Ubuntu sold out to systemd because Debian sold out to systemd.
You missed the part in the middle where they attempted writing "upstart" as their own local "NIH daemon starting/hardware up-bringing" init replacement. And kept trying even after the rest of the Linux world standardized on systemd instead.
(Just like they kept trying bazaar, even after everybody else moved to git) (Just like they decided to not follow the common Wayland efforts, but write their own Mir) etc.
Ubuntu tried, but it didn't work well for them. Everybody else tried systemd and it turned okay for them.
Otherwise, they would have had to do all the work Devuan is doing now, to remove needless systemd dependencies.
If it's so much effort removing, maybe systemd wasn't that much needless.
The article is poor on the details of the methods. But given the size of the data set, and given the kind of hardware used (massive GPU cluster), it's possible that the system relies on deep neural nets (probably of the convnet type, given that it processes visual information) and these things are as close as we can currently get to an actual AI. (i.e.: something that learns on its own, forms its own way to interpret the data, and vaguely works in the same way as visual cortex in animals and human)
it's simplistic AI (it's closer to a collection to various visual cortex columns than a full-sized sentient brain. and even if there are some of the insects crawling in your basement with more brain power than this), but it's still AI.
Just because you *have* the code, doesn't mean you can GPL the code. The author is the only one that has that right (because they own the copyrights) unless they delegate those rights to someone else
or unless they gave you a license. like a BSD license
or is your intention to simply steal it and stick a GPL on it.
which is 100% doable when switching from BSD to GPL (but not the other way around).
BSD is license designed for the freedom of developers developers. it basically states: "you are allowed to do what the fuck you want with the code, even keeping it all for yourself and only releasing binaries, but you can also release it"
GPL is about freedom of end users. A user should always be provided with the freedom o code they are running, so they too can modify it not (not only you as a dev). it thus comes down to "you are allowed to do what the fuck you want to do with the code. but if you give out to someone elsr, they shoukd get the same rights to tinker ad you did".
If you look closely, the second is basically the first, but with the additional requirements to pass down the freedoms to the next. now play close attention : nothing in a permissive license like BSD would prevent you to add this extra clause. it s all about letting you, the dev, do whatever the fuck you want, and adding the requirement is something you might want to. (otoh, copyleft license require that transmit further down under the exact same condition you received. you receive it as GPL, the next one must be getting GPL. And that GPL doesn't allow you to remove that requirement.
relicensing problems do exist, but they concerne smaller details. - older BSD license required a name mention. it's not only "do what you want " (a change to GPl would be ob the list),but also "required to spell name in any product you hand out. (which goes against the GPL. GPL won't rely adding restrictions ) - gpl has versions. some softwar is licensed as "GPL v2-only" so you caan't release under anything else (because copyleft asks you to keep the rights) other explicitly mention future: "gpl v3 or anything more recent" so on the day GPL v6 is out you can use it.
Finland, the Netherlands, France, Switzerland, and the UK already has systems in place to make sure that people doesn't starve or end up homeless.
Yes, they all have social welfare.
But nope, unlike Netherland and Finland who are or were actually running pilot experiment, Switzerland voted against.
Note that Switzerland practice direct democracy. i.e.: no mattter what, the population has always the final say on everything. And in this case, democracy has spoken against UBI: apparently the population was indeed genuinely afraid of rise costs.
The cost is already there. Switching to UBI doesn't necessarily require higher taxes. Especially since you no longer need government workers investigating who is entitled to extra support.
That is the general idea behind UBI : - keep giving out money as before, under welfare programs. - except now you give out the money indiscriminately to absoluetely everyone. - because you blanked give money to everyone, you don't need to pay that many people to take care of the minute details. - saved money = extra money to redistribute.
you can be tracked all over the internet with ONE ID - including all your political commentary!
Technically, this effort (like lots of other similar efforts in the past) aren't targetting forum, but mostly on-line shops, and e-government platforms. - i.e.: things where you already need to identify with your real-world ID for obvious reasons. (e.g.: Because the goods need to be delivered to you in person). They are all platform who already know you, and could (if they wanted to put the effort and collude together) trace you.
You're confusing with OAuth and OpenID platforms (like Google, Facebook, etc.) which are targetting forums.
I would assume prey would be more sensitive. I'd rather miss a meal than be a meal.
Depends on how much their strategy for escaping the "be a meal" role relies on visual system. (As opposed to other strategy, like being massive and difficult to kill, etc.)
An Antelope should be able able indeed to spot the approaching big cat, and start to run.
(Then there are evolutionnary compromises : rod have faster response speed and better sensitivity to low-light condition, but are not colour-sensitive (increasing the risk that a predator could successfully camouflage) cone have colour sensitivity (but only work in broad daylight). Also making a visual system generally costs ressource. The better the eye, the more visual processing must be done by the visual cortex, and thus a bigger brain is a requirement).
Nvidia also doesn't use anymore on their GPU (16nm).
Seems that 28nm was used back in 2010~2011 products.
So this stolen technology won't suddenly enable China to produce AMD/Nvidia GPU clones.
On then other hand, China's own home made Loongson CPU, is currently at 65nm, but some models are scaled down to 28nm (the Longsoon 3B). So this is definitely going to help china produce their own CPUs locally, without needing to rely on some external 28nm-capable fab (currently Loongson are built by STMicroelectronics).
By adding 28nm capability to its portfolio, Huali Microelectronics would be in a position to offer to manufacture loongson locally - probably a worthy contract.
Do not use email or phone as authenticated endpoints.
Well technically, with using correct End-To-End encryption, you could use any channel as 2FA.
E-mail could be usable if correctly encrypted with a trusted openPGP key pair (or if you have a trusted S/MIME authority). But people won't bother fumbling arround with PGP nor even getting Mailveloppe to use with their webmail.
SMS could also be used with a correctly authetified OTR layer. But there are about 2 software in the whole universe using OTR-over-SMS (TextSecure. And there should be another one somewhere). So no serious enterprise is ever going to consider it.
but the point is the data is there, unencrypted, it's a matter of tapping the signal, and pretending you're the display.
