I'm a professional in the business and I was really happy to see that they seem to have gotten everything right! I was prepared to roll my eyes when they showed a cross-section of a bipolar transistor (which they didn't) and their treatment of BEOL processing was outstanding.
Given it was AnandTech, it was expected to be of high quality and well-researched.
While AnandTech is primarily a PC oriented enthusiast website, Anand Lal Shimpi (when he was there) ensured that the content there would be extremely well researched and accurate. The speciality of the site is internals. Want to know how an SSD works? Try any other site. Want to know how the SSD controller works with details on the FTL and how TLC memory works in detail? Anand will tell you.
Not every tech website decided to go with flashy over actual technical content.
Furthernore, having a router that cannot handle that many TCP connections is kind of broken. I'm using a Linux PC as the LAN server/router, and you can blast around what you want, have 10K NATed TCP connections and everything works fine. The cable company's provided "router", OTOH, does not even handle long running ssh connections (especially when they go idle for periods) without any torrent traffic properly. Worse, it does not even send a RST packet, so your local ssh client thinks everything is fine till it tries to send something,...
Depends on the router.
Home users opt for the cheapest of the cheap specials which may have stunningly low connection counts (256 isn't uncommon).. Higher end ones may have 1024, 2048, 4096 or so.
The best of the best now are basically unlimited (I think Linux's maximum) because those have gobs of RAM (128-256MB) to hold all the information down.
It's not just TCP sessions holding entries in the table - UDP ones too. Even if you have 10K entries, connecting to a rather large swarm could easily use up all 10K sessions as your BT client tries to do UDP sessions (which often only time out the connection after 5 minutes). So the table gets bloated with UDP sessions and doesn't clear out.
One common test I've seen people do is start a game of Counter-Strike. Just list all the servers and see if your router collapses as it tries to ping each and every server. Most cheap routers die or reboot.
There is an electric race car that puts out something like 830 HP and 2950 ft. lbs. of torque. I can't even imagine what that kind of power feels like behind the wheel. Electric cars are exciting to gearheads like me who enjoy performance driving. I can't wait until they become more affordable.
The well-known electric racing circuit is the Formula E which uses pure electric race cars. Now, they only last about 10 minutes before drivers have to pit and switch cars, but that just adds a bit more excitement to the mix (how fast you can egress and get in now becomes important, just like how long you spend at pit spots in regular auto racing).
Though, the other thing is just how quiet it is - yeah, I know modern race cars are actually getting a lot quieter to improve mileage (sound energy is wasted energy) and lengthen times between pit stops for refuelling.
Heck, a lot are starting to experiment with hybrid technology for the same reason - pit stops cost time, and if you can go just as fast but use less fuel, then you have a big advantage by skipping a 30-second pit stop (plus having to actually drive through pit row - there's a 60mph speed limit that's strictly enforced. There's a special button on the wheel for this where it limits the max speed to that).
Heck, Formula E has people driving in interesting ways - is it better to be slower and prolong your battery, or go quick and get a sufficient lead for the swap?
And given the low end torque, skill becomes important because wheels that' slip, while impressive, are wasted energy that could be better spent moving.
Exactly, it puzzles me why it not left in officially. Performance issues maybe? A non optimal gui? Or a sinister conspiracy?
1) Extra thing to QA (and apparently there are issues with UIs being unreadable and sequences where they move the camera causing motion sickness)
2) Perhaps they fell for the hype that the Rift would be out for consumers by now when they started coding the engine years ago. Given how few have one, the added support costs couldn't justify the feature (low ROI).
3) With so many competitors coming on the market ahead of the Rift, perhaps there's incompatibility with them
I think what Ubisoft is trying to say here is that it's programmers are shit, and so is it's game engine (AnvilNext). Other publishers manage to do okay, but poor old Ubisoft are stuck with this turd and can't just switch to Unreal or something more competent, so sorry guys you only get 900p on consoles.
I mean, obviously if you are CPU bound the solution is to reduce the load on the GPU by making the game render at 900p. The problem is the AI for the large number of characters in the game, so clearly reducing the pixel count will help with that.
Exactly.
Not only that, but whare the specs going to be on the PC? A quad-core i7 minimum to have the necessary number of threads and processing power? That will exclude a LARGE number of people who have more modest machines.
Yes, the Xbone and PS4 can't hold a lick towards the latest and greatest PC available. But a good number of people don't have PCs as powerful as an Xbone or PS4. Sure, maybe in 10 years time the complaint is valid, but for most PCs out there, the Xbone or PS4 is more powerful.
Oracles control to undo the GPL license Java is under.
Oracle gave a free license for J2SE, whereas Android and mobile devices use J2ME (which costs $$$ and is the money-maker side of Java - Oracle gives squat about J2SE or J2EE because they aren't profit centers).
