MiniTruth in 1984 was supposed to *decide* (i.e.: unilaterally) what is truth and what is not. They rewrite what is considered thruth : "We have always been at war with East-Asia".
Fact-checking is supposed (in theory) to be about finding the sources of some information : They try to find where a thing is comming from : Is the number mentioned by a politician pulled-out-of-his-ass ? or is there a real article/sudy/agency reporting these numbers ?
Well, even if "fact-checking" has recently devolved into "calling each-other liars depending on political agenda".
Just like the European union has mandated USB charging for phones. They have also mandated Mennekes for AC charging and Combo for DC charging. (Some but with 2 extra pins for the DC)
(Also same for the 2 pronged un-earthed mains power, and the shutko-like europlug for earthed mains)
Everything is done so that, no matter where you travel across Europe, you can still plug and charge your electrical device, no matter if its a laptop, a smartphone or an electric car.
Except that the market has already standardized on a different set. Mennekes connectors are the current standard in Europe. (And a similar variant "Combo" exist with an extra pair of DC pins)
Tesla's charging connector is the Apple Lightning port of EV.
Menekes and Combo are the micro USB and USB-3 equivalent. (but with much better interoperability in between: closer to micro-USB 2 to micro-USB 3 rather than the more modern USB-C)
Though for Tesla's defence, even if Mennekes dates back from 2009, it was only declared official european standard in 2013, after the model S got already launched (2012).
On the other hand, from what I've read, Tesla's connector on European cars and european supercharger is modified for better interroperability with Mennekes (Yay !)... but still conveys the DC over the base pins of the european standard Mennekes AC connector, whereas the european standard calls for a Combo conenctor in case of DC.
So at least on the European market, Tesla has put some minimal efforts to be compatible with the rest, but still isn't there yet.
Much more basically, everyone else beside Tesla is standardizing on Menekes connector, on a similar one with additional pair of DC pins, and on Chademo. Nobody else is using Tesla's connector right now.
That means that, for Tesla charger to be usable by other cars, you'd need a converter anyway.
That means there's market for Tesla to starts selling a converter, that enable other cars to charge, by both adapting connector and adapting various power standards. An whose pricing includes charging fee (just as super-charging is currently covered by the "super charger" option when ordering the car. Though apparently "pay as you go" is planned for the future).
Would be even better if there was a practical way to plug other vehicles into the network. (Pay Tesla for a Menekes adapter + fee/plan to access the supercharger network ?...)
And you can expect to find it used in the wild in about a few seconds next...
(At least, luckily it got discovered though public channels : It got published by shadowbrocker and got analysed by experts. So at least our sysadmin have heard about it. Security solutions vendor will try to get ways to detect and neutralize it.
Imagine if instead it was discovered by a few blackhats who reverse engineered a sample, and decided to incorporate the technology into their exploits, without the information ever reaching the public).
But even if only foreigner and rich citizens get vaccinated :
- By being vaccinated, there's a bit less risk that they (or more correctly: their mosquitoes) will transmit malaria to somebody else. (Which would other wise would happen in the time frame between the parasite entering the body and the disease being cured by antimalaric drugs).
It won't be enough to develop herd immunity (see children viruses in developed world - at least until anti-vaxxer started messing around), but it's a tiny bit help in reducing propagation (see influenza virus in the developed world) - even if only a drop in the bucket.
- Also prevention based on vaccination of travelling foreigner, is a lot less likely than the current prevention (dosing traveller with anti-malaric drug in advance) to overload the environment with drug to the point that the parasite has pressure to mutate and develop resistance.
And this has very important long-term ecological benefits. (Less risks of drugs stopping to work eventually)
I'm under the impression that JavaScript doesn't have a standard library outside of its core functionality, but it does have a ton of frameworks available.
Yup, that's exactly my (poorly worded) complain. Tons of semi-usefull frameworks everywhere, but not a basic library of standard functions.
Leading to either tons of copy-pasting, or relying on scattered external modules.
Though, as mentionned by TFA (sorry, I read it, here I'll turn in my/. member card) they are even more obsessed with their privacy, as a significant part of Europe is.
The other country are also European (Italy, France) or similarly obsessed with privacy and averse to intrusion (Japan).
Seems that the US is actually the anomaly, having a high GDP *but* happily providing all their personal information to be abused by marketeers/advertiser, by three-letter agencies, and by pirates leaking databases and personal photo collections.
