🔗 Quick links thread
-
-
-
@pie_flavor I love the fake slo-mos. And the guy in the back sitting with his arms crossed.
-
-
@boomzilla said in 🔗 Quick links thread:
I have a friend who is a chemistry teacher and he uses cooking as real world examples of some of the principles.
-
@Karla said in 🔗 Quick links thread:
@boomzilla said in 🔗 Quick links thread:
I have a friend who is a chemistry teacher and he uses cooking as real world examples of some of the principles.
One of my colleagues actually teaches a "chemistry of cooking" class where you make food and talk about the underlying principles. I use it as an example pretty frequently myself.
-
@Benjamin-Hall Maillard reactions...
-
@Benjamin-Hall said in 🔗 Quick links thread:
@Karla said in 🔗 Quick links thread:
@boomzilla said in 🔗 Quick links thread:
I have a friend who is a chemistry teacher and he uses cooking as real world examples of some of the principles.
One of my colleagues actually teaches a "chemistry of cooking" class where you make food and talk about the underlying principles. I use it as an example pretty frequently myself.
-
Pollute your history with random searches to confuse advertisers etc.
-
Built atop uBlock Origin, AdNauseam quietly clicks on every blocked ad, registering a visit on ad networks' databases. As the collected data gathered shows an omnivorous click-stream, user tracking, targeting and surveillance become futile.
-
@bb36e Boy, Chrome really does not like that extension. I had to edit the registry before Chrome would agree to install it, and even then I still get a message that it's disabled this extension every time I start Chrome.
-
@pie_flavor said in 🔗 Quick links thread:
ChromeGoogle really does not like that extensionbut yeah, using it on chrome is annoying.
-
@bb36e It's working, though. Here's the real question - am I making websites money by visiting them?
-
@bb36e said in 🔗 Quick links thread:
Built atop uBlock Origin, AdNauseam quietly clicks on every blocked ad, registering a visit on ad networks' databases. As the collected data gathered shows an omnivorous click-stream, user tracking, targeting and surveillance become futile.
But it only interacts with the ads you would have seen, right? That way, it doesn't add any outside randomness and they would actually seem to be better able to track what you visit, while not knowing what ads you liked better or worse.
-
-
-
-
@pie_flavor Bonus points for immediately telling what it does on the home page.
-
-
@Gribnit hmm?
-
Interesting.
-
-
Since there's no onebox: It's a decentralized VCS based on patches instead of versions; most Git nonsense is done away with while maintaining essentially the same usage.
-
Generated using this:
-
@Zecc said in 🔗 Quick links thread:
Generated using this:
I kinda want to train it on pony data and see what it does...
-
-
@hungrier said in 🔗 Quick links thread:
Hold up.. is he saying someone actually looks at those Windows bug reports you can send?
-
This is an interesting take on multiplayer scalability in general.
https://minecraft.curseforge.com/projects/tiquality
-
-
@loopback0 Are there any other shapes available?
-
-
@bb36e said in 🔗 Quick links thread:
Being a relatively new doctor, I ask the passing attending physician, "how do these people die?".
He replies, "usually alone" - and continues to walk on into another patient's room.
That is how I would prefer it, so I could focus on the experience.
-
-
@bb36e said in 🔗 Quick links thread:
Being a relatively new doctor, I ask the passing attending physician, "how do these people die?".
He replies, "usually alone" - and continues to walk on into another patient's room.
That calls for a drink!
-
I don't read a lot of ACM articles all the way through, but this one seemed pretty interesting.
-
@jinpa The author oversimplifies, but he's right that the programming models that work for very large numbers of threads are not at all like C. It's just that they're not necessarily like the other things he lists either; they're pretty alien to all current software programming. (Hardware is closer, except they try to use near-global clocks and that's totally not the right approach for scaling up, even really to a whole chip nowadays.)
-
@dkf said in 🔗 Quick links thread:
@jinpa The author oversimplifies, but he's right that the programming models that work for very large numbers of threads are not at all like C. It's just that they're not necessarily like the other things he lists either; they're pretty alien to all current software programming. (Hardware is closer, except they try to use near-global clocks and that's totally not the right approach for scaling up, even really to a whole chip nowadays.)