And it's really a patent license - that as long as your implementation is J2SE compatible, you're good.
Of course, if it's really about whether APIs can be copyrighted, this can have far-reach decisions, because it places a bunch of GPL'd stuff on the line. E.g., in the Linux kernel, there are a bunch of utility functions that are exported to GPL-only kernel modules (EXPORT_SYMBOL_GPL). If copyright doesn't apply, then GPL protections can't apply either (since GPL requires copyright in order to function - the GPL grants you rights if you agree to terms, if you don't agree, you agree to standard copyright). And the kernel devs have looked down on proprietary drivers deliberately working around limitations to call those functions.
GPL'd libraries are fine - since the library implementation is GPL'd. But it also means that someone else can use the same APIs and make a non-GPL'd version of the library for proprietary code, like say a non-free version of readline.
It's a real shame too. Used to be that you could buy a game and then download a clean installer without the DRM malware in it, and enter your code from the retail packaging. Now you have no choice but to accept the malware, even on a legitimate purchase. Well, you can chose not to play of course.
I thought this died out now that everyone's using Steam. You could buy a retail box which is just an installer that copies the bulk of the data to your hard drive (saving you a download), but in the end it was just a retail code you entered into Steam to redeem.
(Which, when you think about it, makes Steam even worse overall than what the Xbone was originally supposed to ship with - at least the Xbone was supposed to allow you to share a game with up to 10 other people, and the ability to transfer the license. Steam lets you share your entire library with someone else but not an individual game, and doesn't have a license transfer mechanism yet ("used game sales").).
I was excited to see a good discussion about software/hardware required to route 1 Gbps. How many people are running pfsense or Linux router with Google Fiber? What do people have for wireless? I want to separate my router and my AP. Consumer routers never seem to actually perform up to theoretical speeds or have problems with a large number of clients.
SmallNetBuilder routinely tests routers in simulated conditions to get their best speeds, so it's not just theoretical, but what they could obtain with their test equipment.
And the modern high end routers like the Nighthawk or the RT-AC68U (Asus) ARE Gigabit capable - they have sufficiently powerful CPUs (multi-core, 800MHz) and RAM (128MB+) to actually be able to route and NAT that fast.
Every viral disease considered to be airborne spreads through droplets. They don't fly around the air like birds. Chickenpox, smallpox, and the flu are all considered airborne diseases
Coughing up blood on someone isn't airborne. Sneezing or coughing on them is. If you can catch Ebola this way, then it is airborne. They are probably saying it's not just to keep the panic level down.
Airborne transmissions include viruses that live on surfaces. Ebola (like HIV) has proven to be too fragile to live on surface for any length of time - a sick patient may cough and sneeze all over a table, but once the mucus dries out, the genetic material breaks down because the shell isn't resilient enough to hold it together.
Yes, this includes if the virus lands on airborne dust - it dries out and becomes ineffective.
The thing is that "airborne" transmissions occur when someone is ill and passes through a space, and then someone else passes through the same and gets ill because of it. E.g., just because no one around you is coughing doesn't mean you can't get the cold by touching an infected surface (the cold and flu viruses CAN survive on surfaces long after they dry out - their protective shells are far more resilient).
If a patient is coughing and sneezing, you already take precautions by putting on protective gear. But airborne is in the case where the sick patient is removed, and the virus is still around, as if it was hanging in the air and you get sick merely by sharing the same space.
AIrborne viruses are far more infectious because of it - the only way to get rid of the threat is extensive cleaning. While non-airborne transmission methods require a failure of taking preventative measures in order to be infected. Remove the sick person and the disease cannot be transmitted to someone who uses the space next.
It's why the cold and flu are extremely infectious - it just takes someone touching an infected surface (easy) then transferring that virus into their body (easiest way is through the eyes when the person goes to rub them, but a cut will also do).
I bet that software pirates already have injected malware in many warez, mainly heavy graphics games. Doing so they could discreetly control a lot of powerful machines.
No, the software itself isn't infected with malware, actually. What happens is they infect the keygens or cracks. This is because most software applications are actually signed, as are installers, so they don't bother infecting that - they distribute the original installers with all the original signatures intact.
But since to use it requires running the crack executable to get the key, well, the user will just double-click it, get their machine infected, and the key to unlock the program they just installed.
And it's been happening a long time - it's why cracks and keygens are long tagged by AV apps - because while there are a few clean cracks and keygens, you can bet most you find on torrent sites and elsewhere are infected.
Whale: Me no likey fishy. Me likey seals. You likey fish? capiche?
Dolphin: Shit, where's this guy from?
In BC, we have two kinds of orcas. The "resident orca" population actually eats fish (salmon) most of the time. The others are "transient" and they are the ones that eat seals.