The problem is that, in practice, in the name of making this as dead-cheap as possible, component tends to be highly integrated. So you have a single chip being fed the encrypted HDCP signal in, and directly spitting the desired signal out. (Much cheaper to make a device than having different chips for each stage of the signal processing talking to each other) So the unencrypted data you refer is actually between 2 cores inside a highly integrated chip.
That's also why you won't find screens with the exact unusuall feature set that you would like as a geek. Because these features aren't done by mixing-and-matching chips with discrete functions, but because the screen will basically use a single integrated chip for everything, and chip maker make chips with the most commonly needed features in.
(I'm oversimplifying. You'll still find some signal converters. e.g.: the chip only accept 3 HDMI inputs, so you need to add 1 DP-to-HDMI converter, 1 VGA-to-HDMI ADC (yup analog-digital-analog wrecks the quality, but that's the cheapest they managed to make it), but basically manufacturer are going to keep the chip count as low as possible. It's very likely that the decryption-core will be integrated tightly with the display core).
just have a word with the manufacturer to obtain some part. As they're all in China I'm sure you will be able to find one that can sell you the parts
Yup, given that there are genuine reasons to have a separate discrete HDCP stage (mostly to split signal between displays), you're sure to be able to find parts in china as you wish. Again, in the name of as-chepa-as-possible-integration you'll also find a bunch of output that you don't need (the additionnal outputs of the splitter), but hey, it works all the same. (e.g.: a ViewHD splitter)
There's dozen "On Screen" recorder software. So if I can read BlueRay on my computer, what kept me from recording the screen?
Probably depends on how the display is composited by the desktop on your screen. And how the real-time video compression hardware used by the recorder is playing along with the video decompression hardware used by the media player. You might end up with a grey rectangle instead. (see the problems of taking a screen shot of a video player that appeared 15 years ago when the first video acceleration overlay started to appear in Super VGA cards. Only much more complicated due to increased complexity of modern hardware and modern compositing desktops)
(Or if you go with the software solution : depends if the OS allows the recorder software to access the screen buffer of the software player, and the player still accepts to play in such an OS. Be ready to have to fumble around with an OS that implements DRM "secure computing", requiring a gpu device driver that implements "secure path" (i.e.: video-buffer memory protection) and a player that requires ton of signed code. You might end with a "content cannot be player currently on your computer" pop-up.)
At that point, it's easier to get a second machine, a box that scrubs HDCP away (e.g.: a vew HD splitter), and a HDMI to USB3 lossless grabber (magewell).
Or go for the whole "disassembly" approach mentioned above by another poster.
At that point the data is unencrypted, as it's set to drive all the individual pixels of your display. All the recorder has to do is collect the values of those pixels and store it again for later playback.
(NOTE: that you'll not find trace that drive all the individual pixels, but only lines and columns of a matrix, and the display is scanned. Also, in *active matrix*, the display doesn't directly drive the pixels, but drives active component (transistors+capacitors) in the cell which are then in charge of keeping their corresponding pixel in its desired state between scans)
On an OLED display : yes, basically it makes 3 different type of protein fluorescent in an electric field, one for each R, G, B. You get one signal for each.
On a modern LCD display : not exactly, what you see is an LCD grid (pixels going on or off) overlaid over an RGB LED light source (That's how the obtain good constrast ratio : in the darker region of the image, the LEDs of the backlight are turned down or off). So for each pixel, you get and R, G, B signal giving what fraction of the backlight should be allowed through, and for each region a LED signal (usually multiplexed: most modern RGB LEDs have a driver directly in the package) telling how much backlight that region of the image should emit.
So you either need to integrate those signal together (determine the composition of light at a given position, given the nearby LEDs in that region ; and then filter by the state of pixels in the LCD grid). Or you go and disassemble an OLED instead. Or you simply record it with a cam and let the integration be done for you by the screen.
Did anybody stop to consider the fact that speakers and microphones by-and-large are not capable of ultrasound frequencies? {...} It takes specially designed transducers to operate in the ultrasound range.
not optimized for ultra sound (like your car's parking range finder)
!= impossible to even pick faint ultra sound.
You don't want to perfectly reproduce ultra-sound music for your dog. You just want "morse-code" level of vague faint ultra sound.
Tiny speakers like the ones in a smartphone are going to hit 18khz at BEST. It's probably closer to 15khz in reality. Even high-end studio monitors only reach 20-22khz.
Yup speaker are mainly optimized for the human hearing range. (mostly in the 10Hz to 15kHz) range. Yet it doesn't go silent at 15001Hz, just less efficient. As stated in the article, they use 18kHz : at that frequency it's hard for human to really notice, but speakers are still not so bad, you can at least transmit simple beep-codes.
Regarding microphone : yes, they *are* sensitive at much higher frequencies. It's hard to do a perfect frequency filter that stops abruptely right at the 15kHz limit. Much easier to use a microphone that picks way higher than this, but at least doesn't muffle in the audible zone at all, record it with a ADC at a high sample rate, and then clean the sound digitally.
(That the actually real reason of 192kHz ADC/DAC you see in most pro-level equipment. Not that it makes any sense to keep a 192kHz audio track all the way to the customer. But at least if the recording stage works at a higher frequency, there's less risk of muffling the interesting frequencies or have distrosion that spill out into the audible range).
So a smartphone could pick a bit of 18+ kHz sound (at least something really easy to distinguish, like beep-codes). Even if it's not optimized to record calls in that range, it's still designed in a way that avoid muffling under the 15kHz range.
This story is complete bullshit.
yet, US communication actually work, and is used in the wild. When Spotify detect a device logged into your account on the same network, it can use ultra sound to match which device/speaker is connected where - i.e.: it can do ultra-sound pairing. (It says clearly on the message box that appears).
Chromecast has also been reported to support such ultrasound pairing.
Seems that dogs couldn't really perceive motion on older TVs because the framerate and resolution were too low.
Carnivore (predator) pets like dogs and cats tend to be much more sensitive to motion. They will *perceive* motion on TV, it will just look more choppy and flickering to them.
