Yes. I understand it, this Intel flaw concerns the kernel being mapped into all processes' address space at the same addresses. The kernel's memory pages are marked "Global" (present for all processes) and "Ring 0" (Kernel access only). This means that a system call does not require a context switch - only a flip of a bit inside the CPU. Both Linux, MS Windows and macOS do/did this before the recent patches.
While macOS uses the Mach microkernel which (because it is a microkernel) had been designed to be small(ish) and run alone in Ring 0, Apple has broken the kernel's protection model and put a whole lot of other stuff in global Ring 0 pages as well to increase performance. This makes their "kernel" almost as large as MS Windows and Linux, providing just as much memory that can be attacked through memory/timing-based attacks. So, in effect, macOS would be just as vulnerable.
That's not what "early adopter" means. Early adoption is when you buy a new kind of product that is not mainstream.
The iPhone line is now ten years old. Buying the first iPhone back in 2007/2008 would have been being an early adopter. Buying the iPhone X is jumping on the bandwagon.
I saw "Google Analytics" listed as one of the sites that Firefox delays. I run Privacy Badger in Chromium, so I checked quickly what it blocks on this site and apparently, Slashdot uses Google Analytics but Privacy Badger does not block it. I suppose that there could be lots of other sites that are let through but which Firefox prioritises down when loading. This means that running Privacy Badger is not a replacement for the prioritisation scheme that Firefox is doing.
Integer types in C are the same size on 32-bit and 64-bit Windows. On most 64-bit Unix:es and Linux, "long int" are 64 bits as opposed to 32 bits.
The big difference is that pointers are 64 bits instead of 32. Pointers and 64-bit integers are also 8-byte aligned instead of 4-byte aligned. (at least on Linux they are) It may not seem much but if you have large arrays of structs with these types, it could add up.
Meanwhile, in my small country, brain tumours are on the rise. Cell phones seem to be more popular here than in the US. The cities are more densely populated.
Anyway, I don't think the risk is so much from cell phones themselves. I would be more concerned for people who are living or working too close to a cell phone tower. In the latter case, it is not just the frequency band that the cell phone is using, when it is being used. That tower is on all the time and could be in 2G, 3G, 4G and 5G all at once and/or long-range high-speed directional microwave links between towers instead of fibre.
The electromagnetic waves from mobile phones do not cause the initial mutation that causes a cell to turn into a cancer cell and multiply out of control. That fact has been long established and is not contested here. What microwaves in the frequency ranges used by cell phones have been shown to do is to promote the growth of existing cancel cells.
Mutations into cancer cells are actually not as uncommon as most people think. All of us have had cancer cells many many times, but what normally happens is that the immune system detects the microtumour and kills it long before it has grown large enough to cause a serious problem. All of us have been exposed to cosmic radiation, practically all of us have had X-rays a few times in our lives, all of us have inhaled and ingested carcinogenic chemicals from car exhaust, pollution, first- or second-hand tobacco smoke etc. etc.
Carcinogens are all around us, but you don't get cancer that easily. Cancer risk is about probabilities: many different factors are at play. And radio frequencies in the part of the microwave spectrum, at frequencies and levels that mobile phones use - have been proven in several studies (I have heard of two beside the California study) to be one of those factors.
I agree. Mara Jade even had an official model that used to show up at sci-fi conventions, as did many of the book authors. The problem with the EU is that were quite a few things in the EU that was of questionable quality, and many things even contradicted each other or contradicted facts in the prequel movies.
At a convention in summer of 2013, Lucasfilm announced that they were going to sort the EU out. They were going to pick the best bits and put the main storyline into a new consistent canon and scrap the rest. (I was there in person and heard it directly from Leland Chee's and Pablo Hidalgo's lips) In early 2014, however... Lucasfilm announced that they were instead going to scrap it all and relabel it as "Legends". They still print and sell new copies of those books though - with a new "Legends" label on the cover.
Dave Filoni (showrunner of Clone Wars and Rebels) did pick a lot out from the old EU to use in his series. Grand Admiral Thrawn in Rebels is one of those. The Nightsisters of Dathomir is from the EU also. Darth Maul's comeback with robotic legs is not even from the EU, but from an alternate universe story where he shows up on Tatooine, wanting revenge on Obi-Wan for having been cut in half on Naboo some thirty years earlier...