Social media - also known as mind hive CIA project
Speaking of which : notice how nearly all cited countries - Germany, France and Italy - are in Europe, and we European tend to be really serious about our privacy. And Japan is similarly concerned with privacy and not intruding onto other people.
And that not only classical social networks (like Facebook). That's also the case with chat systems. WhatsApp seems to be not as popular there are elsewhere in the world. You could find actually lots of german who prefere/have switched to other systems (basically : systems with more green checks on FSF's list)
We shouldn't be asking why people are copying bad code, we should be asking why they need to.
Not sure why you're rephrasing my statement.
There's a subtle difference.
Your statement clearly poses re-implementation of code as the main alternative to copy-pasting. (boils down to "You should intelligently re-implement, instead of blindy copy-pasting").
The above statement simply discourages from copy-pasting (boils down to "Do not copy-paste, why do you even want to ?") but is still open to *any* solution : that includes having a standard library (which was another criticism back during the "#LeftPadGate" ) which is also a valid solution : if there a decent standard library, nobody will need to copy-paste anymore either (but nobody will neither need to re-implement).
The underlying problem is that too many programmers are willing to copy and paste code rather than think through what they need to code. Remember the left-pad crisis that broke the Internet because a developer removed his npm packages over a dispute? How hard is to write a left-pad function?
Sorry but now.
You should not be copy-pasting a left pad function. But, you should not be re-implementing yet another one yourself neither.
Simple trivial task like this *should go into a standard library*.
On any machine on which I fire up a C compiler, I know that at least I can rely on a decent compliant standard library for simple task. If I want to left-pad a number, I just give the appropriate parameter to printf. (Well unless I'm writing kernel code, or unless I'm writing for an tiny embed platform where every byte counts and I need as specific code as possible).
Why does Npm needs to be any different ?
---
Also, left-padding function might be not as trivial as you think. Not every language is english, not every language is written only with the ASCII subset of unicode. Some weird corner cases will start to popup. Think situation where : Number of bytes IS NOT number of unicode code point which in turn IS NOT the number of displayed characters (e.g.: some of the unicode are diacritics or other such modifiers) (To think about worste case scenarios: How do you even left-pad Zalgo ?) But these are indeed extreme cases.
The guy hacked A UNIX NETWORK! I heard those networks are hardcore, some even use the vi protocol to load balance the kernel across multiple NFS loopbacks. It's basically POSIX grade security with layers upon layers of nmaps.
But is this UNIX webscale ? And does it enough Apps to synergize the user experience integration with cyberwarfare cryptosecurity ?
So it is basically Android x86 which is indeed an unstable pile of shit - unless ran on the hardware that the "owner" of the x86 project provides from a parallel commercial company he runs
I strongly suspect that most of the problems with android kernels are the hardware drivers : for the GPU, the wireless chips, etc. (because there aren't much kernel drivers which support android's unusuall ABI. Most free/libre linux kernel drivers use GNU/Linux's DRI API for graphics instead of Android's Flinger)
Here the situation is a bit different :
- in the special case of the GPU, anbox uses a facility which is normally used by the emulators (like QEMU) for accelerated emulated graphics. Graphics command are recorded by the container android compositor, and then sent to a daemon on the host whose role is to draw them on a regular Linux desktop over OpenGL. There's no requirement for the host to support Android's ABI.
- network is isolated. guest can't have access to the actual network device (no direct Wifi chip control). container usually sees only a bridge device.
- Bluetooth isn't currently supported inside the container.
So once anbox leaves beta, it might end up being more stable than x86 android counting on special kernel API/ABI
Not on all of them, and not in all market. But still, Dells are so much popular, that even for the few Lattitude that you can directly get with Ubuntu pre-installed, you can just pop your Suse CD in and install a tumbleweed, because of popularity, lots of people would have tester a tweaked what is necessary for the distro to work out of the box.
And the Microsoft tax is a myth. All the demo software they tend to put on pays for windows plus a bit more so Linux laptops can often be a bit more because they are unsubsidised.
Yup. Totally agree with you. That's why I was saying:
linux laptops not only come *without the Microsoft tax* (making them a bit cheaper), they also come *without the Bloatware/Crapware bonus* (making them not heavily subsidized by "Punch the Monkey to win big prizes !" and "Let's siphon all your data straight to all the marketeer we an find".