Just like human where able to perceive motion in silent film era's 12-16fps, in half-rate/dupe-frame 12-15fps animation, or in "shitty low"-fps GIFs. It looks a lot more choppy, than a 24/30fps or even a 48/60fps. Or just like human *can* see the flicker of a 60Hz CRT monitor when looked at the periphery of the view (i.e.: where there are more rods - sensors with faster response that are also responsible for the pets better motion sensitivity).
I had my cats recognize and react to things on my old 50Hz CRT, even if *I* could notice the flickering.
Hopefully, as more of these patents expire, a Vulcan based open source GPU will emerge.
As the numerous past failures of attempts at opensource GPU or even opensource graphic cards have shown, making a functional and competitive graphic card is EXTREMELY difficult.
There's a reason that nVidia and AMD both suddenly implemented tile based renderers,
Actually, Nvidia has had their own TBR patents for quite some time : - Nvidia bought up 3DFx for their patents and their engineer back when that one went bankrupt. - Before that, 3DFx had bought up Gigapixel, among other for their TBR patents, to be used in future product (forgot the code name) - and HSR (hidden surface removal) tech to be applied much earlier in then current product (in the then VSA-100 / Voodoo4/5/6 and in the upcoming Rampage / Spectre)
So Nvidia indirectly acquired TBR patents. Though for the record, they were more interested in the know-how and engineer which where working on the Rampage GPU ("3DFx Spectre" cards) due to programmable pixel shaders being all the rage, and retained them to work on GeForce FX (speculation backthen that probably the pun in the name was intended... )
So in theory, they could have moved into the field much faster than ATI / AMD. (But back at the Rampage / GeForce FX era, there where some area were TBR was problamatic : e.g. some transparency (i.e.: simple alpha-blend, back then) couldn't be handled in a single pass easily. So probably they decided not to bother. Given that modern games work with tons of subsequent passes (transparent materials cause diffraction/distortion in a separate pass of a pixel-shader), I would suspect that it's not that much a problem anymore).
(Note: I legally obtain the book I'm DeDRM-ing. I'm just removing the DRM because I'm fed up with the Adobe Digital Edition fucking things up on a regular basis and access to my book getting b0rked yet again. This kind of De-DRM-ing is actually tolerated in my local jurisdiction - as it should be everywhere)
Yeah but not by that *much* (though it depends on the quality of both encoders - x264 is incredibly better visually than nearly everything else). At least, not given the quality/bitrate that the industry has decided to use for Ultra HD (where it makes sense or if it's mostly a placebo is an entirely different can of worm).
Also HEVC is patent minefield (and thus hardware HEVC/H265 decoding isn't as widespread as AVC/H264), so perhaps they also want to keep a door open for content producers that can only afford the MPEG patent pool for H264/AVC and would prefer to keep that codec.
Of course, next-next generations codecs like AOMedia's AV-1 are right around the corner, with even better performance, less patent encumbremenent (xiph and google are onboard), and lots of industry supprot (youtube and netflix - ie. a huge chunk of all watched content - are behind it. As are most current hardware constructors), so perhaps will end up soon with actually smaller files (and thus smaller discs or - more likely - even smaller file to torrent)
it could easily fit on dual layer BDs...
And this movie is a nice example : it's 53GB, it DOES NOT fit on 50 GB dual layer BlueRays. You would need more layers, that would drive the price of pressing discs up (i.e.: less obscene margins for the industry). Instead UHD discs have slightly higher densities: the dual layer goes up to 66GB (so you can fit this 54GB movie on 2 layers only).
Systems like PS4 Pro (or even PS4 since it can do 4k at 30/24Hz)
Thus it would depend it the movie is 24-30 fps or 48-60 fps, and/or if there's (alterning frame) 3D. But yeah...
could easily play them if it weren't for the drive requirement
Now the big question : are there really that many differences between vanilla and UHD blueray disks ? or is it only a slightly higher surface density ?
I would suspect that it's only a slight difference of density, and that it should be in theory possible to reflash the drive itself with an upgraded firmware that change the behaviour of the servos of the head and the focus assembly. (The same way it was possible to obtain a little bit higher density with floppy drivers. Except that these where directly controlled so the controlling software actually ran on the PC's main CPU, no flashing required, just a device driver).
Not that any manufacturer is actually going to release such an upgrade (it's much more lucrative to have the users re-buy a new more expensive device which is actually the same device as before (no R&D costs !) but with slightly different constants in the firmware and a different model number written on the box).
I can output the signal and the audio on a HD recorder and there's no protection that will protect you from that.
Not exactly. Nowadays, the output signal is *digital* (HDMI mostly with standalone players, and HDMI mostly with computers). There's an encryption standard HDCP which is supposed to protect this data during its transit to the screen.
In theory, you should NOT be able to directly hook-up the output to a recorder, that recorder will only see an ecrypted stream that only the screen can decrypt. You can only use a *cam* to record the actual screen as suggested above, not the stream itself.
In practice, HDCP is done poorly. Its current form is cracked and can be bypassed, so the only actual real-world is not stopping pirate, but only failing in weird ways for legitimate users.
And in actual practice : nobody gives a damn about the latest cookie-cutter soulless movie. Chances are high that I'll be too busy doing some interesting outdoor sport (or some indoor one) rather than trying to see how I could pirate a copy of Smurfs 2 (what, they even made a *second* one ?)
I never understood why you couldn't just capture the encrypted key and feed it that and let it do it's thing.
And that's exactly how it was done with the very few generations of movies. Some BlueRay player (i think WinDVD ?) stored they decryption key in an insecure memory location, and hackers used to tap there to find which key is used to decrypt a specific BlueRay.
Movie industry noticed and revoked the keys for that player (meaning newer disc produced after that where encrypted with a selection of keys for which the player had no corresponding keys). Cue-in cat and mouse game, until hackers managed to find the master key that sits at the top of the whole encryption pyramid.
