The can also be good and bad at the same company. No company gets to the size of MS etc. without someone doing something evil. Hell even Linux has had at least one murderer work on it; statistically companies that size must have all sorts.
With the possible exception of Oracle no tech. company sets out to be evil.
Microsoft's bad name has mainly come from their cut-throat approach to competitors (of which OSS is one) and the even worse treatment of their supposed partners. However they have historically treated developers well, and respected privacy.
Despite their motto and their many-many good deeds, this is far from the first time Google have deliberately violated people's privacy (Streetview springs to mind), and even when court have never given wholehearted apologies.
As with all things you have to choose your poison, then take responsibility for limiting your own risks.
It is already trivially easy for any App that wants to support WINE to work. Trying to provide an acceptable UX running software whose developers don't care isn't a route to mainstream acceptance.
GDDR3 is an optimized type of DDR2 memory. The 700Mhz in the Xbox360 it is less than half the speed of the DDR3-800Mhz stuff used by Atom's for the last couple of years. Even if you fit them with 4 times the memory they can't get close to the 360's graphics performance? In the context of Haswell you are talking about an entry level of dual-channel DDR3-1600Mhz or around 25Gb/s beating the GDDR5 in bleeding edge top of the line discreet cards from just 4 years ago.
So research combining at least three cutting edge AI problems (each of which have had millions of R&D spent on) in to a new application isn't worth doing? Computer vision and OCR may already have many practical applications but it is far from a 'solved' problem. The existence of Watson and Alpha (both of which require millions of pounds of hardware and hundreds of human PhDs to solve very narrow domain problems) hardly suggests that the is no need for further AI research.
1024 is much faster: ~x5 less operations. The is also an issue with larger keys (especially with chains) not fitting in the first packet.
More specifically Google still uses 1024 keys (despite the fact you and me can't even get them any more), so not supporting them may have been bigger news.
Even the IP and size of the file may be enough to prosecute some things (say a given image was deemed illegal, but happens to be the only image roughly* a given size on Wikipedia at the time you accessed it). A much bigger problem is that if the page (or worse search results) that linked to the forbidden page are not using SSL, the fact that your next hop is to the SSL server is itself fairly incriminating. (A lot of the UK's thought crime stuff is about 'anti-terrorism' stuff were they don't need to prove anything). If every site as you suggest was using SSL this would prevent the second case, but with sites smaller then Wikipedia the first point gets dramatically worse.
Frankly I SERIOUSLY doubt the UK are the first with this sort of tracking (just the first dumb enough to tell everyone). Anyone not already using at least SSL obviously doesn't really care.
*Actually the size of the file is really hard on SSL with pipelining etc.
If MS still have the monopoly then pushing it would result in sanctions. If they don't have the monopoly then the will be just as many Chrome/Ubunto/Tizen manufacturers able to use it as a 'destroy Windows' button (not that that is really any better).
Your description is not so much machine learning as basic math. If you just want each scoring system to have equal weight on the results then compensating for the variation is trivial.
Where machine learning would come in is to find underlying pattens in the data.
This could be used to weed out reviewers to lazily copy scores, or are subject to influence.
It would also allow them to test your scores for some games against the population of reviewers to find reviewers with similar tastes.
You could also use clustering algorithms to find niche games which got a few really strong scores but whose average was really pulled down because they don't have wide appeal.
To start with the gamer/non-gamer distinction is something they want to 'fix'. We don't have filmer/non-filmer, or much tv/non-tv anymore why shouldn't Valve target everyone?
Secondly Microsoft is a very significant competitor. Both as a game studio but much more importantly with XBox/Windows Live. You can dismiss it as emotional if you want, but few successful companies have left their entire fate in the hands of a competitor.
Thirdly they they need to look to the future. They are already far behind on phones and tablets, and probably think that sub-$100 media centers/consoles are reaching a tipping point. It seems most likely that the majority of these will continue to run Linux, and proprietary ones all lock Valve out of the market anyway.
Valve literally can't afford to do anything that might disrupt the TF2 and DOTA2 communities. If a bug in the Linux client caused x-number of games with Windows users to drop they would have a major problem. Not to mention the anti-cheet being by far their most important feature and is genuinely much harder on Linux. By using an old title with limited potential they are simply managing risk. If they can get it to work then TF2 and DOTA2 will be their highest priorities.
The digital hats you will be buying are on their servers not your console. It should be possible to make this 100% secure. Banks don't mind you running Linux on a toaster to connect to your account and they have a lot more to lose. Saying that they are still 100-1000 fold away of being able to get a Tegra3 console out the door. They should have proved the market by sticking a badge on a cheep Chinese ARM.