Both are orcas, but they have completely different diets.
Dolphins though generally eat smaller fish.
OTOH, one wonders if orcas do it to eat dolphins. (They aren't called "killer whales" for nothing).
Seals have also been known to seek the sanctuary of whalewatching boats while escaping a bunch of hungry orcas.
There would be political opposition to anything like that, and some will truly believe not having a corporation making obscene profits and being entrenched monopolies would be immoral.
My guess is, the same people who oppose socialized medicine, would disagree on the same premise. Because they somehow feel society is best left to rot as long as they've got their pile of money.
Of course. There's lots of people that believe if you impose government regulations, it'll cost jobs. If a giant multinational corporation doesn't get the tax breaks it wants, it immediately threatens layoffs. It's led to people believing if the companies can't make (as much) profit, they're going to lay off the entire workforce.
Of course, large companies generally have the worst job creation numbers - the biggest employers are small business where the owners generally work very hard trying to grow their company.
No, it's Slashdot. You have to put é to get é.
I have no idea if beta is similarly broken, but I wouldn't be in the least bit surprised.
Slash (and/.) do support Unicode. The problem was a bunch of trolls abused it to the point where the only solution is to implement a WHITELIST of Unicode codepoints.
If you want to see this and many other comments, search google for strings like "5:erocS" or "erocS". Slashdot broke the displaying of those comments ages ago, but the legacy is still indexed.
(Hint: think what "erocS" might be used on/.. It's used in combination with some Unicode codepoints... )
Has it occurred to you that [users of software] simply don't care?
Has it occurred to you that publishers of software simply don't care about the needs of some of their users? That's why free software is so important: it gives you the flexibility to hire anyone to make a program do what you want.
Some is not a lot of people. It's probably not big enough for publishers to care, nor cater specially for - because it just costs more money to do than they'll get back (ROI).
It's like why PC ports of AAA games are so shitty - because the ROI on PC is poor. There may be more PC users, but there's also a ton more piracy, so what sold well on consoles may barely make up the porting costs on PC.
Free software's great, but I'm sure most people using it are just using it "just because" and less because "it's freedom!". I.e., their job involves developing for Linux, so they have a Linux PC. Doesn't mean they care about free software, it's just that Linux is paying the bills so they'll use it. Heck, some probably debug without realizing the library they're debugging through actually has source code .
I suppose the joke is that in the 80s, 8.2mpg was considered shitty gas mileage while we fast forward and that's what a lot of people are getting and they're happy for it, Or owning vehicles that get less than that.
1 litre / kilometre = 1 sq millimetre
That is another win for the metric system in my book.
Or drop the prefixes. so 1 litre/meter is 1 square meter, keeping all the units in the base unit without prefixes.
(Yes, it's probably the worst part of metric is when people use base units. Like 1Mg (1 megagram) to mean 1000kg. Yes, the former is more correct, but 1000kg is more "intuitive" when you want to talk about the tonne)
I have a GTX Titan, and I *miss* my old work 17" MacBook Pro laptop -- the bigger screen was _nice_. While I appreciate the new slim MacBook Pro and the high resolution Retina brings I _also_ want the fastest nVidia GPU available so I continue to do CUDA research, aka "Big Data on the GPU." when I'm not at home. Current mobile GPU's suck for performance, but they are _finally_ decent.
Why not use the fact that you have two external PCIe connections and well, use it with an external PCIe enclosure?
(The fact that modern laptops expose internal buses should be very exciting - how many times have you wished for "real" peripherals like serial ports and the like instead of having to live with lame USB versions? Well, a Thunderbolt adapter is just as real as one plugged into the PCIe bus. Granted, serial and parallel ports are usually on the LPC bus, but there are bridge chips to PCIe).
If you have access to the ATM physically, why not just take the cash there and then?
Because it's easier to get to the electronics than the cashbox.
Inside these little ATMs is a steel box. Get that steel box open and you have full access to the electronics. But to get to the cash requires opening said box, then opening the safe holding the cash, which is vastly more protected.
The cash is dispensed from within the safe and exits out a slot in the safe (basically the safe carries a number of cash cassettes and the electronics count out the cash, which is why if they mis-load the cassettes, you can be short changed or given more than you expect.
Oh yeah, and the safe has all sorts of safeguards to destroy the cassettes should they be tampered with, making it even harder to get the cash out.
Of course, they assumed the electronics were secure, so the other way to get the cash out is via the front door. Bypasses all the safe security systems and everythign else.
1. Almost all auto dealerships are franchise operations. Tesla owns theirs outright. 2. Most used cars are not sold through a dealership.
Private sales and 3rd party dealerships are common.
Even when somebody trades in used vehicle for a new one it rarely shows up on that dealer's lot. Normally it goes to auction where some other used car dealer buys it up.