The extra colour and dynamic range with 4k, and even more so with 8k is really nice, but to get much out of it you need a dimly lit room and a TV capable of reproducing it. Similarly with audio you need a dedicated room and to then sit in the sweet spot while listening.
Huh... nope. 8k and 10bits colour isn't the equivalent to 200$ monster digital cables and 192kHz sample rate.
- Ears have some physiological limits due to how physics work (your ears can hear very approx in the 20-20'000Hz range. your body can also feel vibrations in the 1-100Hz. There's no receptor in a human body capable of reacting to 90kHz). - Physics of digital signals, and a whole bunch of signal processing science (e.g.: error correction) means that in the digital world, sometime a bit is just a bit, no matter the concentration of gold and diamond powder (sic!) in the cabling it goes through.
No matter the dedicated audio room you're sitting in, you'll never be able to hear ultrasounds (directly. though ultrasounds can cause distorsions in the audible range on some equiement), and monster cables will change nothing to the SPDIF link.
The "cinephile" equivalent of an audiophile insisting on 200$ monster cables and 192kHz rates, would be a guy who insist on movie formats that not only record Red, Green and Blue primary colours, but also infra-red and ultra violet (i.e.: insist on frequencies/wavelenghts for which the human eyes doesn't have any receptors) and on buying a $10'000 silver screen to project projecting onto which, that should also perfectly reflect x-rays, gamma rays and microwaves (completely irrelevant given what is transmitted by the light of the beamer).
No matter if the movie room is dimly lit or not, insisting on wave-lenght outside the human range (like ultra-violets) is useless, as is insisting on a screen optimized for something completely irrelevant.
The same way, no matter the dedicated listening room and it's sweet spot, a human ear lacks receptors for 96kHz sounds.
The value in a messaging system is in the users not the system.
A.k.a. "Network effect". As/. is nearly entirely composed of geeks, who by definition have weird fetishes regarding technology (and not only... ahem...) we are all constantly exposed to this whenever we try to persuade the vanilla people around us to try some new tech.
Tell you what: Right now where I live I can send anyone a whatsapp message, even someone who's number I just read off a business card,
The keyword is "Right now". A couple of years ago it used to be some other system (Apple's iPhone-only iMessage). then Facebook Messenger (still have some contact stuck on this one), now it's WhatsApp, next year it's going to be SnapChat (*) And who knows what awaits us beyond.
---
(*) : As a hobby, I am a ski instructor at my university's ski camps. I've noticed that among first-year student (~19 y.o.), very have WhatsApp installed and most of them spend their time on SnapChat.
i.e.: unlike you, I'm already in the situation where I CANNOT rely on sending a WhatsApp message to any random mobile number I happen to stumble upon. Though (as usual with network effect) this tend to be generational so - if I know the age group of my target - I can have a vague idea on which network I have the most chance of finding them.
---
NOTE that desktop chat went through the exact same kind of madness. ICQ, AIM, Y!M, MSN, Skype, Facebook chat, Google chat/hangouts, WebEx a bunch of more modern WebRTC meeting platforms, Rocket Chat, Slack, etc. At least during some time XMPP brought a relative hope to this madness (even Facebook used to have a XMPP gateway at some point in time), specially since servers started implementing federation (e.g.: you could talk between Google chat and Jabber.org users. Until Google decided "It's too much risk of SPAM" and disabled federation about the same time they introduced hangouts).
Which actually doesn't as much show that WhatsApp has some value (as opposed to any other messaging system), but shows how bad it is to rely on a single centralized service.
(as opposed to things which are a bit more distributed : like e-mail, like XMPP chat server with federation switched on,...)
If the police have strong evidence that the stolen diamonds are in your wall safe, and that the things is rigged to blow if they try to crack it - do you think a judge could NOT compel you to give up the combination ?
...but you still have the right to remain silent. In the US under Miranda's rights, and under similar rights in lots of other jurisdiction elsewhere.
Note that in this case, given the *strong evidence* and and your obstination of not revealing the safe's password, you might start to look really suspicious(*) to the jury/judge (whatever is relevant in your jurisdiction).
Then again, to keep your metaphor, any thief with half a brain cell will very probably NOT stash the diamond in their own safe at home, but try to hide them in some place that nobody would think about. So they can safely show that the safe is empty of diamond, so that police can't manage to find the diamond, so competing thieves can't manage to get a grip on them, and so, once the steam blows off and nobody bother to keep an eye on the case, the thief can discretely retrieve the diamonds and enjoy them.
Here we have some idiots who decided to blackmail a victim *USING THEIR OWN FUCKING PHONES*. The devices that can easily be confiscated by police, with message logs that can be easily traced back to them, and the fact that they chose to remain silent might look a little bit strange(*).
---
(*): though not necessarily. there's a scenario were the person might be innocent, but have some personal detail that they want to keep secret (e.g.: being gay and the publicly official partner being only "a bread". specially in culture that carries a stigma around homosexuality, or worse, in a jurisdiction with "sodomy laws" that considers it illegal)
on the other hand, if all other evidence points to guilty, they case won't be hinging on the sole content of the phone/safe...
I love how CRISPR makes geneticists act like a bunch of little kids who found a working go-cart someone left in a dumpster.
Except this time round, metaphorically they actually use it to really do go-cart laps on a nearby track (though necessarily the same track as the guy who threw it in the dumpster). Unlike all the other times when they decide to re-purpose it for something completely different, like using it as a delivery vehicle with optional autonomous navigation (which actually works more or less), or trying to conquer the world with it and at least managing to build a self-driving tank out of it (by strapping a bazooka on the precedent autonomous delivery vehicle).
This post was brought to you by the Slashdot's "Car Analogy Service(tm)".
Slashdot Mirror
Ubuntu sold out to systemd because Debian sold out to systemd.
You missed the part in the middle where they attempted writing "upstart" as their own local "NIH daemon starting/hardware up-bringing" init replacement.
And kept trying even after the rest of the Linux world standardized on systemd instead.
(Just like they kept trying bazaar, even after everybody else moved to git)
(Just like they decided to not follow the common Wayland efforts, but write their own Mir)
etc.