Whilst it's openness may not be perfect it would love this approach to get a foot in the door of the otherwise draconian console market. Realistically for this project to get anywhere near what people are hoping, it will be it needs to be a Dreamcast level success. To match the Dreamcast this Kickstarter would really need 300,000 backers (it is just reaching 30,000 so who knows).
If you are using the laptop for work then the company pays the extra. If you are using a work laptop that is so locked down you can't run a 'portable' FF or chrome for personal web browsing, you probably shouldn't be.
Easiest solution is probably to use tar's incremental backups. The -g argument creates a relatively small file listing the files already backed up, so future incremental runs can skip them. If you keep the incremental files on the laptop then you can put each days actual tar backup on whatever devices you have handy.
x86 machine code is a publicly standardised, widely know and extensively studied language. By modern standards is actually quite simple. Forget about minified JavaScript. I have seen 'documented' scripts that were harder to understand than machine code. Proof in point JavaScript can execute x86 code anyway: http://bellard.org/jslinux/.
Fine, so if you EVER want to print anything which mentions a religion, politics or any of the millions of other things people have been persecuted, jailed and killed for over the years you will puchase a competitors printer? You know like history homework, or a birthday card to anyone who holds any option on anything.
Of couse they can trace it back to you (or why would they do it), and to make it easier all they need to do is compele everyone to registerer an example page (the is lots of precedent for this with typewriters).
Sure the lots of other important things to worry about, but that dosn't mean it should be let slide. Perticularly when the alternative (don't do it) is even easier and cheeper.
The are LOTS of flaws in your agument. Prehaps the easiest to explain is what happens if the is a revolution in your country and previous 'free-expression' suddenly lands you in jail?
Slashdot Mirror
Jazelle is a bit of a red hearing. Even on the Pi's limited CPU a JIT VM will nearly always be faster.
The can also be good and bad at the same company. No company gets to the size of MS etc. without someone doing something evil. Hell even Linux has had at least one murderer work on it; statistically companies that size must have all sorts.
With the possible exception of Oracle no tech. company sets out to be evil.
Microsoft's bad name has mainly come from their cut-throat approach to competitors (of which OSS is one) and the even worse treatment of their supposed partners. However they have historically treated developers well, and respected privacy.
Despite their motto and their many-many good deeds, this is far from the first time Google have deliberately violated people's privacy (Streetview springs to mind), and even when court have never given wholehearted apologies.
As with all things you have to choose your poison, then take responsibility for limiting your own risks.
It is already trivially easy for any App that wants to support WINE to work. Trying to provide an acceptable UX running software whose developers don't care isn't a route to mainstream acceptance.
A Core 2 era dual-core Xeon would out-muscle them unless they can use the propitiatory GPU. With the GPUs they could spank any single chip.
Not supporting the type of moble App Facebook were trying to write is a feature of HTML5 not a wrinkle to iron out.
GDDR3 is an optimized type of DDR2 memory. The 700Mhz in the Xbox360 it is less than half the speed of the DDR3-800Mhz stuff used by Atom's for the last couple of years. Even if you fit them with 4 times the memory they can't get close to the 360's graphics performance?
In the context of Haswell you are talking about an entry level of dual-channel DDR3-1600Mhz or around 25Gb/s beating the GDDR5 in bleeding edge top of the line discreet cards from just 4 years ago.
So research combining at least three cutting edge AI problems (each of which have had millions of R&D spent on) in to a new application isn't worth doing?
Computer vision and OCR may already have many practical applications but it is far from a 'solved' problem. The existence of Watson and Alpha (both of which require millions of pounds of hardware and hundreds of human PhDs to solve very narrow domain problems) hardly suggests that the is no need for further AI research.
1024 is much faster: ~x5 less operations. The is also an issue with larger keys (especially with chains) not fitting in the first packet.
More specifically Google still uses 1024 keys (despite the fact you and me can't even get them any more), so not supporting them may have been bigger news.
Even the IP and size of the file may be enough to prosecute some things (say a given image was deemed illegal, but happens to be the only image roughly* a given size on Wikipedia at the time you accessed it). A much bigger problem is that if the page (or worse search results) that linked to the forbidden page are not using SSL, the fact that your next hop is to the SSL server is itself fairly incriminating. (A lot of the UK's thought crime stuff is about 'anti-terrorism' stuff were they don't need to prove anything).