Is this a subtle assault on the 1st sale doctrine, keeping the cars within Tesla's system?
I don't see anywhere where it blocks the sale of Teslas by other people. In fact, right now, you can buy used Tesla Model S', but they cost MORE than from the factory.
If you wanted to trade your Model S at a dealer to buy a new Ford or whatever, you certainly can do it. No one said you have to sell your used Tesla back to Tesla.
And lots of dealers have "certified pre-owned" programs - walk into any luxury marque's lot and if you can't afford new, they'll show you their certified pre-owned ones. Lots of dealerships have pre-owned sections of their lots - they may auction off the off-marque cars they took in, but they probably just service their own and put it up for sale.
Cancel their H1B's and 90+% of their workforce (i.e., income) disappears. I'm not sure any company could survive an overnight 90% drop in revenue.
That has secondary side effects in companies that hired the workers.
No, what you do is every H1B they hired is converted into permanent residency or citizenship, sponsored completely by Infosys. Which means even if the company fires the employee, he's still on a permanent resident/citizenship track except now Infosys is responsible for his well-being. Living expenses, housing, etc, like if an American was sponsoring someone.
Get yourself plenty of serial console servers. 48 port ones are normal, so maybe you need 5. Then you can have specific places where each one goes. People go in once place, equipment goes in another. There's nothing worse than working next to hardware.
It's not just having to work next to the noise, it's the lack of space to work - you're plugging your laptop in and it's precariously balanced on something and you need to reach for something else, etc. etc. then OOPS! You bump the laptop, and it jerks the wrong cable and breaks something.
Yes, sometimes you have to do it, but if it's a misconfiguration or something, being able to do it away from the din of the machines, at a desk where you can CALMLY sit down and have reference materials open means fixing problems are less likely to cause new ones.
Trying to reconfigure a misconfigured device while standing and trying to balance your laptop, reference materials and dealing with an already stressful place (the noise), really adds nothing other than needless stress. Being able to fix it in a calm, quiet environment means you're more likely to do it right the first time rather than trying any solution to get you out of the hellhole.
If we could get our homes switched to DC, then we could have these without the extra electronics in them.
Unlikely to happen as it would cost more.
Low voltage DC throughout the home would mean high-currents are required - take say, a 12W LED bulb. If you wanted to run it direct, that would mean around 3V at 4A. Wire 4 up in parallel and that's 16A, which means your cabling in the house is just as thick as it is now. Get 10 bulbs and that's 40A, which requires jumper-cable style thickness of wires.
Wire thickness dictates its ampacity (how much current can be carried for a given temperature rise).
It's cheaper to actually run a higher voltage and use a converter to the lower voltage you need because you're not carrying as much current so you can use thinner cables (cheaper, easier to install) and potentially more efficient because of IIR losses (which goes up with the square of the current).
From the sound of it the rest of the phone is okay but they might as well have put on broken screens by how much they crippled their phone
Have you read the review? The screen is HORRIBLE. And not horrible as in low-res (it's that, but that can slide). It's horrible in that it's literally the cheapest POS on the market - given how narrow a viewing angle It has (landscape mode is almost unusable because the colors shift - it's that narrow). Heck, it reminds me of the old PASSIVE displays of ancient times which had the same issues.
To be honest, perhaps they could've saved the effort and just put in a monochrome LCD instead...
Valve is the sole reason why I'm probably not buying a console this generation, or maybe ever again. Gaming is so much better on PC these days that it just doesn't make sense to lock yourself into the console market anymore. And that's all on Valve.
And yet when the Xbone tried to do a Valve Steam (but with a few improvements) it was soundly killed off. I mean, the original Xbone inspiration was Steam - you could buy a game on disc, load it into your console and that's it. It was tied to your account, just like Steam. Though, Microsoft did add the ability to move games from your account to someone else's.
Funny, that. No one considers PC gaming to be "draconian" when Valve does it, but when Microsoft does it...
SteamOS is interesting, but I hope it doesn't take off, because it'll hold back PC gaming even worse than consoles when people expect their $500 i3-and-intel-graphics SteamBoxes to still play games 10 years later.
Apple places all their bets on the magic chip, and who knows if there is a backdoor in that. The chips are fabbed in China, and it would be trivial to add extra functionality to the mask. In fact, since the entire chip is designed in China, baking in added, undocumented "features" like password recovery or even a vector for bypassing signed code is entirely possible, and likely in place.
Actually, the chips are designed in the US (Apple has acquired numerous ASIC designers and companies) and fabbed in either the US (Samsung - Texas) or Taiwan (TSMC). It's a part of the A7/A8 processor. In fact, given Apple's fab choices, most of the have been fabbed in either Korea or the US (Samsung's SoCs for the iPhone, iPhone 3G, iPhone 3GS, Apple's design since the iPhone4 using Samsung's fabs).