Ubuntu tried, but it didn't work well for them.
Everybody else tried systemd and it turned okay for them.
Otherwise, they would have had to do all the work Devuan is doing now, to remove needless systemd dependencies.
If it's so much effort removing, maybe systemd wasn't that much needless.
Statistical models are not AI.
The article is poor on the details of the methods.
But given the size of the data set, and given the kind of hardware used (massive GPU cluster),
it's possible that the system relies on deep neural nets (probably of the convnet type, given that it processes visual information)
and these things are as close as we can currently get to an actual AI.
(i.e.: something that learns on its own, forms its own way to interpret the data, and vaguely works in the same way as visual cortex in animals and human)
it's simplistic AI (it's closer to a collection to various visual cortex columns than a full-sized sentient brain. and even if there are some of the insects crawling in your basement with more brain power than this), but it's still AI.
Just because you *have* the code, doesn't mean you can GPL the code. The author is the only one that has that right (because they own the copyrights) unless they delegate those rights to someone else
or unless they gave you a license.
like a BSD license
or is your intention to simply steal it and stick a GPL on it.
which is 100% doable when switching from BSD to GPL (but not the other way around).
BSD is license designed for the freedom of developers developers.
it basically states:
"you are allowed to do what the fuck you want with the code, even keeping it all for yourself and only releasing binaries, but you can also release it"
GPL is about freedom of end users. A user should always be provided with the freedom o code they are running, so they too can modify it not (not only you as a dev). it thus comes down to "you are allowed to do what the fuck you want to do with the code. but if you give out to someone elsr, they shoukd get the same rights to tinker ad you did".
If you look closely, the second is basically the first, but with the additional requirements to pass down the freedoms to the next.
now play close attention : nothing in a permissive license like BSD would prevent you to add this extra clause. it s all about letting you, the dev, do whatever the fuck you want, and adding the requirement is something you might want to.
(otoh, copyleft license require that transmit further down under the exact same condition you received. you receive it as GPL, the next one must be getting GPL.
And that GPL doesn't allow you to remove that requirement.
relicensing problems do exist, but they concerne smaller details.
- older BSD license required a name mention. it's not only "do what you want " (a change to GPl would be ob the list),but also "required to spell name in any product you hand out. (which goes against the GPL. GPL won't rely adding restrictions )
- gpl has versions. some softwar is licensed as "GPL v2-only" so you caan't release under anything else (because copyleft asks you to keep the rights)
other explicitly mention future: "gpl v3 or anything more recent" so on the day GPL v6 is out you can use it.
Finland, the Netherlands, France, Switzerland, and the UK already has systems in place to make sure that people doesn't starve or end up homeless.
Yes, they all have social welfare.
But nope, unlike Netherland and Finland who are or were actually running pilot experiment, Switzerland voted against.
Note that Switzerland practice direct democracy. i.e.: no mattter what, the population has always the final say on everything.
And in this case, democracy has spoken against UBI: apparently the population was indeed genuinely afraid of rise costs.
The cost is already there. Switching to UBI doesn't necessarily require higher taxes. Especially since you no longer need government workers investigating who is entitled to extra support.
That is the general idea behind UBI :
- keep giving out money as before, under welfare programs.
- except now you give out the money indiscriminately to absoluetely everyone.
- because you blanked give money to everyone, you don't need to pay that many people to take care of the minute details.
- saved money = extra money to redistribute.
you can be tracked all over the internet with ONE ID - including all your political commentary!
Technically, this effort (like lots of other similar efforts in the past) aren't targetting forum, but mostly on-line shops, and e-government platforms.
- i.e.: things where you already need to identify with your real-world ID for obvious reasons. (e.g.: Because the goods need to be delivered to you in person).
They are all platform who already know you, and could (if they wanted to put the effort and collude together) trace you.
You're confusing with OAuth and OpenID platforms (like Google, Facebook, etc.) which are targetting forums.
I would assume prey would be more sensitive. I'd rather miss a meal than be a meal.
Depends on how much their strategy for escaping the "be a meal" role relies on visual system.
(As opposed to other strategy, like being massive and difficult to kill, etc.)
An Antelope should be able able indeed to spot the approaching big cat, and start to run.
(Then there are evolutionnary compromises :
rod have faster response speed and better sensitivity to low-light condition, but are not colour-sensitive (increasing the risk that a predator could successfully camouflage)
cone have colour sensitivity (but only work in broad daylight).
Also making a visual system generally costs ressource. The better the eye, the more visual processing must be done by the visual cortex, and thus a bigger brain is a requirement).
This has nothing to do with AMD
Even more so, as AMD doesn't use 28nm process anymore, neither on GPU (14nm FinFet), nor CPU (also 14 nm), whereas the chipset is still done with a coarser process AMD 990 (65 nm).
Nvidia also doesn't use anymore on their GPU (16nm).
Seems that 28nm was used back in 2010~2011 products.
So this stolen technology won't suddenly enable China to produce AMD/Nvidia GPU clones.
On then other hand, China's own home made Loongson CPU, is currently at 65nm, but some models are scaled down to 28nm (the Longsoon 3B).
So this is definitely going to help china produce their own CPUs locally, without needing to rely on some external 28nm-capable fab (currently Loongson are built by STMicroelectronics).
By adding 28nm capability to its portfolio, Huali Microelectronics would be in a position to offer to manufacture loongson locally - probably a worthy contract.
Do not use email or phone as authenticated endpoints.
Well technically, with using correct End-To-End encryption, you could use any channel as 2FA.
E-mail could be usable if correctly encrypted with a trusted openPGP key pair (or if you have a trusted S/MIME authority).
But people won't bother fumbling arround with PGP nor even getting Mailveloppe to use with their webmail.
SMS could also be used with a correctly authetified OTR layer.
But there are about 2 software in the whole universe using OTR-over-SMS (TextSecure. And there should be another one somewhere).
So no serious enterprise is ever going to consider it.
but the point is the data is there, unencrypted, it's a matter of tapping the signal, and pretending you're the display.
