The only reason they would want to expand their foundry business would be if it was hugely profitable, and offering Apple even bigger discounts than they already were getting would make it LESS profitable. Further, Samsung already can handle Apple's total chip requirements, so this wouldn't involve an expansion at all.
Samsung might not want to IDLE any of their foundries by losing Apple business, but with Android sales surging to 70% market share world wide, there is little risk of that having any long term effect.
Fabs are expensive. If they're not running 24/7/365, they're costing money. It takes billions to open a fab, only to have in 18 months having to pay out a significant chunk of that money AGAIN to buy all new fab equipment because technology marches on.
It's why there are few people with fabs around the world - Intel, Global Foundries (AMD), TSMC, Samsung, Toshiba and a handful of others.
Oh yeah, each fab is different! The one optimized for making RAM isn't as good for making CPUs, the one making CPUs isn't as good for making RAM, and Flash lines just don't do well for either as well. Each kind of device requires special processing without which makes it suboptimal. (It's why embedded flash or embedded RAM just is never as good - or we'd be sticking RAM and Flash on die already).
So one fab becomes many and each needs new sites to accommodate new nodes (the older nodes are still useful for other things to help amortize the costs).
As for Android? Most of that 70% marketshare are for crap-droids with piss-poor screens, anemic processors, and RAM under 1GB. The flagships, even popular ones like the SGS3 make up only a tiny proportion of actual sales (the SGS3 is probably around 10% of the entire Android market). Most of that 70% goes to other fabs like TSMC (Qualcomm, MediaTek, etc), and will contain whatever crap is on sale that week.
It's why Intel sells foundry services to small companies - it helps top up their business, and also ensures they don't monopolize the fab. Fabs are just very capital intensive to run, and unless you can keep it going with parts continually, become a money sink.
that is going to be a bit difficult , seeing as the traffic IS ENCRYPTED Not strongly , but still encrypted .
yes in a few weeks you can decrypt it to see what is what , but by then a few weeks have past there is no way to "filter" it IN REAL TIME
Only while it stays within the network.
One thing that makes Tor more popular than other darknet systems is that it connects to the regular internet. Sort of how VoIP took off once call gateways to the POTS network were established. Because before that, well, you're just talking amongst yourselves and that's it.
Of course, once you hit the public internet, your traffic is back in plain text that can be logged, recorded and analyzed by the owners of the exit node. (And a lot of traffic contains personal information - like say, bittorrent), as well as being a potential MITM if you're trying to do an SSL connection over Tor.
"The fact Linux is awesome and Linus is an abusive and profane manager doesn't mean the profanity and abuse is necessary to make Linux awesome."
No. But the fact that Linus is sometimes abusive, plus the fact that he thinks that sometimes he has to be abusive, plus the fact that he leads no less than Linux as a testament to his management abilities does mean something.
He has a theory, he practices it and he has success backing him. You have a theory and... what else?
So, basically, the Steve Jobs management model strikes again. First at Apple, and again for Linux.
Of course, the thing is, the Steve Jobs method of being an asshole only works for certain people with certain personality traits. For Jobs, it was his charisma and RDF. For Linus, it's practically hero worship - after all, he created a very useful piece of software.
Of course, there's also a whole pile of people who are put off by such tactics - after all, having physical violence threatened or being cursed out can make one question their motivations. For there are legions of people who were put off by Jobs' assholeishness and many who quit after being chewed out by Jobs (and these people were, for the most part, leaving dream jobs that paid decently - they just couldn't stand the verbal abuse Jobs would dish out).
Will it affect anything? Well, Apple's still around and didn't collapse while Jobs was at the helm, and Linux won't either - the long time people have grown used to it, the newbies know what to expect, and everyone else just works away quietly on their fork of the kernel. And yes, there are tons of forks to the kernel - after all, every Android phone releases a kernel, yet I'm sure 99.99% of those patches never make it back into mainline.
From skimming the site I think it's subsription based. Speaking as a human, I was entertained by the sample clips. Also note that the image is in color which suggests to me that this is geared at least partly toward the owners.
Dogs, despite common belief, don't see in just black and white. They do have limited color vision (usually limited to blue and yellow) as well. Of course, recording in color means the differences in color vision is narrowed down.
Anyhow, I had dogs that reacted to dogs on TV (even upscaled SD programming of dogs). One other one I had just loved watched TV with motion on it). Of course, this was on an older 60Hz TV...
No. It presents itself as 8 cores, OS decides what to do with them. AFAIK currently Linux migrates threads using cpufreq clues. Nothing stops you from using all 8 cores at the same time.
Well, big.LITTLE allows both - a set of quad cores where one can work or the other, but not both (simplifies scheduler logic), or as a set of 8 cores if your scheduler is smarter and can control power requirements to turn off unneeded cores.
Though, the big thing is, you really don't want to run all 8 cores for more than a brief period of time. It turns out the limiting factor is thermal - if you even just run all 4 big cores at the same time, the chip will reach max junction temperature in just a few minutes. Without thermal control, it would overheat, and since you're using PoP packaging (with RAM on top), cooling is somewhat of a big concern.
I've see the result of thermal analysis, and it ended up being you could run two of the four at 100%, but the other two had to be software-modulated to around 50% load, and you'd maintain max temp.