Except virtually no Android apps even bother to make their apps work well on slightly larger tablet screen. {...} But for the vast majority of Android apps, the actual UI experience is designed for an approx. 6" diagonal vertically oriented screen with only touch-input. Virtually no apps are designed or redesigned to even work well on a slightly larger tablet screen.
My personnal experience differs (10" asian tablet here). Of course, your personal annecdotal experience is just one data point as well as mine.
But I've seen several applications which work flawlessly on the 10" portrait (16:9 widescreen) tablet. None of those that I regularily use pose any problem. (Of the top of my head: Firefox, VLC, Google Maps, Google Calendar, the finding/renting software of a couple of car-sharing services, a few games, several chat programs both personnal (Skype, etc.) and professional (Cisco WebEx, Adobe Connect), Orbot (=tor), ConnectBot, ES Files, Netflix, Spotify, etc.) They are either trivial (e.g.: Orbot - it's just a single "Start TOR" button in the middle) or they are correctly adapted.
And again, what the parent poster misses most are a few key apps (banking OTP / 2FA) which tend to have a trivial layout.
Any non-trivial app will absolutely suck to use on a 27" horizontally oriented screen using a keyboard and mouse.
Yes, I get it, like me and most other people on/. you're a dev. We tend to have our IDE open full screen on giant 27" 4k desktops, full with tools box and other small windows which completely clutter the screen. But you, me and the rest of/. aren't typical users.
Most more "regular people" I know - even those with Linux installed on their laptop - tend to be of the 13-15" laptop HD screen crowd. And again, thank to the same extension that enable side-by-side display on tablets, Anbox should allow you to open application in windows. So in the end, it's not the a smartphone game will be so much scaled up, that you can see individual pixels, each the size of your normal desktop icon.
It's more that a laptop user will be able to open a small/medium window with an android app running inside. Not that much different than any flash games.
Hell, WINE or a virtual PC environment will give you much better apps than Android will.
If the app do actually have a Windows desktop port. There are a lot of people who use their iOS or Android smartphone or tablet as their main computing platform. Thus there are a lot of things which are primarily targeting pocket devices. e.g.: There is no such thing as a desktop Windows or Mac OS X verison of Instagram.
the typical non-dev hobbyist who's likely to install Linux on their laptop, would be also more likely to be interested in trying android-only apps on a laptop.
They run a virtual machine. A whole Android computer is emulated. A complete stack, including the anrdoid userlan (of course, that's the point), but also including its own kernel which talks to the virtual hardware that is emulated inside the emulator/virtual machine, and that emulator/virtual machine is a user-land application that in turns runs on the host linux, and talk to yet another kernel, the one running on the real hardware.
In short you have two completey stack kernel+userspace, and a heavy emulation layer that simulates a whole imaginary computer in between. (that layer can even have a CPU emulator, to run ARM apps on AMD64 hardware).
Solution such as Anbox : - Add a few key missing feature to the host's kernel (e.g.: Android needs Binder, desktop seldom have it). - Simply run the Android userland in a chroot - Android talks to the real kernel running on the real hardware.
(Well not exactly a chroot. LXC uses namespace - like chroot but isolates many more things).
The point: Now you avoid an entire extra layer simulating a computer in which you run a whole stack with a kernel. As sais, android userland talk directly to your desktop's kernel.
If you play close attention : on any user space that lacked Android support up until now.
Anbox is a combination of 3 things : - an LXC container containing the Android userland - compiling android-only kernel extensions so the container can actually find them in the kernel (e.g.: Binder) - using a forwarder normally designed for the emulator that will forward a few things (like graphics). Because the container is isolated from nearly all hardware accesses.
As long as your weird user-space can hangle this, you can give it a try. (GNU/Linux of course. But Android might be able to host such a thing. Maybe even find a way to get a Busybox/uLibc user space run this)
Actually there is a limitation, but the above poster is wrong: it's not a kernel one, it's a hardware one. - Android expects hardware floating point (e.g.: armhf). But Raspberry Pi 1 lacks them.
Also this requires Android directly talking to the host kernel. And not all Android Kernel extension (that are needed in this case) have been thoroughly tested with raspberry pi. Expect to stumble unpon even more never-before-seen bugs.
either: - in android's architecture-neutral bytecode ("I can't believe it's not Java(tm) !") - in the same native architecture as your main OS, because there's no emulator, Anbox runs in a container, thus interecting straight with your current kernel.