The problem is that, in practice, in the name of making this as dead-cheap as possible, component tends to be highly integrated.
So you have a single chip being fed the encrypted HDCP signal in, and directly spitting the desired signal out.
(Much cheaper to make a device than having different chips for each stage of the signal processing talking to each other)
So the unencrypted data you refer is actually between 2 cores inside a highly integrated chip.
That's also why you won't find screens with the exact unusuall feature set that you would like as a geek.
Because these features aren't done by mixing-and-matching chips with discrete functions,
but because the screen will basically use a single integrated chip for everything, and chip maker make chips with the most commonly needed features in.
(I'm oversimplifying. You'll still find some signal converters. e.g.: the chip only accept 3 HDMI inputs, so you need to add 1 DP-to-HDMI converter, 1 VGA-to-HDMI ADC (yup analog-digital-analog wrecks the quality, but that's the cheapest they managed to make it), but basically manufacturer are going to keep the chip count as low as possible. It's very likely that the decryption-core will be integrated tightly with the display core).
just have a word with the manufacturer to obtain some part. As they're all in China I'm sure you will be able to find one that can sell you the parts
Yup, given that there are genuine reasons to have a separate discrete HDCP stage (mostly to split signal between displays), you're sure to be able to find parts in china as you wish.
Again, in the name of as-chepa-as-possible-integration you'll also find a bunch of output that you don't need (the additionnal outputs of the splitter), but hey, it works all the same.
(e.g.: a ViewHD splitter)
There's dozen "On Screen" recorder software. So if I can read BlueRay on my computer, what kept me from recording the screen?
Probably depends on how the display is composited by the desktop on your screen. And how the real-time video compression hardware used by the recorder is playing along with the video decompression hardware used by the media player.
You might end up with a grey rectangle instead.
(see the problems of taking a screen shot of a video player that appeared 15 years ago when the first video acceleration overlay started to appear in Super VGA cards. Only much more complicated due to increased complexity of modern hardware and modern compositing desktops)
(Or if you go with the software solution :
depends if the OS allows the recorder software to access the screen buffer of the software player, and the player still accepts to play in such an OS.
Be ready to have to fumble around with an OS that implements DRM "secure computing", requiring a gpu device driver that implements "secure path" (i.e.: video-buffer memory protection) and a player that requires ton of signed code.
You might end with a "content cannot be player currently on your computer" pop-up.)
At that point, it's easier to get a second machine, a box that scrubs HDCP away (e.g.: a vew HD splitter), and a HDMI to USB3 lossless grabber (magewell).
Or go for the whole "disassembly" approach mentioned above by another poster.
At that point the data is unencrypted, as it's set to drive all the individual pixels of your display. All the recorder has to do is collect the values of those pixels and store it again for later playback.
(NOTE: that you'll not find trace that drive all the individual pixels, but only lines and columns of a matrix, and the display is scanned.
Also, in *active matrix*, the display doesn't directly drive the pixels, but drives active component (transistors+capacitors) in the cell which are then in charge of keeping their corresponding pixel in its desired state between scans)
On an OLED display : yes, basically it makes 3 different type of protein fluorescent in an electric field, one for each R, G, B. You get one signal for each.
On a modern LCD display : not exactly, what you see is an LCD grid (pixels going on or off) overlaid over an RGB LED light source (That's how the obtain good constrast ratio : in the darker region of the image, the LEDs of the backlight are turned down or off).
So for each pixel, you get and R, G, B signal giving what fraction of the backlight should be allowed through, and for each region a LED signal (usually multiplexed: most modern RGB LEDs have a driver directly in the package) telling how much backlight that region of the image should emit.
So you either need to integrate those signal together (determine the composition of light at a given position, given the nearby LEDs in that region ; and then filter by the state of pixels in the LCD grid).
Or you go and disassemble an OLED instead.
Or you simply record it with a cam and let the integration be done for you by the screen.
Did anybody stop to consider the fact that speakers and microphones by-and-large are not capable of ultrasound frequencies? {...} It takes specially designed transducers to operate in the ultrasound range.
not optimized for ultra sound (like your car's parking range finder)
!=
impossible to even pick faint ultra sound.
You don't want to perfectly reproduce ultra-sound music for your dog.
You just want "morse-code" level of vague faint ultra sound.
Tiny speakers like the ones in a smartphone are going to hit 18khz at BEST. It's probably closer to 15khz in reality. Even high-end studio monitors only reach 20-22khz.
Yup speaker are mainly optimized for the human hearing range. (mostly in the 10Hz to 15kHz) range.
Yet it doesn't go silent at 15001Hz, just less efficient.
As stated in the article, they use 18kHz : at that frequency it's hard for human to really notice, but speakers are still not so bad, you can at least transmit simple beep-codes.
Regarding microphone : yes, they *are* sensitive at much higher frequencies.
It's hard to do a perfect frequency filter that stops abruptely right at the 15kHz limit.
Much easier to use a microphone that picks way higher than this, but at least doesn't muffle in the audible zone at all,
record it with a ADC at a high sample rate, and then clean the sound digitally.
(That the actually real reason of 192kHz ADC/DAC you see in most pro-level equipment. Not that it makes any sense to keep a 192kHz audio track all the way to the customer. But at least if the recording stage works at a higher frequency, there's less risk of muffling the interesting frequencies or have distrosion that spill out into the audible range).
So a smartphone could pick a bit of 18+ kHz sound (at least something really easy to distinguish, like beep-codes).
Even if it's not optimized to record calls in that range, it's still designed in a way that avoid muffling under the 15kHz range.
This story is complete bullshit.
yet, US communication actually work, and is used in the wild.
When Spotify detect a device logged into your account on the same network, it can use ultra sound to match which device/speaker is connected where - i.e.: it can do ultra-sound pairing.
(It says clearly on the message box that appears).
Chromecast has also been reported to support such ultrasound pairing.
At least Spotify clearly states when it uses ultra-sounds to identify which device is connected to which speaker within which range.
Seems that dogs couldn't really perceive motion on older TVs because the framerate and resolution were too low.