Of course, a few minutes is all you really need if you want to benchmark - no CPU power benchmark really stresses the chip to requiring thermal amangement. Would be nice to graph out though to see maximum practical speed you can get in a certain configuration.
I never understood the practical reasons for locating mission control in Houston. Seems to me it was a purely political play by Johnson. Wouldn't you want to be as close to the equator as possible to maximize the benefits of the earth's rotation? Wouldn't that mean Key West? And why wouldn't you consider Puerto Rico?
Why? Mission control doesn't leave earth, so why put it there? Mission control can be anywhere really - since you're already relying on a huge network of remote antennas to communicate and get telemetry, all you really need is a location with good infrastructure. There's very little advantage to being physically close to where the launch site is
Responsibility requires effort, but empowers the user. To basically give away that control simply because it's "too hard" to know how your files are saved and where on local storage, smacks of going backwards.
Guess what? People don't care.
Not everyone in the world, hell, the vast majority of people in the world, give one shit about computing. They only have PCs and all that because they need it. They really don't care for it, and you'll see it in their behaviors. It's why tablets, smartphones and alternative computing platforms like Chromebooks are popular.
It's like owning a car - there are a ton of cars out there where all one needs to do is put in gas and bring it to the dealer every 6 months. Because automakers have realized that not everyone wants to be a mechanic, and despite the amount of independent garages and the like with all the same training and equipment, the vast majority still just bring it to the dealer and pay dealer pricing.
Why?
Because people don't all have the same interests, or wants. A computer is like a car - if you ask most people, if they could avoid having one, they would. Computers, like car, require lots of maintenance. But that gets in the way of doing other things they want in life, so if the car demands a lot of maintenance, surprise surprise, they don't get it.
Hell, the mechanics at said garages may have computerized diagnostic equipment. That stuff may run Linux internally, or Windows, or whatever. The mechanics don't care about it because they don't have to - that equipment's on a service contract and if it fails, they call in support. Your mechanic won't compile new kernels for it because they shouldn't need to (and they probably don't care - you might, if you're having to pay for said kernel compile).
Most people's data collections consist of music, movies, photos and documents. The latter two are almost always generated by themselves, while the former are generally disposable and obtained from a cloud service that'll re-stream it to them.
And most people don't backup at all. Hell, if you visit your parents for holidays, go do a backup of their PC for them. Because more often than not, that backup hard drive you bought them was used once and sat in the box since then as they forgot to do it again.
Bose do not aim to accurately reproduce the sound. They aim to sound good to people who don't care about accuracy, which is probably most people.
Exactly. They're also into aesthetics - heavily so.
They know people would pay a lot for those tiny cube speakers - which given all their size, aren't really suited for anything more than the range humans "care" about. The big subwoofer helps get to the lower frequencies, while the cube speakers that make the system "cute" and has high WAF really get you the parts that sound nice.
Put another way, that subwoofer gets you the thumps in the bass, the cube speakers only really do 1-8kHz well "enough" that with suitable content sound "good". Things like movies where you have to have good dialog (most human voice energy is between 3-6kHz or so) and a loud booming explosion style of content, and common top 40s music where there aren't any highs nor lows to begin with (other than a thumping bass).
Listen to anything else outside of that style of content and the limitations become clear.
The only thing I think is a gaping limitation is the lack of IP printing without a middleman. It's kind of stupid that i need to have an XP machine running somewhere in order to print. Organizations looking to supplement their hardware options with chromebooks shouldn't need to buy special printers to go with them.
Just buy a better printer. A lot of newer ones support practically every print protocol available - from de-facto lpr/lpd, IPP, AirPrint, Google Cloud Print (for Chromebooks, this is what you need), Bonjour, Windows,...
because the shipping company doesn't worry at all about overloaded containers or ships at all.
Why should they? They're insured.
Unfortunately, they lose out still because of opportunity costs. You can't go out and buy a new cargo ship at Wal-Mart, well, at least, full size ones.
No, it takes anywhere from a year to several years to build a cargo ship, including design time and waiting for the shipyard to have free capacity (yes, ships are scheduled for building).
And the insurance companies don't pay while the investigation continues - so the shipping company is out of raw revenue from shipping.
That self-interest generally keeps shipping companies honest - losing a boat is not a fun thing, and downtime costs money. Of course, that doesn't mean they don't try to get away with it and push their luck...
Well, let's say you love to shop Amazon (and admit it, you do).
Basically this extension sees what you're trying to buy and sees if it can find it on Amazon cheaper and then popup a message saying such.
Perhaps you're shopping Newegg and find some product you want. The Amazon thingy pops up and can tell you if Amazon has it cheaper so go shop there. Or if you're wanting to buy something and never clicked the checkout, it can pop up showing you that it's on sale.
It's like that Amazon app for your smartphone - you scan the barcode, and tap Buy and Amazon ships it to you, all while you're browsing in the store. Except instead of just B&M stores, Amazon now does it for online stores as well.
ARM has been focusing on mobile platform architecture for much longer than intel, its like Honda trying to make a truck, sure they did do it but its nothing like a company making nothing but trucks from the get-go. Intel needs to stay right where they are king and keep that crown and stop trying to follow rabbit holes thinking they might find money where ARM is the clear leader down in those places.
Problem is, Intel's kingdom is shrinking and while it's the bread and butter, it won't be large enough to sustain them in the long term.