The two main complains I have seen about The Last Jedi have been:
* It contained too much comedy. It reminded them too much of a Saturday morning cartoon than a Star Wars flick. That's perfectly fine for a movie -- if that is the general tone of the franchise but this is the eight movie in a series and that would in many older viewers' eyes require it to follow an established format or it would get close to jumping the shark. For the casual viewer, however, and for most reviewers -- who don't view the movie more deeply in context -- this does not matter at all.
* At the end, not much has been accomplished. The state of the universe is pretty much as it was when the movie started. I don't agree with that though.
* It is too long and could have been shortened. This movie is two and a half hours where as every other Star Wars has been around two hours.
10K is not 16:9 but an ultra-wide variant of Ultra HD. 10240 * 4320.
There is no such thing as 16K, yet. And if someone tells you they have 16K then they probably have only four times Ultra HD... which is 15K ! 4 * 3840 = 15 * 1024.
Usenet in its prime was more or less an elitist forum - for academics at universities and large corporations. The less fortunate computer nerds used to dial up BBS systems, which had organised its own FidoNet for inter-BBS communication.
Usenet was not succeeded by Facebook. Facebook was made for the less tech-savvy. When Usenet became too spammy, interest groups instead went back to the BBS model, but on the web: specialised forums that they themselves had control over, often running PHPBB... or even SlashCode.
BTW, the microkernel at the bottom of Fuchsia has changed name from Magenta to Zircon. Maybe because there's a Magenta Linux... I dunno.
Zircon has security based on capabilities (which it calls "handles", rightly so IMHO) for pretty much everything. This could support sandboxing of new sub-processes that you own, but it lacks revocation of rights from running processes that would be used as services -- which I find to be a serious omission. IPC is very much like Unix domain sockets: with streams and queued asynchronous message passing... which means that it is never going to be faster than Mach... and Linux is demonstrably faster than Mach. seL4 would have been a better foundation IMHO.
It's actually the opposite. Cell phone speakers these days are so small that overdriving them is considered normal. The DSP in the audio codec/amplifier is supposed to compensate enough to avoid artefacts and to avoid damaging the speaker but it is walking a tight line. I was told this by engineers at Cirrus Logic, which makes audio chips for Apple and other cell phone manufacturers.
DRM servers have killed the used games market on PCs. There are lots and lots of games out there that are over a decade old even that you can't buy used because they have been "activated" on an old DRM server and can't be reactivated on another machine.
In several cases, legitimate copies of a game can't be played at all on any PC because the game had demanded to contact a now discontinued DRM server even to start.
Studies made in þe olde typewriter era have shown that having a wider surface on top of the keys leads to more unintentional key presses. The standard surface width is 1/2" or 12 mm, and with standard width (what is usually meant when talking about "key spacing") being 3/4".
I think that what Rock21k is actually referring to is what is called "island keys" or "chiclet" keys. I don't think that whether the keys' skirts are angled or go straight down matters that much. The problems are rather that chiclet keyboard tend to have flatter surfaces but more often entirely flat, wider surfaces and less key travel than other keyboards. MacBook "Pro"'s keyboards with its ultra-flat "butterfly" scissor mechanism is especially bad.
Also, some popular chiclet laptop keyboards (such as MS Surface "Type Cover") have very low surface friction, so fingers slip more often. Low surface friction wouldn't have been so detrimental to keyboard feel if the keys had been dished and had more space between them. Older keyboard keys tended to be made of plastic, such as ABS or PBT which has good surface feel even when glossy but backlit keys tend to be painted with a slippery paint layer with laser-ablated legends.
... and PCID is not an instruction. The feature means that there is a "process ID" tag on each entry in the TLB to avoid having to flush them unnecessarily. The intended benefit is that all entries would not necessarily have to be reloaded from page tables in RAM (or cache) whenever there is a context switch.
"Tagged TLB"s have been available on other CPU architectures for decades -- and have been used by the Linux kernels for those architectures. The feature is pretty recent on Intel x86 CPUs though. Correct me if I'm mistaken but I think AMD's x86 CPUs do not have PCID specifically but has support for "virtual machine ID" tags on the hypervisor's second-level TLB.