In fact, if you're claiming Apple's SoCs are vulnerable, that implies EVERYONE's SoCs are vulnerable - the secure enclave is really a fancy name for secure memory and secure encryption processors which you can find in every modern SoC, be it Qualcomm (fabbed in Taiwan), Broadcom (TSMC), NVidia (TSMC), or Samsung (Korea).
And the Taiwanese may refer to themselves as Chinese, but any association with mainland China generally brings up nothing but hatred since the two countries/regions/whatever are quite separate. Fabs like TSMC are very careful because they know they're dead if there's even a hint of Chinese (mainland China) spying.
Oh yeah, Apple renewed their contract with Samsung on their fabs as well.
End result is either every phone is vulnerable (unlikely), or there really aren't any Chinese backdoors. (Because well, Pengu would be out for iOS 8 already)
It isn't a matter of "subshell" -- which gets a copy of the parent process's memory from fork() -- but how functions are exported for use by future shells. For example, your shell (bash) runs less or vi (or python or php...), and from there a shell is started. That's not a subshell, it's a new process; it wasn't created by fork() so it doesn't have a copy of the original bash memory. The logical question is why anyone would need that functionality? In general, no one does, but it does offer some optimization for complex shell environments (read: the idiotic bash_completion BS. That's why bash takes an eon and a half to load.) Yes, my first response was "turn that crap off; nobody uses it anyway."
The "namespace" patch does absolutely NOTHING to address this bug, or security in general. So I now have to name my variable BASH_FUNC_foo() (or some variant.) If inputs aren't validated, the system is still vulnerable.
No, the bug IS related to subshells. Environment variables are "exported" to make them available to subshells to use. POSIX defines the mechanism (and the environment is passed in as parameteres to the exec() family of functions)
Now, bash has a design flaw in how it exports FUNCTIONS to subshells, namely if it starts with () { then bash treats it as a function declaration.
Of course, the problem is not related to subshells - but it stems from there. Because if you invoke bash, and pass it a function declaration like that, bash will think it's inheriting a function (which could've been started many, many processes ago because exported functions in bash look like exported environment variables).
So you can do the bash-->httpd-->bash thing and if you export a function in the partent bash, even though you're going through httpd, the sub-bash (probably started as CGI) will see the parent exported function.
And CGI typically calls system() to perform the script execution, which by definition launches the system shell to run the command you passed to system().
The bug originates from a design flaw in POSIX that fails to specify how to securely pass in exported functions, so the bash devs made up some method. Shellshock basically relies on this feature.
The best way, of course, is to disable exporting of functions, but there may be many reasons why this may not be possible - including entrenched software that relies on this behavior and has for many years (it's a design flaw dating back over 20 years). And one should also note that 20-30 years ago, Unix security was practically non-existent.
Things like this were a product of a different era.
Slashdot Mirror
Given it was AnandTech, it was expected to be of high quality and well-researched.
While AnandTech is primarily a PC oriented enthusiast website, Anand Lal Shimpi (when he was there) ensured that the content there would be extremely well researched and accurate. The speciality of the site is internals. Want to know how an SSD works? Try any other site. Want to know how the SSD controller works with details on the FTL and how TLC memory works in detail? Anand will tell you.
Not every tech website decided to go with flashy over actual technical content.
Depends on the router.
Home users opt for the cheapest of the cheap specials which may have stunningly low connection counts (256 isn't uncommon).. Higher end ones may have 1024, 2048, 4096 or so.
The best of the best now are basically unlimited (I think Linux's maximum) because those have gobs of RAM (128-256MB) to hold all the information down.
It's not just TCP sessions holding entries in the table - UDP ones too. Even if you have 10K entries, connecting to a rather large swarm could easily use up all 10K sessions as your BT client tries to do UDP sessions (which often only time out the connection after 5 minutes). So the table gets bloated with UDP sessions and doesn't clear out.
One common test I've seen people do is start a game of Counter-Strike. Just list all the servers and see if your router collapses as it tries to ping each and every server. Most cheap routers die or reboot.
The well-known electric racing circuit is the Formula E which uses pure electric race cars. Now, they only last about 10 minutes before drivers have to pit and switch cars, but that just adds a bit more excitement to the mix (how fast you can egress and get in now becomes important, just like how long you spend at pit spots in regular auto racing).
Though, the other thing is just how quiet it is - yeah, I know modern race cars are actually getting a lot quieter to improve mileage (sound energy is wasted energy) and lengthen times between pit stops for refuelling.
