Judging by the FAQ, they give parameters about the network to the computer and the computer comes up with a better algoritm, but when the network does not behave within the give parameters the algoritm performance starts to suck quickly.
I think I read/heared somewhere that they can do this because the top part is handled by Mozilla and only the bottom part, like the kernel is handled by the carriers.
When was the last time you've seen an exploit only written in Javascript only (not abuse plugins and so on) that would own a computer in the wild on an up to date browser ?
TeamViewer traffic gets routed through their servers too (unless you are using it on the same network, but you don't control that).
Keep an eye on WebRTC protocol, they will add screen sharing (not just for your browser, there are libraries for mobile and desktop apps too). Then you are in control.
Evercookie and profiling prove that anything technical can be used to track you on the web.
If you don't want to be tracked, you'll need every browser in the world to be the same implementation (make, brand, code and version !) on the same screen on the same hardware and using Tor.
Then, maybe, all browsers will look the same and they'll not be able to track you.
So if you believe that, then there is only one solution, a way for the user to communicate to the site he/she doesn't want to be tracked and a law which forces companies to comply with the wishes of the users.
There are 8 million webdevelopers and a couple of hundred thousand developers for iOS and Android.
That is why some think it might have some future, will it be the third platform ? It's goal is to give people all over the world better access to the web on mobile. It isn't trying to be first, second, maybe not even the third platform.
I've done webdevelopment for years, there are things wrong with Javascript, there are things wrong with many languages. And if you stick to the subset people call 'the good parts' it's not bad.
Your argument about untyped code is a non-argument, it has never been an issue. It's a matter of programming style. With most styles of programming you'll never encounter a problem.
What it does do is create a performance ceiling, because untyped is hard to optimize for a VM.
But that is what asm.js is for now, for the performance critical code paths.
I had a look at how YouGov works (based on Wikipedia at least) they work with panels with as much diversity as possible, they are active in the US since 2007, so I guess a 1000 might be OK. I thought it was some online site where people vote on their own initiative, it isn't.
What I don't understand is how the a poll with only 1000 people could be somehow regarded as representative for 300.000.000+ people.
A poll of a 1000 people isn't even thought of as representative for my country which only has 16.000.000+ people.
That is what the first comment should have been about.
These types of things tell me how people in the US have lost touch with reality, please, please be more critical of the media and everything else. Apply more common sense.
The problem is, like in the US, I've not seen any large political party in any country say they wanted to get rid of this. So you'll have to behead all people in government to get rid of this.
The four largest net contributors in absolute terms are Germany, France, Italy, UK The four largest net contributors in per capita terms are Denmark, Finland, Germany, Italy.
The previous person working on that code also had a fear, a fear of losing his job. So he or she thought job security could be increased by not documenting and not adding tests;-)
That was my first thought too, but I don't see it in the Chromium code base.
The change that is refered to in the article is just enabling the use of the application. Which might mean probably it's just available in ChromeOS (and later in Chrome).
This is the text of a slide of the presentation linked below:
"A Unified Approach? Should we compile entire VMs from C/C++ to JavaScript, and implement JavaScript-emitting JITs? Seems the only way to run most languages with perfect semantics + maximum speed This is why I believe C/C++ to JavaScript translation is the core issue regarding compilation to JavaScript"
Slashdot Mirror
Judging by the FAQ, they give parameters about the network to the computer and the computer comes up with a better algoritm, but when the network does not behave within the give parameters the algoritm performance starts to suck quickly.
That doesn't sound all that revolutionary.
I think I read/heared somewhere that they can do this because the top part is handled by Mozilla and only the bottom part, like the kernel is handled by the carriers.
When was the last time you've seen an exploit only written in Javascript only (not abuse plugins and so on) that would own a computer in the wild on an up to date browser ?
TeamViewer traffic gets routed through their servers too (unless you are using it on the same network, but you don't control that).
Keep an eye on WebRTC protocol, they will add screen sharing (not just for your browser, there are libraries for mobile and desktop apps too). Then you are in control.
They have a profile about you even without gmail, google plus, google account.
They don't care about your name, they want to make profile they can can target for advertising.
Evercookie and profiling prove that anything technical can be used to track you on the web.
If you don't want to be tracked, you'll need every browser in the world to be the same implementation (make, brand, code and version !) on the same screen on the same hardware and using Tor.