Currently supported architecture lists: AMD64 (obviously), but also ARMHF and ARM64.
So you can install an ARM app, as long as you do it on a compatible Raspberry Pi, or Pyra, etc..
But again, the whole thing is currently alpha. So for the next few months don't excpet much except a lot of crashes, specially if you're not running the same kind of configuration as most other testers. (You'll find way more bugs)
So, you can use apps "designed" for a pretty small screen with a touch-only UI, on a large monitor with a keyboard/mouse?
Actually, android also runs on tablets so not only small screens.
Nearly every last android device has a OTG port, and Android supports mouse and keyboard input out of the box (yup, plugin a powered hub, an USB keyboard and USB mouse and, suprise, everything works as expected.) (There are even small accessory manufacturer specializing in hardware keyboards with USB and Bluetooth for Android).
Not only that, but since recently Android even support variable-sized windows. On actual real tablets, this is used to enable to tile application side-by-side. But in Anbox it's used to enable application to run as arbitrary windows on the desktop.
And the parent was asking about an App that simply works as a glorified OTP. So that's definitely something which isn't entirle built around a phone UX.
I think they did some years ago, it's called "git".
Sparkle Share is an example of document tracking built aroudn "git".
Also there are examples such as ownCloud / NextCloud. This last one is getting so much popular that it has seen official deployment in some universities. (e.g.: Switch is providing country-wide installation for Swiss Universities)
I understand wanting to support the Linux community, but I thought one of the "big advantages" of Linux was that it was cheaper? Yet here, even without the Microsoft tax, it costs a lot more.
The thing is, unlike your custom self-build linux workstation, linux laptops not only come *without the Microsoft tax* (making them a bit cheaper), they also come *without the Bloatware/Crapware bonus* (making them not heavily subsidized by "Punch the Monkey to win big prizes !" and "Let's siphon all your data straight to all the marketeer we an find".
They also don't come with the *integrated by chinese almost-slave labour rebate*. Laptop tend to be complex and weird (embed controllers, etc.) which requires a tiny bit of adaptation to make them linux-worthy.
- When you buy a big popular brand like Lenovo's Thinkpad Ts, Dell's Lattitude, etc. someone else would have done the debugging already (see ressources like Thinkwiki) and by that time it'll probably be upstream in vanilla kernel and standard distros. So you can probably just pop in a CD of Ubuntu or Linux Mint and it will install flawlessly.
- When you start with less popular manufacturer, you'd be in for a few small surprise : screen not turning on, kernel crashing at boot when trying to enumerate hardware, UEFI-Secure refusing the signature of your bootloader's shim, etc. You could be needed fixes in the firmware and/or workaround patched in the kernel. It might something really simple (just hacking a bit some settings). But even that "simple" will by done by some who isn't paid in cents per day range.
up until now System76 were selling - laptops which were simply re-branded laptops from other brand, to which they changed firmware and OS to a more free option - desktop which were mostly of the shelf beige-boxes i.e.: they were selling mostly 3rd party hardware
starting from next year, they also want : - laptop that they make themselves (well, most likely they will be still produced in china. but the idea is that the models are now made by System76. Not Lenovo models with an alternate firmware and OS). This is interesting because in the end it will enable them to better choose the component inside the laptop for Linux compatibility (avoid too much weird embed controllers) - desktop designed by themselves too. (that won't be a much big change from the current beige box trend. A motherboard is still a motherboard). but at least it will help with brand identity and will also help testing their design pipeline on a smaller scale before tackling the laptops.
Their blog post make it clearer (I swear I didn't click TFA's link ! I just clicked last week, when it was on Phoronix. Am I still/.-worthy ?:-D ) Sadly the summery sounds like you need desktop cases specially made compatible with Linux.
Slashdot Mirror
There's a bit of a difference.
MiniTruth in 1984 was supposed to *decide* (i.e.: unilaterally) what is truth and what is not.
They rewrite what is considered thruth : "We have always been at war with East-Asia".
Fact-checking is supposed (in theory) to be about finding the sources of some information :
They try to find where a thing is comming from : Is the number mentioned by a politician pulled-out-of-his-ass ? or is there a real article/sudy/agency reporting these numbers ?