Heck, a lot are starting to experiment with hybrid technology for the same reason - pit stops cost time, and if you can go just as fast but use less fuel, then you have a big advantage by skipping a 30-second pit stop (plus having to actually drive through pit row - there's a 60mph speed limit that's strictly enforced. There's a special button on the wheel for this where it limits the max speed to that).
Heck, Formula E has people driving in interesting ways - is it better to be slower and prolong your battery, or go quick and get a sufficient lead for the swap?
And given the low end torque, skill becomes important because wheels that' slip, while impressive, are wasted energy that could be better spent moving.
1) Extra thing to QA (and apparently there are issues with UIs being unreadable and sequences where they move the camera causing motion sickness)
2) Perhaps they fell for the hype that the Rift would be out for consumers by now when they started coding the engine years ago. Given how few have one, the added support costs couldn't justify the feature (low ROI).
3) With so many competitors coming on the market ahead of the Rift, perhaps there's incompatibility with them
4) People got sick.
Exactly.
Not only that, but whare the specs going to be on the PC? A quad-core i7 minimum to have the necessary number of threads and processing power? That will exclude a LARGE number of people who have more modest machines.
Yes, the Xbone and PS4 can't hold a lick towards the latest and greatest PC available. But a good number of people don't have PCs as powerful as an Xbone or PS4. Sure, maybe in 10 years time the complaint is valid, but for most PCs out there, the Xbone or PS4 is more powerful.
Oracle gave a free license for J2SE, whereas Android and mobile devices use J2ME (which costs $$$ and is the money-maker side of Java - Oracle gives squat about J2SE or J2EE because they aren't profit centers).
And it's really a patent license - that as long as your implementation is J2SE compatible, you're good.
Of course, if it's really about whether APIs can be copyrighted, this can have far-reach decisions, because it places a bunch of GPL'd stuff on the line. E.g., in the Linux kernel, there are a bunch of utility functions that are exported to GPL-only kernel modules (EXPORT_SYMBOL_GPL). If copyright doesn't apply, then GPL protections can't apply either (since GPL requires copyright in order to function - the GPL grants you rights if you agree to terms, if you don't agree, you agree to standard copyright). And the kernel devs have looked down on proprietary drivers deliberately working around limitations to call those functions.
GPL'd libraries are fine - since the library implementation is GPL'd. But it also means that someone else can use the same APIs and make a non-GPL'd version of the library for proprietary code, like say a non-free version of readline.
I thought this died out now that everyone's using Steam. You could buy a retail box which is just an installer that copies the bulk of the data to your hard drive (saving you a download), but in the end it was just a retail code you entered into Steam to redeem.
(Which, when you think about it, makes Steam even worse overall than what the Xbone was originally supposed to ship with - at least the Xbone was supposed to allow you to share a game with up to 10 other people, and the ability to transfer the license. Steam lets you share your entire library with someone else but not an individual game, and doesn't have a license transfer mechanism yet ("used game sales").).
SmallNetBuilder routinely tests routers in simulated conditions to get their best speeds, so it's not just theoretical, but what they could obtain with their test equipment.
And the modern high end routers like the Nighthawk or the RT-AC68U (Asus) ARE Gigabit capable - they have sufficiently powerful CPUs (multi-core, 800MHz) and RAM (128MB+) to actually be able to route and NAT that fast.
Airborne transmissions include viruses that live on surfaces. Ebola (like HIV) has proven to be too fragile to live on surface for any length of time - a sick patient may cough and sneeze all over a table, but once the mucus dries out, the genetic material breaks down because the shell isn't resilient enough to hold it together.
Yes, this includes if the virus lands on airborne dust - it dries out and becomes ineffective.
The thing is that "airborne" transmissions occur when someone is ill and passes through a space, and then someone else passes through the same and gets ill because of it. E.g., just because no one around you is coughing doesn't mean you can't get the cold by touching an infected surface (the cold and flu viruses CAN survive on surfaces long after they dry out - their protective shells are far more resilient).
If a patient is coughing and sneezing, you already take precautions by putting on protective gear. But airborne is in the case where the sick patient is removed, and the virus is still around, as if it was hanging in the air and you get sick merely by sharing the same space.
AIrborne viruses are far more infectious because of it - the only way to get rid of the threat is extensive cleaning. While non-airborne transmission methods require a failure of taking preventative measures in order to be infected. Remove the sick person and the disease cannot be transmitted to someone who uses the space next.
It's why the cold and flu are extremely infectious - it just takes someone touching an infected surface (easy) then transferring that virus into their body (easiest way is through the eyes when the person goes to rub them, but a cut will also do).
No, the software itself isn't infected with malware, actually. What happens is they infect the keygens or cracks. This is because most software applications are actually signed, as are installers, so they don't bother infecting that - they distribute the original installers with all the original signatures intact.