Like buggy whip manufacturers, Intel's having to evolve,and that's what they're doing. There's tons of crap that's wrong with x86, and perhaps Intel's stubbornness to the architecture is the wrong approach, but they have to try something.
Right now, Intel's trying to shoehorn x86 into ARM's territory. Will they be successful? So far, not so much, but who knows?
And a little competition for ARM can't be too bad a thing. Heck, ARM's trying to do a bit of the same by going 64-bit and entering the server room.
(a) forced them to speak English well (because air traffic controllers speak English worldwide, apparently),
At civil airports, English is mandatory. It's an ICAO requirement, actually, that all communications take place in English using standard phraseology.
In fact, the requirement has gone up to require ALL pilots and controllers be tested for English proficiency - even if you're in an English-speaking country and speak it natively. Yes, you have to submit to a (relatively simple) English proficiency test as part of your license.
Apparently, native speakers who score the max (Expert) are exempt from future tests - those who score one below (Operational) must re-take the test yearly. Operational is the minimum required to pass.
Note this only applies to civil aviation. Military airports and airfields are completely different beasts.
And in Canada, Quebec likes to be different so all their controllers tend to greet initially in French and grudgingly speak English to Canadian aircraft. (International aircraft they'll happily speak English to).
I am an EE, and like every other EE I know, I advise my children to stay the hell out of engineering.
Why? There's a lot of EE positions that are paying extremely well because there's no one tulfill the positions. Think fresh grads with near 6-digit salaries being hired at the ceremony type stuff.
EE is a HUGE field. If you narrowly consider EE to be software, then yeah, maybe you have a point. But there's a TON more stuff to EE. And a lot of it can't be outsourced, either.
Want such a field? Go into Power Engineering. Electric utilities are looking for tons of people to help maintain the aging grid, and there's so few power engineers coming out (it's not a very glamorous field) that they're not replacing the people retiring.
There's plenty more if you want to work in electromagnetics (especially these days when high-speed design is all electromagnetics). Or analog IC design (digital gets all the exposure, but the analog side is hurting for people, especially now that a lot of digital relies on analog effects).
Sometimes computer and software engineering gets lumped into EE, which may be fields to avoid. But there's plenty more to it than the popular ones (and it's possible that even at big universities, you end up being the ONLY graduate in that specialization, despite sharing courses with everyone else).
Heck, an emerging EE field merges EE and ME together - electromechanical engineering where you combine both EE designs with ME designs and have them work harmoniously together. Usually it's more towards robotics (as the most obvious application) but even complex mechanical designs now often use a fair bit of electronics and electrical to augment what would be difficult to achieve otherwise.
In other words, the non Fed attendees can't be trusted to act like adults, so the con is warning the Feds not to attend. That really engenders trust in the hacker community. Good move!
Even worse, it would be reported as hackers acting like 12-year-old jackasses thus tarring every "hacker" as a 12-year-old immature idiot. Or a great way to prejudice anyone caught "hacking".
Fragmenting an existing standard creates a new standard that can draw in $$$. Everything else, national security, national pride, etc., are just excused to rip public funds. US or China.
Or in China, it's more a case of having to work with the Chinese government on technology.
China's gone their own way on a lot of stuff - one important one not mentioned is WiFi. China's got a similar-but-quite-unlike-WPA2 security standard that only China uses. And every WiFi devise there has to support it. Of course, to get a copy of the spec, you have to work with the Chinese government. Oh, and they also request you disable WPA2.
You must have it to sell your product in China. You can guess one possible reason why this is so.
Quite likely, the Chinese will fragment it as a China-only option for various reasons. One of which is obvious, the other is advantaging Chinese companies (who get early access to the specs)...
From looking at some of those problems, it seems to me that it's more important to be a better mathematician than a programmer.
Which is the foundation for computer science. These are computer science contests, not programming contests. There's a huge difference.
Computer science is a science - it's the study of computing. Coding is required, but relatively secondary to the whole thing. You can apply what you learn from CS to practical things - that's what computer engineering is about. Then there's just the garden variety coder who takes a course on Java and that's it - that's a trade.
These problems are about generating algorithmic solutions and realizing when one may outdo another (and sometimes, realizing that while nice solution is great, sometimes brute force is the order of the day).
If you are a consumer and you have a 4 year old version of Android on your phone, you will not exactly be shocked that it wont run apps that need a new phone. If you cared, you would upgrade. People with 2.1 are not going to buy your app. Fragmentation would be if people were launching products with wildly different versions of the OS: eg with HTC or Samsung's own APIs. Yes they do have manuf specific APIs. just dont use them and you will be fine. The generic APIs will be there and work.
You do realize that... you can get Android phones shipping with Gingerbread today, right?
These aren't ancient Android phones running Gingerbread because the user hasn't upgraded (in fact, most people upgrade phones when their contracts expire - because if you're taking on a new contract, you might as well get a new phone!). They're running ancient Android because that's what they shipped with! And what carriers are pushing out.
Basically, the Android phones that move are the free ones (++ Android Fragmentation - for making phones available in a wide range of pricing, -- Android Fragmentation for making crap-droids with crappy UXs possible), and why not? Featurephones cost just as much as a smartphone these days (i.e., free), and the lower end free crap-droids you don't even need a data plan for.