Well, even if "fact-checking" has recently devolved into "calling each-other liars depending on political agenda".
So plugs are an exception now?
Yes, plugs are actually an exception. Really.
Just like the European union has mandated USB charging for phones.
They have also mandated Mennekes for AC charging and Combo for DC charging. (Some but with 2 extra pins for the DC)
(Also same for the 2 pronged un-earthed mains power, and the shutko-like europlug for earthed mains)
Everything is done so that, no matter where you travel across Europe, you can still plug and charge your electrical device, no matter if its a laptop, a smartphone or an electric car.
Except that the market has already standardized on a different set.
Mennekes connectors are the current standard in Europe.
(And a similar variant "Combo" exist with an extra pair of DC pins)
Tesla's charging connector is the Apple Lightning port of EV.
Menekes and Combo are the micro USB and USB-3 equivalent.
(but with much better interoperability in between: closer to micro-USB 2 to micro-USB 3 rather than the more modern USB-C)
Though for Tesla's defence, even if Mennekes dates back from 2009, it was only declared official european standard in 2013, after the model S got already launched (2012).
On the other hand, from what I've read, Tesla's connector on European cars and european supercharger is modified for better interroperability with Mennekes (Yay !)... but still conveys the DC over the base pins of the european standard Mennekes AC connector, whereas the european standard calls for a Combo conenctor in case of DC.
So at least on the European market, Tesla has put some minimal efforts to be compatible with the rest, but still isn't there yet.
Much more basically, everyone else beside Tesla is standardizing on Menekes connector, on a similar one with additional pair of DC pins, and on Chademo.
Nobody else is using Tesla's connector right now.
That means that, for Tesla charger to be usable by other cars, you'd need a converter anyway.
That means there's market for Tesla to starts selling a converter, that enable other cars to charge, by both adapting connector and adapting various power standards. An whose pricing includes charging fee (just as super-charging is currently covered by the "super charger" option when ordering the car. Though apparently "pay as you go" is planned for the future).
Would be even better if there was a practical way to plug other vehicles into the network.
(Pay Tesla for a Menekes adapter + fee/plan to access the supercharger network ?...)
And you can expect to find it used in the wild in about a few seconds next...
(At least, luckily it got discovered though public channels : It got published by shadowbrocker and got analysed by experts.
So at least our sysadmin have heard about it.
Security solutions vendor will try to get ways to detect and neutralize it.
Imagine if instead it was discovered by a few blackhats who reverse engineered a sample, and decided to incorporate the technology into their exploits, without the information ever reaching the public).
But even if only foreigner and rich citizens get vaccinated :
- By being vaccinated, there's a bit less risk that they (or more correctly: their mosquitoes) will transmit malaria to somebody else.
(Which would other wise would happen in the time frame between the parasite entering the body and the disease being cured by antimalaric drugs).
It won't be enough to develop herd immunity (see children viruses in developed world - at least until anti-vaxxer started messing around), but it's a tiny bit help in reducing propagation (see influenza virus in the developed world) - even if only a drop in the bucket.
- Also prevention based on vaccination of travelling foreigner, is a lot less likely than the current prevention (dosing traveller with anti-malaric drug in advance) to overload the environment with drug to the point that the parasite has pressure to mutate and develop resistance.
And this has very important long-term ecological benefits. (Less risks of drugs stopping to work eventually)
I'm under the impression that JavaScript doesn't have a standard library outside of its core functionality, but it does have a ton of frameworks available.
Yup, that's exactly my (poorly worded) complain.
Tons of semi-usefull frameworks everywhere,
but not a basic library of standard functions.
Leading to either tons of copy-pasting, or relying on scattered external modules.
does not necessarily lead to productivity.
Though, as mentionned by TFA (sorry, I read it, here I'll turn in my /. member card) they are even more obsessed with their privacy, as a significant part of Europe is.
The other country are also European (Italy, France) or similarly obsessed with privacy and averse to intrusion (Japan).
Seems that the US is actually the anomaly, having a high GDP *but* happily providing all their personal information to be abused by marketeers/advertiser, by three-letter agencies, and by pirates leaking databases and personal photo collections.
Social media - also known as mind hive CIA project
Speaking of which : notice how nearly all cited countries - Germany, France and Italy - are in Europe, and we European tend to be really serious about our privacy.