I would say that before C++11, all C++ code could be compiled into the equivalent C code without loss of performance.. which is also how the very first C++ compilers worked. In C++11 came move-semantics and better synchronisation primitives that C did not have. In C, you may be able to use intrinsics for synchronisation and atomic access though but move-semantics even for simple value types can improve low-level performance.
While I am also primarily a C++ programmer, I do mostly agree with the AC. Most of the so called "idioms" in C++ are what would be called "hacks" in any other programming language.
C is more readable than C++. In C++, you often have to spend more mental energy trying to find out what exactly is going on in other people's code. For instance, because what looks like one type of call might actually be something else involving operators that takes place between types higher up in the class hierarchies, and what looks like value-semantics may actually be pointer semantics because references are used. You are often completely dependent on the IntelliSense (or whatever it is called) in the code editor you are using to not get lost.
Next, shared_ptr needs multiple objects for a friggin pointer and C++ lack the concept of "borrowing" which I would say is absolutely essential in a language with unique pointers.
Indeed. I have also tried to add an article about a new subject only to find that it already exists as a rejected and locked page because someone at one time several years ago misunderstood what it was and deemed it "not notable". Also, with no possibility of changing it or adding to the "talk" page.
The oldest product that I have heard of was the Neonode phone, first exhibited in 2002. It had vertical slide-to-unlock, albeit with three different sliders (left, centre, right) that activated three different functions.
The Neonode wasn't from Microsoft but it ran Windows CE underneath its own GUI.
Slashdot Mirror
Yes. I understand it, this Intel flaw concerns the kernel being mapped into all processes' address space at the same addresses. The kernel's memory pages are marked "Global" (present for all processes) and "Ring 0" (Kernel access only). This means that a system call does not require a context switch - only a flip of a bit inside the CPU.
Both Linux, MS Windows and macOS do/did this before the recent patches.
While macOS uses the Mach microkernel which (because it is a microkernel) had been designed to be small(ish) and run alone in Ring 0, Apple has broken the kernel's protection model and put a whole lot of other stuff in global Ring 0 pages as well to increase performance.
This makes their "kernel" almost as large as MS Windows and Linux, providing just as much memory that can be attacked through memory/timing-based attacks.
So, in effect, macOS would be just as vulnerable.
That's not what "early adopter" means. Early adoption is when you buy a new kind of product that is not mainstream.
The iPhone line is now ten years old. Buying the first iPhone back in 2007/2008 would have been being an early adopter. Buying the iPhone X is jumping on the bandwagon.
I saw "Google Analytics" listed as one of the sites that Firefox delays. I run Privacy Badger in Chromium, so I checked quickly what it blocks on this site and apparently, Slashdot uses Google Analytics but Privacy Badger does not block it.
I suppose that there could be lots of other sites that are let through but which Firefox prioritises down when loading.
This means that running Privacy Badger is not a replacement for the prioritisation scheme that Firefox is doing.
Integer types in C are the same size on 32-bit and 64-bit Windows. On most 64-bit Unix:es and Linux, "long int" are 64 bits as opposed to 32 bits.
The big difference is that pointers are 64 bits instead of 32.
Pointers and 64-bit integers are also 8-byte aligned instead of 4-byte aligned. (at least on Linux they are)
It may not seem much but if you have large arrays of structs with these types, it could add up.
Meanwhile, in my small country, brain tumours are on the rise. Cell phones seem to be more popular here than in the US. The cities are more densely populated.
Anyway, I don't think the risk is so much from cell phones themselves.
I would be more concerned for people who are living or working too close to a cell phone tower. In the latter case, it is not just the frequency band that the cell phone is using, when it is being used. That tower is on all the time and could be in 2G, 3G, 4G and 5G all at once and/or long-range high-speed directional microwave links between towers instead of fibre.
That's not how it works.
The electromagnetic waves from mobile phones do not cause the initial mutation that causes a cell to turn into a cancer cell and multiply out of control. That fact has been long established and is not contested here.