Since when has the industry had any interest in broadening their sales? if anything, the reverse. see region coding, or try using netflix outside of the USA (hint, you won't find the same selection) The industry won't wake up and embrace the global economy, they will go down kicking and screaming, and they intend to take the rest of the world with them.
Exactly. That's why we have Blu-Ray instead of HD-DVD - the movie companies were seriously PO'd about HD-DVDs being imported in countries (HD-DVD got rid of all region codes) where the movies haven't even had theatrical releases that they started delaying HD-DVD releases (so the DVD would come out, or the DVD+Blu-Ray). In the end, they were much happier with the more restrictions that Blu-Ray had that the HD-DVD camp gave up.
i forgot for a second there that this is slashdot, and logic and reason do not apply when talking about copyrights, and supporting anything other than free software and content for everyone is automatically equated with supporting the MPAA (which i dont) and trolling.
The ironic thing is that while true, copyleft requires copyright, and the more reasoned copyleft advocates do not want an abolition of copyright at all. Because without copyright, licenses like the GPL become effectively BSD (if you choose not to accept the GPL, your rights to the software are the same as de-facto copyright, i.e., all rights reserved. So you can't distribute GPL'd software, for example).
Of course, the same advocates against the RIAA/MPAA tactics will probably welcome using them on GPL violators...
Big whoop indeed. How often do you buy some expensive shit that doesn't work after only five years? A typical computer gets about a decade of support, that's what Microsoft does at least. Even then, once the official support runs out, you can still get software running on that computer. e.g. you can download tens of thousands MS-DOS apps and games, if that's what you want. That makes the 5-year-old gizmo less capable than a 386 which is a bit ridiculous.
Yes, but will a system from 2003 be as useful, even though the OS is still supported? Remember, you're talking about systems from the Pentium III era, 750MHz or less, with 512MB or less of RAM.
Such a system is fairly useless today - even running a recent Firefox on Windows XP on something like that is an exercise in patience. If you can even install all you need, given the average size of hard drive is 20GB, and expansion via USB 1.1.
And thats trying to run modern software without flaws.
Now a PC from 5 years ago is still decent today - Core2Duo, 4GB of RAM, 250GB hard drive. Pop in an SSD and it'll still be fairly useful, but that's more a case of the user not being able to tax it than anything else - it's "powerful enough".
Of course, smartphones have evolved rapidly the past 6 years thanks to the iPhone - before 2007, I had a nice smartphone - it had a 200MHz CPU with 16MB of RAM and ran some apps. Post iPhone though, with CPUs ramping up from 400MHz to nearly 2GHz, single to quad+ cores, RAM going from 128MB to 2GB, and from nonexistent 3D to practically console quality graphics on screens that were Half-VGA to full 1080p displays or higher.
Of course, smartphones have now matured quite a bit and there isn't much growth left, so a smartphone from 5 years ago is fairly useless, but a smartphone from 2 years ago is pretty damn useful today. And most likely, in 2 years, today's smartphones will probably work just fine as improvements have been slim.
The 'next step' is probably not printing with eutectic InGal alloys, but with other metals that exhibit the same properties but at higher melting points. One of the paper authors casually mentioned on Reddit that Aluminium happens to have the right properties.
If you're willing to accept long print times, melting microlitre/picolitre quantities of Aluminium with an induction heating nozzle seems entirely possible to do at home.
Why not use existing Tin-alloys? the melting point of those isn't that high to begin with, and we have plenty of that in cheap quantities (it's also known as... solder).
it's still lazy and ends up with juggling between the two cores, but that's not arms problem so they went with it.
but this is prett much the 4+1 core solution from nvidia anyways. it would be far better if they could just shut down parts of the one core to stop leakage. article blurb is just stupid though, it implies this would be a way towards cheaper, while it obviously isn't since it uses more die space.
Except it's easier in software to use big.LITTLE. If you wanted to switch from A7 to A15 and back, as long as the cores lined up and you had a cache coherent bus (a requirement anyways), all you'd really do was activate the new cores, force a context save on all 4 cores, then restore the context on the new cores. Shut down the old cores, and let the caches in the new cores warm up through cache coherency, then issue one final flush on the old cores and you're done.
As far as the OS is concerned, nothing changed. The scheduler doesn't even need to be super-aware of the change (it could, which makes life easier, but if it isn't, you can use an upper level library to do it underneath the OS).
NVidia's solution requires scheduler awareness, and runs the risk that the +1 is not sufficient to handle existing tasks thus keeping the 4 cores unnecessarily up.
The big problem though isn't power draw - the A15 is a huge power hog, true, but the problem with that is thermal issues.
Yes, thermal. You run 4 A15 cores full tilt and the chip heats up significantly - easily reaching max junction temperature in a few minutes. And there's very little in the way of cooling - you can't have huge heat planes on the PCB because you're talking high-density BGA parts, and the top of the chip is covered by a PoP memory chip.
I've seen thermal analysis done - the best that could be done would be 2 A15s running full tilt, with the other two software-modulated to run under 50% load after a few minutes, which would keep the chip at max temperature. And the system would try to be progressive - as the chip heated up, the cores would increasingly require being idle to maintain thermal limits.
Of course, a few minutes is all you need when you're e-peening GeekBench and other related benchmarks.
F-14s could engage the drones at 20 miles. The unmanned forces would never get a shot off.