But since to use it requires running the crack executable to get the key, well, the user will just double-click it, get their machine infected, and the key to unlock the program they just installed.
And it's been happening a long time - it's why cracks and keygens are long tagged by AV apps - because while there are a few clean cracks and keygens, you can bet most you find on torrent sites and elsewhere are infected.
In BC, we have two kinds of orcas. The "resident orca" population actually eats fish (salmon) most of the time. The others are "transient" and they are the ones that eat seals.
Both are orcas, but they have completely different diets.
Dolphins though generally eat smaller fish.
OTOH, one wonders if orcas do it to eat dolphins. (They aren't called "killer whales" for nothing).
Seals have also been known to seek the sanctuary of whalewatching boats while escaping a bunch of hungry orcas.
Of course. There's lots of people that believe if you impose government regulations, it'll cost jobs. If a giant multinational corporation doesn't get the tax breaks it wants, it immediately threatens layoffs. It's led to people believing if the companies can't make (as much) profit, they're going to lay off the entire workforce.
Of course, large companies generally have the worst job creation numbers - the biggest employers are small business where the owners generally work very hard trying to grow their company.
Slash (and /.) do support Unicode. The problem was a bunch of trolls abused it to the point where the only solution is to implement a WHITELIST of Unicode codepoints.
If you want to see this and many other comments, search google for strings like "5:erocS" or "erocS". Slashdot broke the displaying of those comments ages ago, but the legacy is still indexed.
(Hint: think what "erocS" might be used on /.. It's used in combination with some Unicode codepoints... )
Some is not a lot of people. It's probably not big enough for publishers to care, nor cater specially for - because it just costs more money to do than they'll get back (ROI).
It's like why PC ports of AAA games are so shitty - because the ROI on PC is poor. There may be more PC users, but there's also a ton more piracy, so what sold well on consoles may barely make up the porting costs on PC.
Free software's great, but I'm sure most people using it are just using it "just because" and less because "it's freedom!". I.e., their job involves developing for Linux, so they have a Linux PC. Doesn't mean they care about free software, it's just that Linux is paying the bills so they'll use it. Heck, some probably debug without realizing the library they're debugging through actually has source code .
I suppose the joke is that in the 80s, 8.2mpg was considered shitty gas mileage while we fast forward and that's what a lot of people are getting and they're happy for it, Or owning vehicles that get less than that.
Or drop the prefixes. so 1 litre/meter is 1 square meter, keeping all the units in the base unit without prefixes.
(Yes, it's probably the worst part of metric is when people use base units. Like 1Mg (1 megagram) to mean 1000kg. Yes, the former is more correct, but 1000kg is more "intuitive" when you want to talk about the tonne)
Why not use the fact that you have two external PCIe connections and well, use it with an external PCIe enclosure?
(The fact that modern laptops expose internal buses should be very exciting - how many times have you wished for "real" peripherals like serial ports and the like instead of having to live with lame USB versions? Well, a Thunderbolt adapter is just as real as one plugged into the PCIe bus. Granted, serial and parallel ports are usually on the LPC bus, but there are bridge chips to PCIe).
Because it's easier to get to the electronics than the cashbox.
Inside these little ATMs is a steel box. Get that steel box open and you have full access to the electronics. But to get to the cash requires opening said box, then opening the safe holding the cash, which is vastly more protected.
The cash is dispensed from within the safe and exits out a slot in the safe (basically the safe carries a number of cash cassettes and the electronics count out the cash, which is why if they mis-load the cassettes, you can be short changed or given more than you expect.
Oh yeah, and the safe has all sorts of safeguards to destroy the cassettes should they be tampered with, making it even harder to get the cash out.
Of course, they assumed the electronics were secure, so the other way to get the cash out is via the front door. Bypasses all the safe security systems and everythign else.
I don't see anywhere where it blocks the sale of Teslas by other people. In fact, right now, you can buy used Tesla Model S', but they cost MORE than from the factory.
If you wanted to trade your Model S at a dealer to buy a new Ford or whatever, you certainly can do it. No one said you have to sell your used Tesla back to Tesla.
And lots of dealers have "certified pre-owned" programs - walk into any luxury marque's lot and if you can't afford new, they'll show you their certified pre-owned ones. Lots of dealerships have pre-owned sections of their lots - they may auction off the off-marque cars they took in, but they probably just service their own and put it up for sale.
That has secondary side effects in companies that hired the workers.
No, what you do is every H1B they hired is converted into permanent residency or citizenship, sponsored completely by Infosys. Which means even if the company fires the employee, he's still on a permanent resident/citizenship track except now Infosys is responsible for his well-being. Living expenses, housing, etc, like if an American was sponsoring someone.