What microwaves in the frequency ranges used by cell phones have been shown to do is to promote the growth of existing cancel cells.
Mutations into cancer cells are actually not as uncommon as most people think. All of us have had cancer cells many many times, but what normally happens is that the immune system detects the microtumour and kills it long before it has grown large enough to cause a serious problem.
All of us have been exposed to cosmic radiation, practically all of us have had X-rays a few times in our lives, all of us have inhaled and ingested carcinogenic chemicals from car exhaust, pollution, first- or second-hand tobacco smoke etc. etc.
Carcinogens are all around us, but you don't get cancer that easily.
Cancer risk is about probabilities: many different factors are at play.
And radio frequencies in the part of the microwave spectrum, at frequencies and levels that mobile phones use - have been proven in several studies (I have heard of two beside the California study) to be one of those factors.
DAB radio does not provide more channels and better quality.
It provides the option between more channels or better quality: pick one!
And we all know what gets picked every time.
Shouldn't the flag icon in the title be at half mast?
I agree. Mara Jade even had an official model that used to show up at sci-fi conventions, as did many of the book authors.
The problem with the EU is that were quite a few things in the EU that was of questionable quality, and many things even contradicted each other or contradicted facts in the prequel movies.
At a convention in summer of 2013, Lucasfilm announced that they were going to sort the EU out. They were going to pick the best bits and put the main storyline into a new consistent canon and scrap the rest. (I was there in person and heard it directly from Leland Chee's and Pablo Hidalgo's lips) ... Lucasfilm announced that they were instead going to scrap it all and relabel it as "Legends". They still print and sell new copies of those books though - with a new "Legends" label on the cover.
In early 2014, however
Dave Filoni (showrunner of Clone Wars and Rebels) did pick a lot out from the old EU to use in his series. Grand Admiral Thrawn in Rebels is one of those. The Nightsisters of Dathomir is from the EU also. ...
Darth Maul's comeback with robotic legs is not even from the EU, but from an alternate universe story where he shows up on Tatooine, wanting revenge on Obi-Wan for having been cut in half on Naboo some thirty years earlier
The two main complains I have seen about The Last Jedi have been:
* It contained too much comedy. It reminded them too much of a Saturday morning cartoon than a Star Wars flick.
That's perfectly fine for a movie -- if that is the general tone of the franchise but this is the eight movie in a series and that would in many older viewers' eyes require it to follow an established format or it would get close to jumping the shark.
For the casual viewer, however, and for most reviewers -- who don't view the movie more deeply in context -- this does not matter at all.
* At the end, not much has been accomplished. The state of the universe is pretty much as it was when the movie started.
I don't agree with that though.
* It is too long and could have been shortened. This movie is two and a half hours where as every other Star Wars has been around two hours.
Unfortunately, distracted drivers tend to remove innocent passengers and pedestrians from the gene pool as well.
10K is not 16:9 but an ultra-wide variant of Ultra HD. 10240 * 4320.
There is no such thing as 16K, yet. ... which is 15K ! 4 * 3840 = 15 * 1024.
And if someone tells you they have 16K then they probably have only four times Ultra HD
Usenet in its prime was more or less an elitist forum - for academics at universities and large corporations. The less fortunate computer nerds used to dial up BBS systems, which had organised its own FidoNet for inter-BBS communication.
Usenet was not succeeded by Facebook. Facebook was made for the less tech-savvy. ... or even SlashCode.
When Usenet became too spammy, interest groups instead went back to the BBS model, but on the web: specialised forums that they themselves had control over, often running PHPBB
Try to search Youtube for "Kunt and the Gang Sexy Kids" and you won't find his duet with Jimmy Saville's ghost.
Kids at a birthday party have been shown to be in a different mood if served sweet soft drinks than if they had been served tepid tap water.
BTW, the microkernel at the bottom of Fuchsia has changed name from Magenta to Zircon.
Maybe because there's a Magenta Linux... I dunno.
Zircon has security based on capabilities (which it calls "handles", rightly so IMHO) for pretty much everything. This could support sandboxing of new sub-processes that you own, but it lacks revocation of rights from running processes that would be used as services -- which I find to be a serious omission. ... and Linux is demonstrably faster than Mach. seL4 would have been a better foundation IMHO.