I also thought his implied point was interesting that there's a sort of deadlock (my paraphrase). It's not profitable to market better hardware that doesn't support the C-related illusions he mentioned, because UNIX-family OS's rely so heavily on C, but OTOH as long as the hardware has those C-features, C-based software (OS-related, primarily) will continue to be used.
-
@dkf It's an interesting consideration, especially since the high-volume data processing I work with day-to-day is mostly embarrassingly parallel.
-
@jinpa There's still a lot of work that needs doing that is best described serially, and nobody can really comprehend systems with millions of threads.
-
@dkf said in 🔗 Quick links thread:
Hardware is closer, except they try to use near-global clocks and that's totally not the right approach for scaling up, even really to a whole chip nowadays.
This is true. A modern chip may have several clocks running at various frequencies for core logic, memory interfaces, and various bus interfaces. Even at a single frequency, a clock may be split into many domains that can be independently enabled or disabled as sections of the chip are active or idle — even down to the level of individual flip-flops1.
1 A flip-flop is a device that stores a logic value (0/1)2 for later use. It samples the value of its data input when its clock input changes from false to true and keeps that value on its output until the next time its clock input changes false to true. In older designs, the clocks typically toggled continuously, and a control signal determined whether to update the flip-flop's value by connecting its data input to either whatever logic was generating the new value, or to its own output to retain its current value. In modern, low power designs, the control signal is instead used to generate a clock pulse only when the flip-flop is to be updated. This saves power, and sometimes makes the chip very slightly smaller, but creates additional difficulty in making sure the chip meets its specs — although this is common practice nowadays, so I'd be surprised if the tools haven't been enhanced to handle it without any additional human effort (it's been several years since I've had any involvement with this part of the process).
2 A single bit. Multiple flip-flops may be used to make multi-bit registers. For example, a 64-bit CPU register is 64 flip-flops sharing common clock and control signals; each bit of the data is connected to one of the flip-flops.
-
@dkf said in 🔗 Quick links thread:
nobody can really comprehend systems with millions of threads.
Except hardware is inherently massively multithreaded, and so are the software models of the hardware.
There are three things that keep us hardware engineers somewhat sane[citation needed]. All of those threads are synchronized on every clock tick (for varying definitions of all and every), the scheduling is well defined in the language specification, and all of the thread scheduling and synchronization implementation is hidden in the runtime stuff that costs $10000/license.
-
@HardwareGeek … and all of those are techniques that cannot possibly work in software. Except the $10k/license, and even then that doesn't really get a lot of takers these days.
Our system, when people think at the level of it where that makes sense, can quite easily have a lot of millions of threads. It's programmed in Python (at that level) but that's really just a way to write a description of the program; the program itself is just a graph of threads and communication channels connecting them. It's also totally alien to the way people normally write programs, and very different from how people make hardware too. Perhaps it is more like how large-scale ISPs operate at the high level, except that the reliability guarantees are totally different too and that changes everything…
-
@dkf said in 🔗 Quick links thread:
a lot of millions of threads. It's programmed in Python
:face_with_raised_eyebrow_while_not_looking_like_a_pervert:
-
@dkf said in 🔗 Quick links thread:
can quite easily have a lot of millions of threads. It's programmed in Python
I need further explanation of that. Is the Python part only managing the threads, or is the code that runs on threads also Python? If so, what about Global Interpreter Lock?
-
@Gąska said in 🔗 Quick links thread:
Is the Python part only managing the threads
It's building a description of the system (effectively compiling the program in a sense) and pushing the program it builds onto the hardware that can actually run that many threads. Standard desktop computers really aren't built to run with a squillion software threads. The Python's just really doing C&C.
Standard Python isn't really all that happy running two threads at once, let alone a million…
-
@dkf yeah, that's exactly why I asked. Yes, if it's just a generator-slash-compiler, it doesn't really matter what it's written in.
-
@Gąska said in 🔗 Quick links thread:
it doesn't really matter what it's written in
Except for two factors, yes: the first factor is the historical one — the code is already written in Python and it's a heck of a lot of work to translate it into something else — and the second is that our users expect to use Python to interact with the system.
-
@dkf whenever I say anything about anything in IT, especially about software solutions, there's always the implied "except for historical reasons". It's so common that it's not even worth mentioning.
-
Holy carp, there's a lot of stuff in space around Earth! Fortunately, there's this interactive live 3D map of it…
-
@dkf Some of those orbits are quite elliptical...