Carnivore (predator) pets like dogs and cats tend to be much more sensitive to motion.
They will *perceive* motion on TV, it will just look more choppy and flickering to them.
Just like human where able to perceive motion in silent film era's 12-16fps, in half-rate/dupe-frame 12-15fps animation, or in "shitty low"-fps GIFs.
It looks a lot more choppy, than a 24/30fps or even a 48/60fps.
Or just like human *can* see the flicker of a 60Hz CRT monitor when looked at the periphery of the view (i.e.: where there are more rods - sensors with faster response that are also responsible for the pets better motion sensitivity).
I had my cats recognize and react to things on my old 50Hz CRT, even if *I* could notice the flickering.
Hopefully, as more of these patents expire, a Vulcan based open source GPU will emerge.
As the numerous past failures of attempts at opensource GPU or even opensource graphic cards have shown, making a functional and competitive graphic card is EXTREMELY difficult.
There's a reason that nVidia and AMD both suddenly implemented tile based renderers,
Actually, Nvidia has had their own TBR patents for quite some time :
- Nvidia bought up 3DFx for their patents and their engineer back when that one went bankrupt.
- Before that, 3DFx had bought up Gigapixel, among other for their TBR patents, to be used in future product (forgot the code name) - and HSR (hidden surface removal) tech to be applied much earlier in then current product (in the then VSA-100 / Voodoo4/5/6 and in the upcoming Rampage / Spectre)
So Nvidia indirectly acquired TBR patents.
Though for the record, they were more interested in the know-how and engineer which where working on the Rampage GPU ("3DFx Spectre" cards) due to programmable pixel shaders being all the rage, and retained them to work on GeForce FX (speculation backthen that probably the pun in the name was intended... )
So in theory, they could have moved into the field much faster than ATI / AMD.
(But back at the Rampage / GeForce FX era, there where some area were TBR was problamatic : e.g. some transparency (i.e.: simple alpha-blend, back then) couldn't be handled in a single pass easily. So probably they decided not to bother.
Given that modern games work with tons of subsequent passes (transparent materials cause diffraction/distortion in a separate pass of a pixel-shader), I would suspect that it's not that much a problem anymore).
What is this 'DRM' thingy you seem to be afraid of ?
(Note: I legally obtain the book I'm DeDRM-ing.
I'm just removing the DRM because I'm fed up with the Adobe Digital Edition fucking things up on a regular basis and access to my book getting b0rked yet again.
This kind of De-DRM-ing is actually tolerated in my local jurisdiction - as it should be everywhere)
when HVEC is much smaller than H264 anyway
Yeah but not by that *much* (though it depends on the quality of both encoders - x264 is incredibly better visually than nearly everything else).
At least, not given the quality/bitrate that the industry has decided to use for Ultra HD (where it makes sense or if it's mostly a placebo is an entirely different can of worm).
Also HEVC is patent minefield (and thus hardware HEVC/H265 decoding isn't as widespread as AVC/H264), so perhaps they also want to keep a door open for content producers that can only afford the MPEG patent pool for H264/AVC and would prefer to keep that codec.
Of course, next-next generations codecs like AOMedia's AV-1 are right around the corner, with even better performance, less patent encumbremenent (xiph and google are onboard), and lots of industry supprot (youtube and netflix - ie. a huge chunk of all watched content - are behind it. As are most current hardware constructors), so perhaps will end up soon with actually smaller files (and thus smaller discs or - more likely - even smaller file to torrent)
it could easily fit on dual layer BDs...
And this movie is a nice example : it's 53GB, it DOES NOT fit on 50 GB dual layer BlueRays. You would need more layers, that would drive the price of pressing discs up (i.e.: less obscene margins for the industry).
Instead UHD discs have slightly higher densities: the dual layer goes up to 66GB (so you can fit this 54GB movie on 2 layers only).
Systems like PS4 Pro (or even PS4 since it can do 4k at 30/24Hz)
Thus it would depend it the movie is 24-30 fps or 48-60 fps, and/or if there's (alterning frame) 3D.
But yeah...
could easily play them if it weren't for the drive requirement
Now the big question :
are there really that many differences between vanilla and UHD blueray disks ? or is it only a slightly higher surface density ?
I would suspect that it's only a slight difference of density, and that it should be in theory possible to reflash the drive itself with an upgraded firmware that change the behaviour of the servos of the head and the focus assembly.
(The same way it was possible to obtain a little bit higher density with floppy drivers. Except that these where directly controlled so the controlling software actually ran on the PC's main CPU, no flashing required, just a device driver).
Not that any manufacturer is actually going to release such an upgrade (it's much more lucrative to have the users re-buy a new more expensive device which is actually the same device as before (no R&D costs !) but with slightly different constants in the firmware and a different model number written on the box).
if it's on my screen, I can record it.
Yes, as in "record it with a camera".
I can output the signal and the audio on a HD recorder and there's no protection that will protect you from that.
Not exactly.
Nowadays, the output signal is *digital* (HDMI mostly with standalone players, and HDMI mostly with computers).
There's an encryption standard HDCP which is supposed to protect this data during its transit to the screen.
In theory, you should NOT be able to directly hook-up the output to a recorder, that recorder will only see an ecrypted stream that only the screen can decrypt.
You can only use a *cam* to record the actual screen as suggested above, not the stream itself.
In practice, HDCP is done poorly. Its current form is cracked and can be bypassed, so the only actual real-world is not stopping pirate, but only failing in weird ways for legitimate users.
And in actual practice : nobody gives a damn about the latest cookie-cutter soulless movie. Chances are high that I'll be too busy doing some interesting outdoor sport (or some indoor one) rather than trying to see how I could pirate a copy of Smurfs 2 (what, they even made a *second* one ?)
I never understood why you couldn't just capture the encrypted key and feed it that and let it do it's thing.
And that's exactly how it was done with the very few generations of movies.
Some BlueRay player (i think WinDVD ?) stored they decryption key in an insecure memory location, and hackers used to tap there to find which key is used to decrypt a specific BlueRay.