And Japan is similarly concerned with privacy and not intruding onto other people.
And that not only classical social networks (like Facebook).
That's also the case with chat systems. WhatsApp seems to be not as popular there are elsewhere in the world. You could find actually lots of german who prefere/have switched to other systems (basically : systems with more green checks on FSF's list)
We shouldn't be asking why people are copying bad code, we should be asking why they need to.
Not sure why you're rephrasing my statement.
There's a subtle difference.
Your statement clearly poses re-implementation of code as the main alternative to copy-pasting. (boils down to "You should intelligently re-implement, instead of blindy copy-pasting").
The above statement simply discourages from copy-pasting (boils down to "Do not copy-paste, why do you even want to ?") but is still open to *any* solution :
that includes having a standard library (which was another criticism back during the "#LeftPadGate" ) which is also a valid solution : if there a decent standard library, nobody will need to copy-paste anymore either (but nobody will neither need to re-implement).
The underlying problem is that too many programmers are willing to copy and paste code rather than think through what they need to code.
Remember the left-pad crisis that broke the Internet because a developer removed his npm packages over a dispute? How hard is to write a left-pad function?
Sorry but now.
You should not be copy-pasting a left pad function.
But, you should not be re-implementing yet another one yourself neither.
Simple trivial task like this *should go into a standard library*.
On any machine on which I fire up a C compiler, I know that at least I can rely on a decent compliant standard library for simple task.
If I want to left-pad a number, I just give the appropriate parameter to printf.
(Well unless I'm writing kernel code, or unless I'm writing for an tiny embed platform where every byte counts and I need as specific code as possible).
Why does Npm needs to be any different ?
---
Also, left-padding function might be not as trivial as you think. Not every language is english, not every language is written only with the ASCII subset of unicode. Some weird corner cases will start to popup. Think situation where : Number of bytes IS NOT number of unicode code point which in turn IS NOT the number of displayed characters (e.g.: some of the unicode are diacritics or other such modifiers)
(To think about worste case scenarios: How do you even left-pad Zalgo ?)
But these are indeed extreme cases.
The guy hacked A UNIX NETWORK! I heard those networks are hardcore, some even use the vi protocol to load balance the kernel across multiple NFS loopbacks. It's basically POSIX grade security with layers upon layers of nmaps.
But is this UNIX webscale ? And does it enough Apps to synergize the user experience integration with cyberwarfare cryptosecurity ?
So it is basically Android x86 which is indeed an unstable pile of shit - unless ran on the hardware that the "owner" of the x86 project provides from a parallel commercial company he runs
I strongly suspect that most of the problems with android kernels are the hardware drivers : for the GPU, the wireless chips, etc.
(because there aren't much kernel drivers which support android's unusuall ABI.
Most free/libre linux kernel drivers use GNU/Linux's DRI API for graphics instead of Android's Flinger)
Here the situation is a bit different :
- in the special case of the GPU, anbox uses a facility which is normally used by the emulators (like QEMU) for accelerated emulated graphics.
Graphics command are recorded by the container android compositor, and then sent to a daemon on the host whose role is to draw them on a regular Linux desktop over OpenGL.
There's no requirement for the host to support Android's ABI.
- network is isolated. guest can't have access to the actual network device (no direct Wifi chip control). container usually sees only a bridge device.
- Bluetooth isn't currently supported inside the container.
So once anbox leaves beta, it might end up being more stable than x86 android counting on special kernel API/ABI
Dell sells Linux laptops.
Not on all of them, and not in all market.
But still, Dells are so much popular, that even for the few Lattitude that you can directly get with Ubuntu pre-installed, you can just pop your Suse CD in and install a tumbleweed, because of popularity, lots of people would have tester a tweaked what is necessary for the distro to work out of the box.
And the Microsoft tax is a myth. All the demo software they tend to put on pays for windows plus a bit more so Linux laptops can often be a bit more because they are unsubsidised.
Yup. Totally agree with you. That's why I was saying :
linux laptops not only come *without the Microsoft tax* (making them a bit cheaper), they also come *without the Bloatware/Crapware bonus* (making them not heavily subsidized by "Punch the Monkey to win big prizes !" and "Let's siphon all your data straight to all the marketeer we an find".