Then B-52s would overfly the control center and drop 200 iron bombs on each building every hour for six weeks.
And how long would that last?
If the enemy sent up 100 drones (cheap, mostly unarmed) and your fighters all downed them from 20 miles away, you've just depleted 100 missiles. If they launch another 100 drones, do you have another 100 missiles at the ready?
Heck, a tactic would be to send up 100 extra cheap unarmed drones on the first wave, then send up the 100 armed ones afterwards after the missiles have been depleted.
It's like one aspect of warfare - overwhelm. Unless the fighters have a way to disarm huge arrays of drones with one missile, there's a chance of breakthrough.
Especially since fighters cost a lot more than drones.
Binary formats save development time since you don't have to parse them or convert to them. However if you're on a big system with lots of memory maybe it's no big deal to link in a library to do this. If you're on a tiny system however this is problematic.
Not parsing a binary format leads to all sorts of problems.
In the old days, yes, binary formats weren't parsed - you'd just cast a pointer to the structure and call read()/write() to it and everything would magically work and your structure would be populated.
The problem with this is obvious: security and interoperability.
This was fine in single user days where the same program that wrote the file would be the one reading it back. But now you have multiple architectures, endianess, padding and alignment, etc, such that it's impossible to actually just do this.
Not that you'd want to, because many security flaws were exposed because below-average parsers that didn't check for errors in the spec. Hell, many TCP problems emerged from this and many TCP stacks had to have patches and addons to handle various flaws in the way TCP packets were handled.
Binary formats are faster to parse, since you can rely less on string comparisons (e.g., instead of parsing "GET/", you might parse "0x01 '/'") where it's easy to read the command as reading one byte rather than doing a string compare (you still need to validate the byte and string are valid, though).
But if you're hoping to parse something by casting a struct to a binary blob, you'll be in for a world of hurt if you're not verifying every element in the structure is sane. Of course, given the transformations needed, it's probably easier to do parsing byte-by-byte like you would text anyhow.
That makes far too much sense. And to top it off, where such a philosophy would have been normal in the early days of the personal computer industry, cooperation between businesses to where everyone can succeed does not align with the modern douchebag corporate philosophy, where it is not enough for a company to be a success, but all competitors must ultimately fail. And if that is not possible, scorch the land.
But how much cooperation is "too much"?
Would the eBook publishers "cooperating" to make eBooks more successful be considered "too much"? The DoJ thought so.
Would Apple, Microsoft, Google all banding together make for a better place? Or perhaps if Apple "gave up" on OS X, and joined Microsoft to help make Windows better? You know, after all, it had like 98% marketshare at that point over a decade and a half ago.
A form of cooperation is collusion, and trust me, even in the early PC industry, competition was king. Sure a lot of people helped each other, but a lot of others competed. The early days of the Homebrew Computer Club were all about bragging rights, a form of competition.
Perhaps Apple and Google should've cooperated and they could fold Android into iOS, and leave us with one true smartphone OS.
When businesses start acting nice to each other, it's time to worry.
Slashdot Mirror
Fabs are expensive. If they're not running 24/7/365, they're costing money. It takes billions to open a fab, only to have in 18 months having to pay out a significant chunk of that money AGAIN to buy all new fab equipment because technology marches on.
It's why there are few people with fabs around the world - Intel, Global Foundries (AMD), TSMC, Samsung, Toshiba and a handful of others.
Oh yeah, each fab is different! The one optimized for making RAM isn't as good for making CPUs, the one making CPUs isn't as good for making RAM, and Flash lines just don't do well for either as well. Each kind of device requires special processing without which makes it suboptimal. (It's why embedded flash or embedded RAM just is never as good - or we'd be sticking RAM and Flash on die already).
So one fab becomes many and each needs new sites to accommodate new nodes (the older nodes are still useful for other things to help amortize the costs).
As for Android? Most of that 70% marketshare are for crap-droids with piss-poor screens, anemic processors, and RAM under 1GB. The flagships, even popular ones like the SGS3 make up only a tiny proportion of actual sales (the SGS3 is probably around 10% of the entire Android market). Most of that 70% goes to other fabs like TSMC (Qualcomm, MediaTek, etc), and will contain whatever crap is on sale that week.
It's why Intel sells foundry services to small companies - it helps top up their business, and also ensures they don't monopolize the fab. Fabs are just very capital intensive to run, and unless you can keep it going with parts continually, become a money sink.
Only while it stays within the network.
One thing that makes Tor more popular than other darknet systems is that it connects to the regular internet. Sort of how VoIP took off once call gateways to the POTS network were established. Because before that, well, you're just talking amongst yourselves and that's it.
Of course, once you hit the public internet, your traffic is back in plain text that can be logged, recorded and analyzed by the owners of the exit node. (And a lot of traffic contains personal information - like say, bittorrent), as well as being a potential MITM if you're trying to do an SSL connection over Tor.
So, basically, the Steve Jobs management model strikes again. First at Apple, and again for Linux.
Of course, the thing is, the Steve Jobs method of being an asshole only works for certain people with certain personality traits. For Jobs, it was his charisma and RDF. For Linus, it's practically hero worship - after all, he created a very useful piece of software.