IPC is very much like Unix domain sockets: with streams and queued asynchronous message passing... which means that it is never going to be faster than Mach
It's actually the opposite. Cell phone speakers these days are so small that overdriving them is considered normal. The DSP in the audio codec/amplifier is supposed to compensate enough to avoid artefacts and to avoid damaging the speaker but it is walking a tight line.
I was told this by engineers at Cirrus Logic, which makes audio chips for Apple and other cell phone manufacturers.
DRM servers have killed the used games market on PCs.
There are lots and lots of games out there that are over a decade old even that you can't buy used because they have been "activated" on an old DRM server and can't be reactivated on another machine.
In several cases, legitimate copies of a game can't be played at all on any PC because the game had demanded to contact a now discontinued DRM server even to start.
Studies made in þe olde typewriter era have shown that having a wider surface on top of the keys leads to more unintentional key presses.
The standard surface width is 1/2" or 12 mm, and with standard width (what is usually meant when talking about "key spacing") being 3/4".
I think that what Rock21k is actually referring to is what is called "island keys" or "chiclet" keys.
I don't think that whether the keys' skirts are angled or go straight down matters that much. The problems are rather that chiclet keyboard tend to have flatter surfaces but more often entirely flat, wider surfaces and less key travel than other keyboards.
MacBook "Pro"'s keyboards with its ultra-flat "butterfly" scissor mechanism is especially bad.
Also, some popular chiclet laptop keyboards (such as MS Surface "Type Cover") have very low surface friction, so fingers slip more often.
Low surface friction wouldn't have been so detrimental to keyboard feel if the keys had been dished and had more space between them.
Older keyboard keys tended to be made of plastic, such as ABS or PBT which has good surface feel even when glossy but backlit keys tend to be painted with a slippery paint layer with laser-ablated legends.
... and PCID is not an instruction. The feature means that there is a "process ID" tag on each entry in the TLB to avoid having to flush them unnecessarily.
The intended benefit is that all entries would not necessarily have to be reloaded from page tables in RAM (or cache) whenever there is a context switch.
"Tagged TLB"s have been available on other CPU architectures for decades -- and have been used by the Linux kernels for those architectures. The feature is pretty recent on Intel x86 CPUs though.
Correct me if I'm mistaken but I think AMD's x86 CPUs do not have PCID specifically but has support for "virtual machine ID" tags on the hypervisor's second-level TLB.
... and I have come across a rare dialect of Haskell named "H" intended for embedded systems, but there may of course be more.
I would say that before C++11, all C++ code could be compiled into the equivalent C code without loss of performance .. which is also how the very first C++ compilers worked.
In C++11 came move-semantics and better synchronisation primitives that C did not have. In C, you may be able to use intrinsics for synchronisation and atomic access though but move-semantics even for simple value types can improve low-level performance.
While I am also primarily a C++ programmer, I do mostly agree with the AC.
Most of the so called "idioms" in C++ are what would be called "hacks" in any other programming language.
C is more readable than C++. In C++, you often have to spend more mental energy trying to find out what exactly is going on in other people's code.
For instance, because what looks like one type of call might actually be something else involving operators that takes place between types higher up in the class hierarchies, and what looks like value-semantics may actually be pointer semantics because references are used.
You are often completely dependent on the IntelliSense (or whatever it is called) in the code editor you are using to not get lost.
Next, shared_ptr needs multiple objects for a friggin pointer and C++ lack the concept of "borrowing" which I would say is absolutely essential in a language with unique pointers.
... and the "D" you linked to is only one of several C-like languages named "D".
There are also multiple languages each named "E" and "G".
Indeed. I have also tried to add an article about a new subject only to find that it already exists as a rejected and locked page because someone at one time several years ago misunderstood what it was and deemed it "not notable". Also, with no possibility of changing it or adding to the "talk" page.
The oldest product that I have heard of was the Neonode phone, first exhibited in 2002. It had vertical slide-to-unlock, albeit with three different sliders (left, centre, right) that activated three different functions.
The Neonode wasn't from Microsoft but it ran Windows CE underneath its own GUI.