Except virtually no Android apps even bother to make their apps work well on slightly larger tablet screen. {...} But for the vast majority of Android apps, the actual UI experience is designed for an approx. 6" diagonal vertically oriented screen with only touch-input. Virtually no apps are designed or redesigned to even work well on a slightly larger tablet screen.
My personnal experience differs (10" asian tablet here).
Of course, your personal annecdotal experience is just one data point as well as mine.
But I've seen several applications which work flawlessly on the 10" portrait (16:9 widescreen) tablet.
None of those that I regularily use pose any problem.
(Of the top of my head: Firefox, VLC, Google Maps, Google Calendar, the finding/renting software of a couple of car-sharing services, a few games, several chat programs both personnal (Skype, etc.) and professional (Cisco WebEx, Adobe Connect), Orbot (=tor), ConnectBot, ES Files, Netflix, Spotify, etc.)
They are either trivial (e.g.: Orbot - it's just a single "Start TOR" button in the middle) or they are correctly adapted.
And again, what the parent poster misses most are a few key apps (banking OTP / 2FA) which tend to have a trivial layout.
Any non-trivial app will absolutely suck to use on a 27" horizontally oriented screen using a keyboard and mouse.
Yes, I get it, like me and most other people on /. you're a dev. We tend to have our IDE open full screen on giant 27" 4k desktops, full with tools box and other small windows which completely clutter the screen. /. aren't typical users.
But you, me and the rest of
Most more "regular people" I know - even those with Linux installed on their laptop - tend to be of the 13-15" laptop HD screen crowd.
And again, thank to the same extension that enable side-by-side display on tablets, Anbox should allow you to open application in windows.
So in the end, it's not the a smartphone game will be so much scaled up, that you can see individual pixels, each the size of your normal desktop icon.
It's more that a laptop user will be able to open a small/medium window with an android app running inside.
Not that much different than any flash games.
Hell, WINE or a virtual PC environment will give you much better apps than Android will.
If the app do actually have a Windows desktop port.
There are a lot of people who use their iOS or Android smartphone or tablet as their main computing platform.
Thus there are a lot of things which are primarily targeting pocket devices.
e.g.: There is no such thing as a desktop Windows or Mac OS X verison of Instagram.
the typical non-dev hobbyist who's likely to install Linux on their laptop, would be also more likely to be interested in trying android-only apps on a laptop.
I will explain it to you - it runs an entire Android userspace.
Yeah, and that's the entire point.
Before, you would need to fire up something like Qemu, and emulated a complete android machine with a running kernel inside.
Now said Android userspace can talk straight to the kernel, without any emulation layer.
That's the whole fucking point of this.
Solution to run Android APPs before :
They run a virtual machine. A whole Android computer is emulated.
A complete stack, including the anrdoid userlan (of course, that's the point), but also including its own kernel which talks to the virtual hardware that is emulated inside the emulator/virtual machine, and that emulator/virtual machine is a user-land application that in turns runs on the host linux, and talk to yet another kernel, the one running on the real hardware.
In short you have two completey stack kernel+userspace, and a heavy emulation layer that simulates a whole imaginary computer in between.
(that layer can even have a CPU emulator, to run ARM apps on AMD64 hardware).
Solution such as Anbox :
- Add a few key missing feature to the host's kernel (e.g.: Android needs Binder, desktop seldom have it).
- Simply run the Android userland in a chroot
- Android talks to the real kernel running on the real hardware.
(Well not exactly a chroot. LXC uses namespace - like chroot but isolates many more things).
The point: Now you avoid an entire extra layer simulating a computer in which you run a whole stack with a kernel.
As sais, android userland talk directly to your desktop's kernel.
If you play close attention :
on any user space that lacked Android support up until now.
Anbox is a combination of 3 things :
- an LXC container containing the Android userland
- compiling android-only kernel extensions so the container can actually find them in the kernel (e.g.: Binder)
- using a forwarder normally designed for the emulator that will forward a few things (like graphics). Because the container is isolated from nearly all hardware accesses.
As long as your weird user-space can hangle this, you can give it a try.
(GNU/Linux of course. But Android might be able to host such a thing. Maybe even find a way to get a Busybox/uLibc user space run this)
Actually there is a limitation, but the above poster is wrong: it's not a kernel one, it's a hardware one.
- Android expects hardware floating point (e.g.: armhf). But Raspberry Pi 1 lacks them.