Then, maybe, all browsers will look the same and they'll not be able to track you.
So if you believe that, then there is only one solution, a way for the user to communicate to the site he/she doesn't want to be tracked and a law which forces companies to comply with the wishes of the users.
They want to bring it back and I'm fairly certain they are going to try it too.
You are doing an awesome job, thank you and your employer for sponsoring your work.
It is, not only that, I think the split between developed by Mozilla and the community is 60%/40%, I just don't remeber which had the larger share.
There are 8 million webdevelopers and a couple of hundred thousand developers for iOS and Android.
That is why some think it might have some future, will it be the third platform ? It's goal is to give people all over the world better access to the web on mobile. It isn't trying to be first, second, maybe not even the third platform.
I've done webdevelopment for years, there are things wrong with Javascript, there are things wrong with many languages. And if you stick to the subset people call 'the good parts' it's not bad.
Your argument about untyped code is a non-argument, it has never been an issue. It's a matter of programming style. With most styles of programming you'll never encounter a problem.
What it does do is create a performance ceiling, because untyped is hard to optimize for a VM.
But that is what asm.js is for now, for the performance critical code paths.
The reason is there are 8 million webdevelopers, they like to use what they know.
It is as simple as that.
I had a look at how YouGov works (based on Wikipedia at least) they work with panels with as much diversity as possible, they are active in the US since 2007, so I guess a 1000 might be OK. I thought it was some online site where people vote on their own initiative, it isn't.
What I don't understand is how the a poll with only 1000 people could be somehow regarded as representative for 300.000.000+ people.
A poll of a 1000 people isn't even thought of as representative for my country which only has 16.000.000+ people.
That is what the first comment should have been about.
These types of things tell me how people in the US have lost touch with reality, please, please be more critical of the media and everything else. Apply more common sense.
The problem is, like in the US, I've not seen any large political party in any country say they wanted to get rid of this. So you'll have to behead all people in government to get rid of this.
Mickey Mouse from Disney is the biggest reason they've been winning, the time it takes for copyright to expire has been streched again and again.
Currently copyright expires at 75 years after the death of the author. Guess how long ago the creator of Mickey Mouse died ?
Guess what will happen again some day ?
http://en.wikipedia.org/wiki/Mickey_Mouse#Legal_issues
They are just small potatoes.
Here is a fact: the harddisk manufacturers make more money than the movie/film and music industry combined.
And have fun watching this video:
https://www.youtube.com/watch?v=zL2FOrx41N0
Per capita is what is important here:
The four largest net contributors in absolute terms are Germany, France, Italy, UK
The four largest net contributors in per capita terms are Denmark, Finland, Germany, Italy.
http://en.wikipedia.org/wiki/Budget_of_the_European_Union#Net_contributors_and_recipients
The previous person working on that code also had a fear, a fear of losing his job. So he or she thought job security could be increased by not documenting and not adding tests ;-)
Which this seems to be non of them.
This particular solution isn't an open format and the implementation doesn't seem to be open source software as far as I can see.
Maybe eventually it will be.
That was my first thought too, but I don't see it in the Chromium code base.
The change that is refered to in the article is just enabling the use of the application. Which might mean probably it's just available in ChromeOS (and later in Chrome).
Maybe I'm just looking in the wrong place...
Yes, I believe the WebGL implementations even in Chrome and Firefox on Windows are implemented in Direct3D/DirectX/whatever.
Don't count on it the wrapper is probably in IE itself.
Use a WebView and you can use WebGL, that solves your problem. ;-)
Maybe why they aren't bothered to much, as far I know all WelGL implementation on Windows use DirectX.
Well, it depends.
This is the text of a slide of the presentation linked below:
"A Unified Approach?
Should we compile entire VMs from C/C++ to JavaScript, and implement JavaScript-emitting JITs?
Seems the only way to run most languages with perfect semantics + maximum speed
This is why I believe C/C++ to JavaScript translation is the core issue regarding compilation to JavaScript"
http://www.ustream.tv/recorded/29324270
http://kripken.github.io/mloc_emscripten_talk/
On the issue of threading:
That is what webworkers is for.
I know it wouldn't cover all the uses of threading, but it fit a lot of use cases.