar is fake, watching it fill up makes the time seem shorter than staring at a spinner.
Actually improving loading time is important, of course, but the ROI is tiny compared to basic user feedback.
Why not? Should I not be permitted to destroy myself if I so choose?