If every site as you suggest was using SSL this would prevent the second case, but with sites smaller then Wikipedia the first point gets dramatically worse.
Frankly I SERIOUSLY doubt the UK are the first with this sort of tracking (just the first dumb enough to tell everyone). Anyone not already using at least SSL obviously doesn't really care.
*Actually the size of the file is really hard on SSL with pipelining etc.
It is fairly trivial for white-box clusters to reach five nines. When was Google's last downtime?
If MS still have the monopoly then pushing it would result in sanctions. If they don't have the monopoly then the will be just as many Chrome/Ubunto/Tizen manufacturers able to use it as a 'destroy Windows' button (not that that is really any better).
Your description is not so much machine learning as basic math. If you just want each scoring system to have equal weight on the results then compensating for the variation is trivial.
Where machine learning would come in is to find underlying pattens in the data.
This could be used to weed out reviewers to lazily copy scores, or are subject to influence.
It would also allow them to test your scores for some games against the population of reviewers to find reviewers with similar tastes.
You could also use clustering algorithms to find niche games which got a few really strong scores but whose average was really pulled down because they don't have wide appeal.
To start with the gamer/non-gamer distinction is something they want to 'fix'. We don't have filmer/non-filmer, or much tv/non-tv anymore why shouldn't Valve target everyone?
Secondly Microsoft is a very significant competitor. Both as a game studio but much more importantly with XBox/Windows Live. You can dismiss it as emotional if you want, but few successful companies have left their entire fate in the hands of a competitor.
Thirdly they they need to look to the future. They are already far behind on phones and tablets, and probably think that sub-$100 media centers/consoles are reaching a tipping point. It seems most likely that the majority of these will continue to run Linux, and proprietary ones all lock Valve out of the market anyway.
Valve literally can't afford to do anything that might disrupt the TF2 and DOTA2 communities. If a bug in the Linux client caused x-number of games with Windows users to drop they would have a major problem. Not to mention the anti-cheet being by far their most important feature and is genuinely much harder on Linux.
By using an old title with limited potential they are simply managing risk. If they can get it to work then TF2 and DOTA2 will be their highest priorities.
The digital hats you will be buying are on their servers not your console. It should be possible to make this 100% secure. Banks don't mind you running Linux on a toaster to connect to your account and they have a lot more to lose.
Saying that they are still 100-1000 fold away of being able to get a Tegra3 console out the door. They should have proved the market by sticking a badge on a cheep Chinese ARM.
https://play.google.com/store/devices/details?id=nexus_7_8gb
Whilst it's openness may not be perfect it would love this approach to get a foot in the door of the otherwise draconian console market. Realistically for this project to get anywhere near what people are hoping, it will be it needs to be a Dreamcast level success. To match the Dreamcast this Kickstarter would really need 300,000 backers (it is just reaching 30,000 so who knows).
Then why not call it 93.5 square inches, not only would that be higher, it would also be more useful across different aspect ratios.
So Developers and System Admins know WFT the user is talking about when they take support calls the first day it is in the shops?
Assuming the iPad fonts were chosen to be the optimal size, 20% smaller is likely going to be an issue.
If you are using the laptop for work then the company pays the extra.
If you are using a work laptop that is so locked down you can't run a 'portable' FF or chrome for personal web browsing, you probably shouldn't be.
Easiest solution is probably to use tar's incremental backups. The -g argument creates a relatively small file listing the files already backed up, so future incremental runs can skip them. If you keep the incremental files on the laptop then you can put each days actual tar backup on whatever devices you have handy.
x86 machine code is a publicly standardised, widely know and extensively studied language. By modern standards is actually quite simple.
Forget about minified JavaScript. I have seen 'documented' scripts that were harder to understand than machine code.
Proof in point JavaScript can execute x86 code anyway: http://bellard.org/jslinux/.
Fine, so if you EVER want to print anything which mentions a religion, politics or any of the millions of other things people have been persecuted, jailed and killed for over the years you will puchase a competitors printer? You know like history homework, or a birthday card to anyone who holds any option on anything.
Of couse they can trace it back to you (or why would they do it), and to make it easier all they need to do is compele everyone to registerer an example page (the is lots of precedent for this with typewriters).
Sure the lots of other important things to worry about, but that dosn't mean it should be let slide. Perticularly when the alternative (don't do it) is even easier and cheeper.
The are LOTS of flaws in your agument. Prehaps the easiest to explain is what happens if the is a revolution in your country and previous 'free-expression' suddenly lands you in jail?