Of course, there's also a whole pile of people who are put off by such tactics - after all, having physical violence threatened or being cursed out can make one question their motivations. For there are legions of people who were put off by Jobs' assholeishness and many who quit after being chewed out by Jobs (and these people were, for the most part, leaving dream jobs that paid decently - they just couldn't stand the verbal abuse Jobs would dish out).
Will it affect anything? Well, Apple's still around and didn't collapse while Jobs was at the helm, and Linux won't either - the long time people have grown used to it, the newbies know what to expect, and everyone else just works away quietly on their fork of the kernel. And yes, there are tons of forks to the kernel - after all, every Android phone releases a kernel, yet I'm sure 99.99% of those patches never make it back into mainline.
Dogs, despite common belief, don't see in just black and white. They do have limited color vision (usually limited to blue and yellow) as well. Of course, recording in color means the differences in color vision is narrowed down.
Anyhow, I had dogs that reacted to dogs on TV (even upscaled SD programming of dogs). One other one I had just loved watched TV with motion on it). Of course, this was on an older 60Hz TV...
Well, big.LITTLE allows both - a set of quad cores where one can work or the other, but not both (simplifies scheduler logic), or as a set of 8 cores if your scheduler is smarter and can control power requirements to turn off unneeded cores.
Though, the big thing is, you really don't want to run all 8 cores for more than a brief period of time. It turns out the limiting factor is thermal - if you even just run all 4 big cores at the same time, the chip will reach max junction temperature in just a few minutes. Without thermal control, it would overheat, and since you're using PoP packaging (with RAM on top), cooling is somewhat of a big concern.
I've see the result of thermal analysis, and it ended up being you could run two of the four at 100%, but the other two had to be software-modulated to around 50% load, and you'd maintain max temp.
Of course, a few minutes is all you really need if you want to benchmark - no CPU power benchmark really stresses the chip to requiring thermal amangement. Would be nice to graph out though to see maximum practical speed you can get in a certain configuration.
Why? Mission control doesn't leave earth, so why put it there? Mission control can be anywhere really - since you're already relying on a huge network of remote antennas to communicate and get telemetry, all you really need is a location with good infrastructure. There's very little advantage to being physically close to where the launch site is
Guess what? People don't care.
Not everyone in the world, hell, the vast majority of people in the world, give one shit about computing. They only have PCs and all that because they need it. They really don't care for it, and you'll see it in their behaviors. It's why tablets, smartphones and alternative computing platforms like Chromebooks are popular.
It's like owning a car - there are a ton of cars out there where all one needs to do is put in gas and bring it to the dealer every 6 months. Because automakers have realized that not everyone wants to be a mechanic, and despite the amount of independent garages and the like with all the same training and equipment, the vast majority still just bring it to the dealer and pay dealer pricing.
Why?
Because people don't all have the same interests, or wants. A computer is like a car - if you ask most people, if they could avoid having one, they would. Computers, like car, require lots of maintenance. But that gets in the way of doing other things they want in life, so if the car demands a lot of maintenance, surprise surprise, they don't get it.
Hell, the mechanics at said garages may have computerized diagnostic equipment. That stuff may run Linux internally, or Windows, or whatever. The mechanics don't care about it because they don't have to - that equipment's on a service contract and if it fails, they call in support. Your mechanic won't compile new kernels for it because they shouldn't need to (and they probably don't care - you might, if you're having to pay for said kernel compile).
Most people's data collections consist of music, movies, photos and documents. The latter two are almost always generated by themselves, while the former are generally disposable and obtained from a cloud service that'll re-stream it to them.
And most people don't backup at all. Hell, if you visit your parents for holidays, go do a backup of their PC for them. Because more often than not, that backup hard drive you bought them was used once and sat in the box since then as they forgot to do it again.
Exactly. They're also into aesthetics - heavily so.
They know people would pay a lot for those tiny cube speakers - which given all their size, aren't really suited for anything more than the range humans "care" about. The big subwoofer helps get to the lower frequencies, while the cube speakers that make the system "cute" and has high WAF really get you the parts that sound nice.
Put another way, that subwoofer gets you the thumps in the bass, the cube speakers only really do 1-8kHz well "enough" that with suitable content sound "good". Things like movies where you have to have good dialog (most human voice energy is between 3-6kHz or so) and a loud booming explosion style of content, and common top 40s music where there aren't any highs nor lows to begin with (other than a thumping bass).
Listen to anything else outside of that style of content and the limitations become clear.
Just buy a better printer. A lot of newer ones support practically every print protocol available - from de-facto lpr/lpd, IPP, AirPrint, Google Cloud Print (for Chromebooks, this is what you need), Bonjour, Windows, ...
Unfortunately, they lose out still because of opportunity costs. You can't go out and buy a new cargo ship at Wal-Mart, well, at least, full size ones.
No, it takes anywhere from a year to several years to build a cargo ship, including design time and waiting for the shipyard to have free capacity (yes, ships are scheduled for building).
And the insurance companies don't pay while the investigation continues - so the shipping company is out of raw revenue from shipping.
That self-interest generally keeps shipping companies honest - losing a boat is not a fun thing, and downtime costs money. Of course, that doesn't mean they don't try to get away with it and push their luck...
Problem is, Intel's kingdom is shrinking and while it's the bread and butter, it won't be large enough to sustain them in the long term.