It's not just having to work next to the noise, it's the lack of space to work - you're plugging your laptop in and it's precariously balanced on something and you need to reach for something else, etc. etc. then OOPS! You bump the laptop, and it jerks the wrong cable and breaks something.
Yes, sometimes you have to do it, but if it's a misconfiguration or something, being able to do it away from the din of the machines, at a desk where you can CALMLY sit down and have reference materials open means fixing problems are less likely to cause new ones.
Trying to reconfigure a misconfigured device while standing and trying to balance your laptop, reference materials and dealing with an already stressful place (the noise), really adds nothing other than needless stress. Being able to fix it in a calm, quiet environment means you're more likely to do it right the first time rather than trying any solution to get you out of the hellhole.
Unlikely to happen as it would cost more.
Low voltage DC throughout the home would mean high-currents are required - take say, a 12W LED bulb. If you wanted to run it direct, that would mean around 3V at 4A. Wire 4 up in parallel and that's 16A, which means your cabling in the house is just as thick as it is now. Get 10 bulbs and that's 40A, which requires jumper-cable style thickness of wires.
Wire thickness dictates its ampacity (how much current can be carried for a given temperature rise).
It's cheaper to actually run a higher voltage and use a converter to the lower voltage you need because you're not carrying as much current so you can use thinner cables (cheaper, easier to install) and potentially more efficient because of IIR losses (which goes up with the square of the current).
Have you read the review? The screen is HORRIBLE. And not horrible as in low-res (it's that, but that can slide). It's horrible in that it's literally the cheapest POS on the market - given how narrow a viewing angle It has (landscape mode is almost unusable because the colors shift - it's that narrow). Heck, it reminds me of the old PASSIVE displays of ancient times which had the same issues.
To be honest, perhaps they could've saved the effort and just put in a monochrome LCD instead...
And yet when the Xbone tried to do a Valve Steam (but with a few improvements) it was soundly killed off. I mean, the original Xbone inspiration was Steam - you could buy a game on disc, load it into your console and that's it. It was tied to your account, just like Steam. Though, Microsoft did add the ability to move games from your account to someone else's.
Funny, that. No one considers PC gaming to be "draconian" when Valve does it, but when Microsoft does it...
SteamOS is interesting, but I hope it doesn't take off, because it'll hold back PC gaming even worse than consoles when people expect their $500 i3-and-intel-graphics SteamBoxes to still play games 10 years later.
Actually, the chips are designed in the US (Apple has acquired numerous ASIC designers and companies) and fabbed in either the US (Samsung - Texas) or Taiwan (TSMC). It's a part of the A7/A8 processor. In fact, given Apple's fab choices, most of the have been fabbed in either Korea or the US (Samsung's SoCs for the iPhone, iPhone 3G, iPhone 3GS, Apple's design since the iPhone4 using Samsung's fabs).
In fact, if you're claiming Apple's SoCs are vulnerable, that implies EVERYONE's SoCs are vulnerable - the secure enclave is really a fancy name for secure memory and secure encryption processors which you can find in every modern SoC, be it Qualcomm (fabbed in Taiwan), Broadcom (TSMC), NVidia (TSMC), or Samsung (Korea).
And the Taiwanese may refer to themselves as Chinese, but any association with mainland China generally brings up nothing but hatred since the two countries/regions/whatever are quite separate. Fabs like TSMC are very careful because they know they're dead if there's even a hint of Chinese (mainland China) spying.
Oh yeah, Apple renewed their contract with Samsung on their fabs as well.
End result is either every phone is vulnerable (unlikely), or there really aren't any Chinese backdoors. (Because well, Pengu would be out for iOS 8 already)
No, the bug IS related to subshells. Environment variables are "exported" to make them available to subshells to use. POSIX defines the mechanism (and the environment is passed in as parameteres to the exec() family of functions)
Now, bash has a design flaw in how it exports FUNCTIONS to subshells, namely if it starts with () { then bash treats it as a function declaration.
Of course, the problem is not related to subshells - but it stems from there. Because if you invoke bash, and pass it a function declaration like that, bash will think it's inheriting a function (which could've been started many, many processes ago because exported functions in bash look like exported environment variables).
So you can do the bash-->httpd-->bash thing and if you export a function in the partent bash, even though you're going through httpd, the sub-bash (probably started as CGI) will see the parent exported function.
And CGI typically calls system() to perform the script execution, which by definition launches the system shell to run the command you passed to system().
The bug originates from a design flaw in POSIX that fails to specify how to securely pass in exported functions, so the bash devs made up some method. Shellshock basically relies on this feature.
The best way, of course, is to disable exporting of functions, but there may be many reasons why this may not be possible - including entrenched software that relies on this behavior and has for many years (it's a design flaw dating back over 20 years). And one should also note that 20-30 years ago, Unix security was practically non-existent.
Things like this were a product of a different era.