Also this requires Android directly talking to the host kernel.
And not all Android Kernel extension (that are needed in this case) have been thoroughly tested with raspberry pi.
Expect to stumble unpon even more never-before-seen bugs.
The summary is badly worded.
you can install apps but they need to be
either:
- in android's architecture-neutral bytecode ("I can't believe it's not Java(tm) !")
- in the same native architecture as your main OS, because there's no emulator, Anbox runs in a container, thus interecting straight with your current kernel.
Currently supported architecture lists: AMD64 (obviously), but also ARMHF and ARM64.
So you can install an ARM app, as long as you do it on a compatible Raspberry Pi, or Pyra, etc..
But again, the whole thing is currently alpha. So for the next few months don't excpet much except a lot of crashes, specially if you're not running the same kind of configuration as most other testers.
(You'll find way more bugs)
So, you can use apps "designed" for a pretty small screen with a touch-only UI, on a large monitor with a keyboard/mouse?
Actually, android also runs on tablets so not only small screens.
Nearly every last android device has a OTG port, and Android supports mouse and keyboard input out of the box (yup, plugin a powered hub, an USB keyboard and USB mouse and, suprise, everything works as expected.) (There are even small accessory manufacturer specializing in hardware keyboards with USB and Bluetooth for Android).
Not only that, but since recently Android even support variable-sized windows.
On actual real tablets, this is used to enable to tile application side-by-side.
But in Anbox it's used to enable application to run as arbitrary windows on the desktop.
And the parent was asking about an App that simply works as a glorified OTP. So that's definitely something which isn't entirle built around a phone UX.
I think they did some years ago, it's called "git".
Sparkle Share is an example of document tracking built aroudn "git".
Also there are examples such as ownCloud / NextCloud.
This last one is getting so much popular that it has seen official deployment in some universities.
(e.g.: Switch is providing country-wide installation for Swiss Universities)
I understand wanting to support the Linux community, but I thought one of the "big advantages" of Linux was that it was cheaper? Yet here, even without the Microsoft tax, it costs a lot more.
The thing is, unlike your custom self-build linux workstation, linux laptops not only come *without the Microsoft tax* (making them a bit cheaper), they also come *without the Bloatware/Crapware bonus* (making them not heavily subsidized by "Punch the Monkey to win big prizes !" and "Let's siphon all your data straight to all the marketeer we an find".
They also don't come with the *integrated by chinese almost-slave labour rebate*.
Laptop tend to be complex and weird (embed controllers, etc.) which requires a tiny bit of adaptation to make them linux-worthy.
- When you buy a big popular brand like Lenovo's Thinkpad Ts, Dell's Lattitude, etc. someone else would have done the debugging already (see ressources like Thinkwiki) and by that time it'll probably be upstream in vanilla kernel and standard distros. So you can probably just pop in a CD of Ubuntu or Linux Mint and it will install flawlessly.
- When you start with less popular manufacturer, you'd be in for a few small surprise : screen not turning on, kernel crashing at boot when trying to enumerate hardware, UEFI-Secure refusing the signature of your bootloader's shim, etc. You could be needed fixes in the firmware and/or workaround patched in the kernel. It might something really simple (just hacking a bit some settings).
But even that "simple" will by done by some who isn't paid in cents per day range.
So it adds up to the costs.
The summary is badly worder.
The thing is :
up until now System76 were selling
- laptops which were simply re-branded laptops from other brand, to which they changed firmware and OS to a more free option
- desktop which were mostly of the shelf beige-boxes
i.e.: they were selling mostly 3rd party hardware
starting from next year, they also want :
- laptop that they make themselves (well, most likely they will be still produced in china. but the idea is that the models are now made by System76. Not Lenovo models with an alternate firmware and OS).
This is interesting because in the end it will enable them to better choose the component inside the laptop for Linux compatibility (avoid too much weird embed controllers)
- desktop designed by themselves too. (that won't be a much big change from the current beige box trend. A motherboard is still a motherboard).
but at least it will help with brand identity and will also help testing their design pipeline on a smaller scale before tackling the laptops.
Their blog post make it clearer (I swear I didn't click TFA's link ! I just clicked last week, when it was on Phoronix. Am I still /.-worthy ? :-D )
Sadly the summery sounds like you need desktop cases specially made compatible with Linux.