Like buggy whip manufacturers, Intel's having to evolve ,and that's what they're doing. There's tons of crap that's wrong with x86, and perhaps Intel's stubbornness to the architecture is the wrong approach, but they have to try something.
Right now, Intel's trying to shoehorn x86 into ARM's territory. Will they be successful? So far, not so much, but who knows?
And a little competition for ARM can't be too bad a thing. Heck, ARM's trying to do a bit of the same by going 64-bit and entering the server room.
At civil airports, English is mandatory. It's an ICAO requirement, actually, that all communications take place in English using standard phraseology.
In fact, the requirement has gone up to require ALL pilots and controllers be tested for English proficiency - even if you're in an English-speaking country and speak it natively. Yes, you have to submit to a (relatively simple) English proficiency test as part of your license.
Apparently, native speakers who score the max (Expert) are exempt from future tests - those who score one below (Operational) must re-take the test yearly. Operational is the minimum required to pass.
Note this only applies to civil aviation. Military airports and airfields are completely different beasts.
And in Canada, Quebec likes to be different so all their controllers tend to greet initially in French and grudgingly speak English to Canadian aircraft. (International aircraft they'll happily speak English to).
An example set of questions and responses:
http://www.tc.gc.ca/eng/civilaviation/standards/general-personnel-test_taker_guide-2296.htm
Why? There's a lot of EE positions that are paying extremely well because there's no one tulfill the positions. Think fresh grads with near 6-digit salaries being hired at the ceremony type stuff.
EE is a HUGE field. If you narrowly consider EE to be software, then yeah, maybe you have a point. But there's a TON more stuff to EE. And a lot of it can't be outsourced, either.
Want such a field? Go into Power Engineering. Electric utilities are looking for tons of people to help maintain the aging grid, and there's so few power engineers coming out (it's not a very glamorous field) that they're not replacing the people retiring.
There's plenty more if you want to work in electromagnetics (especially these days when high-speed design is all electromagnetics). Or analog IC design (digital gets all the exposure, but the analog side is hurting for people, especially now that a lot of digital relies on analog effects).
Sometimes computer and software engineering gets lumped into EE, which may be fields to avoid. But there's plenty more to it than the popular ones (and it's possible that even at big universities, you end up being the ONLY graduate in that specialization, despite sharing courses with everyone else).
Heck, an emerging EE field merges EE and ME together - electromechanical engineering where you combine both EE designs with ME designs and have them work harmoniously together. Usually it's more towards robotics (as the most obvious application) but even complex mechanical designs now often use a fair bit of electronics and electrical to augment what would be difficult to achieve otherwise.
Even worse, it would be reported as hackers acting like 12-year-old jackasses thus tarring every "hacker" as a 12-year-old immature idiot. Or a great way to prejudice anyone caught "hacking".
Like Schwartz or Snowden.
Or in China, it's more a case of having to work with the Chinese government on technology.
China's gone their own way on a lot of stuff - one important one not mentioned is WiFi. China's got a similar-but-quite-unlike-WPA2 security standard that only China uses. And every WiFi devise there has to support it. Of course, to get a copy of the spec, you have to work with the Chinese government. Oh, and they also request you disable WPA2.
You must have it to sell your product in China. You can guess one possible reason why this is so.
Quite likely, the Chinese will fragment it as a China-only option for various reasons. One of which is obvious, the other is advantaging Chinese companies (who get early access to the specs)...
Which is the foundation for computer science. These are computer science contests, not programming contests. There's a huge difference.
Computer science is a science - it's the study of computing. Coding is required, but relatively secondary to the whole thing. You can apply what you learn from CS to practical things - that's what computer engineering is about. Then there's just the garden variety coder who takes a course on Java and that's it - that's a trade.
These problems are about generating algorithmic solutions and realizing when one may outdo another (and sometimes, realizing that while nice solution is great, sometimes brute force is the order of the day).
You do realize that... you can get Android phones shipping with Gingerbread today, right?
These aren't ancient Android phones running Gingerbread because the user hasn't upgraded (in fact, most people upgrade phones when their contracts expire - because if you're taking on a new contract, you might as well get a new phone!). They're running ancient Android because that's what they shipped with! And what carriers are pushing out.
Basically, the Android phones that move are the free ones (++ Android Fragmentation - for making phones available in a wide range of pricing, -- Android Fragmentation for making crap-droids with crappy UXs possible), and why not? Featurephones cost just as much as a smartphone these days (i.e., free), and the lower end free crap-droids you don't even need a data plan for.
Exactly. That's why we have Blu-Ray instead of HD-DVD - the movie companies were seriously PO'd about HD-DVDs being imported in countries (HD-DVD got rid of all region codes) where the movies haven't even had theatrical releases that they started delaying HD-DVD releases (so the DVD would come out, or the DVD+Blu-Ray). In the end, they were much happier with the more restrictions that Blu-Ray had that the HD-DVD camp gave up.
The ironic thing is that while true, copyleft requires copyright, and the more reasoned copyleft advocates do not want an abolition of copyright at all. Because without copyright, licenses like the GPL become effectively BSD (if you choose not to accept the GPL, your rights to the software are the same as de-facto copyright, i.e., all rights reserved. So you can't distribute GPL'd software, for example).
Of course, the same advocates against the RIAA/MPAA tactics will probably welcome using them on GPL violators...