Movie industry noticed and revoked the keys for that player (meaning newer disc produced after that where encrypted with a selection of keys for which the player had no corresponding keys).
Cue-in cat and mouse game, until hackers managed to find the master key that sits at the top of the whole encryption pyramid.
The extra colour and dynamic range with 4k, and even more so with 8k is really nice, but to get much out of it you need a dimly lit room and a TV capable of reproducing it.
Similarly with audio you need a dedicated room and to then sit in the sweet spot while listening.
Huh... nope. 8k and 10bits colour isn't the equivalent to 200$ monster digital cables and 192kHz sample rate.
- Ears have some physiological limits due to how physics work (your ears can hear very approx in the 20-20'000Hz range. your body can also feel vibrations in the 1-100Hz. There's no receptor in a human body capable of reacting to 90kHz).
- Physics of digital signals, and a whole bunch of signal processing science (e.g.: error correction) means that in the digital world, sometime a bit is just a bit, no matter the concentration of gold and diamond powder (sic!) in the cabling it goes through.
No matter the dedicated audio room you're sitting in, you'll never be able to hear ultrasounds (directly. though ultrasounds can cause distorsions in the audible range on some equiement), and monster cables will change nothing to the SPDIF link.
The "cinephile" equivalent of an audiophile insisting on 200$ monster cables and 192kHz rates,
would be a guy who insist on movie formats that not only record Red, Green and Blue primary colours, but also infra-red and ultra violet (i.e.: insist on frequencies/wavelenghts for which the human eyes doesn't have any receptors) and on buying a $10'000 silver screen to project projecting onto which, that should also perfectly reflect x-rays, gamma rays and microwaves (completely irrelevant given what is transmitted by the light of the beamer).
No matter if the movie room is dimly lit or not, insisting on wave-lenght outside the human range (like ultra-violets) is useless, as is insisting on a screen optimized for something completely irrelevant.
The same way, no matter the dedicated listening room and it's sweet spot, a human ear lacks receptors for 96kHz sounds.
The value in a messaging system is in the users not the system.
A.k.a. "Network effect". /. is nearly entirely composed of geeks, who by definition have weird fetishes regarding technology (and not only... ahem...)
As
we are all constantly exposed to this whenever we try to persuade the vanilla people around us to try some new tech.
Tell you what: Right now where I live I can send anyone a whatsapp message, even someone who's number I just read off a business card,
The keyword is "Right now".
A couple of years ago it used to be some other system (Apple's iPhone-only iMessage).
then Facebook Messenger (still have some contact stuck on this one),
now it's WhatsApp,
next year it's going to be SnapChat (*)
And who knows what awaits us beyond.
---
(*) : As a hobby, I am a ski instructor at my university's ski camps.
I've noticed that among first-year student (~19 y.o.), very have WhatsApp installed and most of them spend their time on SnapChat.
i.e.: unlike you, I'm already in the situation where I CANNOT rely on sending a WhatsApp message to any random mobile number I happen to stumble upon.
Though (as usual with network effect) this tend to be generational so - if I know the age group of my target - I can have a vague idea on which network I have the most chance of finding them.
---
NOTE that desktop chat went through the exact same kind of madness. ICQ, AIM, Y!M, MSN, Skype, Facebook chat, Google chat/hangouts, WebEx a bunch of more modern WebRTC meeting platforms, Rocket Chat, Slack, etc.
At least during some time XMPP brought a relative hope to this madness (even Facebook used to have a XMPP gateway at some point in time), specially since servers started implementing federation (e.g.: you could talk between Google chat and Jabber.org users. Until Google decided "It's too much risk of SPAM" and disabled federation about the same time they introduced hangouts).
WhatsApp is valuable
Which actually doesn't as much show that WhatsApp has some value (as opposed to any other messaging system),
but shows how bad it is to rely on a single centralized service.
(as opposed to things which are a bit more distributed : like e-mail, like XMPP chat server with federation switched on, ...)
If the police have strong evidence that the stolen diamonds are in your wall safe, and that the things is rigged to blow if they try to crack it - do you think a judge could NOT compel you to give up the combination ?
...but you still have the right to remain silent. In the US under Miranda's rights, and under similar rights in lots of other jurisdiction elsewhere.
Note that in this case, given the *strong evidence* and and your obstination of not revealing the safe's password, you might start to look really suspicious(*) to the jury/judge (whatever is relevant in your jurisdiction).
Then again, to keep your metaphor, any thief with half a brain cell will very probably NOT stash the diamond in their own safe at home, but try to hide them in some place that nobody would think about. So they can safely show that the safe is empty of diamond, so that police can't manage to find the diamond, so competing thieves can't manage to get a grip on them, and so, once the steam blows off and nobody bother to keep an eye on the case, the thief can discretely retrieve the diamonds and enjoy them.
Here we have some idiots who decided to blackmail a victim *USING THEIR OWN FUCKING PHONES*. The devices that can easily be confiscated by police, with message logs that can be easily traced back to them, and the fact that they chose to remain silent might look a little bit strange(*).
---
(*): though not necessarily.
there's a scenario were the person might be innocent, but have some personal detail that they want to keep secret (e.g.: being gay and the publicly official partner being only "a bread". specially in culture that carries a stigma around homosexuality, or worse, in a jurisdiction with "sodomy laws" that considers it illegal)
on the other hand, if all other evidence points to guilty, they case won't be hinging on the sole content of the phone/safe...
I love how CRISPR makes geneticists act like a bunch of little kids who found a working go-cart someone left in a dumpster.
Except this time round, metaphorically they actually use it to really do go-cart laps on a nearby track (though necessarily the same track as the guy who threw it in the dumpster).
Unlike all the other times when they decide to re-purpose it for something completely different, like using it as a delivery vehicle with optional autonomous navigation (which actually works more or less),
or trying to conquer the world with it and at least managing to build a self-driving tank out of it (by strapping a bazooka on the precedent autonomous delivery vehicle).
This post was brought to you by the Slashdot's "Car Analogy Service(tm)".