Yes, but will a system from 2003 be as useful, even though the OS is still supported? Remember, you're talking about systems from the Pentium III era, 750MHz or less, with 512MB or less of RAM.
Such a system is fairly useless today - even running a recent Firefox on Windows XP on something like that is an exercise in patience. If you can even install all you need, given the average size of hard drive is 20GB, and expansion via USB 1.1.
And thats trying to run modern software without flaws.
Now a PC from 5 years ago is still decent today - Core2Duo, 4GB of RAM, 250GB hard drive. Pop in an SSD and it'll still be fairly useful, but that's more a case of the user not being able to tax it than anything else - it's "powerful enough".
Of course, smartphones have evolved rapidly the past 6 years thanks to the iPhone - before 2007, I had a nice smartphone - it had a 200MHz CPU with 16MB of RAM and ran some apps. Post iPhone though, with CPUs ramping up from 400MHz to nearly 2GHz, single to quad+ cores, RAM going from 128MB to 2GB, and from nonexistent 3D to practically console quality graphics on screens that were Half-VGA to full 1080p displays or higher.
Of course, smartphones have now matured quite a bit and there isn't much growth left, so a smartphone from 5 years ago is fairly useless, but a smartphone from 2 years ago is pretty damn useful today. And most likely, in 2 years, today's smartphones will probably work just fine as improvements have been slim.
Why not use existing Tin-alloys? the melting point of those isn't that high to begin with, and we have plenty of that in cheap quantities (it's also known as... solder).
Except it's easier in software to use big.LITTLE. If you wanted to switch from A7 to A15 and back, as long as the cores lined up and you had a cache coherent bus (a requirement anyways), all you'd really do was activate the new cores, force a context save on all 4 cores, then restore the context on the new cores. Shut down the old cores, and let the caches in the new cores warm up through cache coherency, then issue one final flush on the old cores and you're done.
As far as the OS is concerned, nothing changed. The scheduler doesn't even need to be super-aware of the change (it could, which makes life easier, but if it isn't, you can use an upper level library to do it underneath the OS).
NVidia's solution requires scheduler awareness, and runs the risk that the +1 is not sufficient to handle existing tasks thus keeping the 4 cores unnecessarily up.
The big problem though isn't power draw - the A15 is a huge power hog, true, but the problem with that is thermal issues.
Yes, thermal. You run 4 A15 cores full tilt and the chip heats up significantly - easily reaching max junction temperature in a few minutes. And there's very little in the way of cooling - you can't have huge heat planes on the PCB because you're talking high-density BGA parts, and the top of the chip is covered by a PoP memory chip.
I've seen thermal analysis done - the best that could be done would be 2 A15s running full tilt, with the other two software-modulated to run under 50% load after a few minutes, which would keep the chip at max temperature. And the system would try to be progressive - as the chip heated up, the cores would increasingly require being idle to maintain thermal limits.
Of course, a few minutes is all you need when you're e-peening GeekBench and other related benchmarks.
And how long would that last?
If the enemy sent up 100 drones (cheap, mostly unarmed) and your fighters all downed them from 20 miles away, you've just depleted 100 missiles. If they launch another 100 drones, do you have another 100 missiles at the ready?
Heck, a tactic would be to send up 100 extra cheap unarmed drones on the first wave, then send up the 100 armed ones afterwards after the missiles have been depleted.
It's like one aspect of warfare - overwhelm. Unless the fighters have a way to disarm huge arrays of drones with one missile, there's a chance of breakthrough.
Especially since fighters cost a lot more than drones.
Not parsing a binary format leads to all sorts of problems.
In the old days, yes, binary formats weren't parsed - you'd just cast a pointer to the structure and call read()/write() to it and everything would magically work and your structure would be populated.
The problem with this is obvious: security and interoperability.
This was fine in single user days where the same program that wrote the file would be the one reading it back. But now you have multiple architectures, endianess, padding and alignment, etc, such that it's impossible to actually just do this.
Not that you'd want to, because many security flaws were exposed because below-average parsers that didn't check for errors in the spec. Hell, many TCP problems emerged from this and many TCP stacks had to have patches and addons to handle various flaws in the way TCP packets were handled.
Binary formats are faster to parse, since you can rely less on string comparisons (e.g., instead of parsing "GET /", you might parse "0x01 '/'") where it's easy to read the command as reading one byte rather than doing a string compare (you still need to validate the byte and string are valid, though).
But if you're hoping to parse something by casting a struct to a binary blob, you'll be in for a world of hurt if you're not verifying every element in the structure is sane. Of course, given the transformations needed, it's probably easier to do parsing byte-by-byte like you would text anyhow.
But how much cooperation is "too much"?
Would the eBook publishers "cooperating" to make eBooks more successful be considered "too much"? The DoJ thought so.
Would Apple, Microsoft, Google all banding together make for a better place? Or perhaps if Apple "gave up" on OS X, and joined Microsoft to help make Windows better? You know, after all, it had like 98% marketshare at that point over a decade and a half ago.
A form of cooperation is collusion, and trust me, even in the early PC industry, competition was king. Sure a lot of people helped each other, but a lot of others competed. The early days of the Homebrew Computer Club were all about bragging rights, a form of competition.
Perhaps Apple and Google should've cooperated and they could fold Android into iOS, and leave us with one true smartphone OS.
When businesses start acting nice to each other, it's time to worry.