But what happens when you run into people like me on slashdot talking about <nonsense> how Obama is great even though he's an African socialist Muslim because he's killed 6x more people with drones than Bush?
As Deng Xiao Ping said "Who cares if the cat is black or white so long as it catches mice?". Obama may be black but goddamn if that boy ain't a good mouser. Maybe we need a Muslim President because he knows how them Muslims think because he's one of 'em and that's what makes him a better killer.
It's like Deckard in Blade Runner. "He does a man's job". Maybe we shouldn't worry if he's a man or not.
When your younger, you don't mind the beater car that you have to repair all the time. You dream of the day when its perfectly restored, but you never get there. One day you just realize, you have other things you want to do, so you buy a new car that just works. If you're lucky you can now afford one because you stopped f'ing with linux and started f'ing with something else that you can make a good living at. And if you're really lucky, you pick up some pile of junk to work on solely as a hobby and without the stress of wondering if the f'ing thing is going to get you to work on time.
Now, I'd rather just pop in a DVD or download a binary blob and drag it to/Applications. My family time is limited and I'd rather be spending it with them.
That's what the Germans probably said. "I don't care about elections and a free press and all that stuff. I'd rather just have the NAZIS sort things out so I can spend time with my Kinder, Kuche, Kirche"
Now you may say the comparison between Apple and the NAZIS is a bit hyperbolic. But is it really? Both Apple fans and the SS wore mostly black clothes and are almost entirely Caucasian. Sure there are some Asians in there, but then NAZIS were quite keen on the Japanese.
The more you think of it, the more you realise that buying an Apple produced device is exactly the same as voting for the NAZIS.
Still it's better than using Linux or bloody Windows 8.
x86 has done great things for them, with both patents and it's general insane difficulty to implement creating huge barriers to entry for others helping them maintain their monopoly.
AMD invented x64 and have licensed it to Via and Transmeta and have apparently said they will license it to anyone. The bits of x86 you need to implement user mode x64 (aka "run third party applications") are getting a bit old now it seems like in the long run x64 will be an architecture that anyone can license from AMD or Intel (who get access to any AMD patents for free) for a fee.
The hard bit is SSE. x64 originally includes SSE which was released in 2000. In fact an x64 chip must include SSE because SSE is baked into the ABI.
Still 17 years from 2000 that means that patents expire in 2017 - i.e. about 4 years from now.
So it seems like we're only four years away from NVidia being able to pay a license fee to AMD and build an x64 chip. In fact if they pay a fee to Intel for SSE2 they could probably build one now.
I.e. in the long run x64 will be a licensable architecture just like ARM is now. The older bits of x86 are already old enough to be out of patent.
What about the kernel mode stuff? There I think it is not as critical. For a fee Microsoft will write a hardware abstraction layer for pretty much any processor. But you need to implement all the registers and instructions in user mode so you can run third party applications. It's not even good enough to trap some as invalid instructions and execute them in software if you look at Mips v Lexra.
I suspect that part of the reason Intel invented Itanium is that a totally new and very weird architecture means a new set of patents for the core stuff that you need to make a processor which can run user mode code written to the standard ABI.
I think you could argue that Risc was a very good fit for the technological limitations of the 1990's - i.e. single issue, single cycle machines with a short pipeline. Also memory latency was low.
It's not at all clear that it is much of a win when you have a wide super scalar machine with a humongous pipeline with loads of memory latency.
That was the inspiration behind EPIC and Itanium. Intel wanted to out-Risc the Risc chips by doing all the wide multi issue scheduling in software instead of hardware. They had predication to try to hide memory latency.
Now as is usual for radical Intel designs it was all let down by compilers.
Still look at the state of the art - Intel and AMD have some excellent server grade 64 chips. No one has done anything like that for ARM. We don't even know what the performance of AArch64 will be.
x86 has some benefits if you care about single thread performance. The code density is surprisingly good so it is cache efficient (though Arm now has Thumb2 which has good code density too - but not for 64 bit mode). You can transform it into UOPS which actually get executed, and each microarchitecture can define its own UOPS scheme. This all takes a lot of silicon to do quickly of course, but on a server grade chip the die is mostly cache. So the look up table that does the transformation is small compared to that.
Modern x86 or particularly x64 manages to be reasonably dense whilst having a decent number of registers and being 64 bit.
And I suspect being able to do a transformation from instructions to UOPS and also having free range to design the micro architecture that executes the UOPS - what Hennessy and Patterson called a 'decoupled architecture' - allows for a lot of flexibility that ARM implementers do not have. Perhaps you can get all the parallelism out of typical programs with this scheme and so EPIC style explicit parallelism isn't really needed.
In any case it's noticeable that since Intel and AMD started to build decoupled x86 and x64 chips those chips have been leaders in single threaded performance. And most of the time the fastest ARM implementation has barely outperformed the slowest x86 one - the Intel Atom.
ARM is nice and all - Qualcomm have some excellent lower power chips for your cellphone. I just don't see it being competitive in single threaded performance with x64 in servers.
Now don't get me wrong - that doesn't mean people aren't going to use ARM in servers. There are probably a lot of places where you don't need an x86 or x64. But for the heaving lifting stuff I think x64 is here to stay.
I worked on an system that was a more or less standard x86 PC with a bunch of PCI cards with i960's running vxWorks and FPGAs that did the critical stuff. It was set up so the PC could crash and reboot while the system kept doing its mission critical stuff because all the intelligence was in the PCI cards.
It was a very interesting project if you liked hacking around vxWorks and Windows and the i960 was used because it came in IO accelerator chips with PCI-PCI bridges. We set things up so the PC saw a memory window it used to communicate with the system but everything downstream of that was hidden. The reason for that is that Windows will reconfigure everything on the PCI bus when it boots. Still if you set things up just right everything was hidden from Windows and could go on running.
Incidentally this is very similar to what happens on a mobile phone. You have an application processor running Android or iOS. Then you have an modem which runs a realtime OS. They typically have some shared memory to communicate but the network stack is isolated from all the crap and chaos on the application side.
The reason for that is - as someone put it - you don't want a bug report saying "When I play this specific Britney Spears song I lose connectivity".
So you can have a real time domain where people write in C with very strict rules on disabling interrupts and not fragmenting the heap and a craplication domain where anything goes, even Java.
Let me tell you a story. A bunch of Swedish guys stay in a hotel in the US. Their manager speaks Spanish and chats to the staff. The staff complain the Swedes don't tip. So the manager talks to them and explains they should all put a dollar bill on the table each day. Some of them leave change and the cleaners tell the manager this is unacceptable. Eventually all but one of them do the crisp $1 per day thing. The one that doesn't claims that tipping is feudal and turns the cleaners into supplicants, the hotel should pay the staff a decent wage like in Sweden, the US should have a social democratic party like in Sweden to stick up for the workers and so on and so on and refuses to do it.
When he checks out he finds out the cleaners have put on the porn channel every day after he left the room and turned it off just before he got back.
I think we can all learn a lesson from that story, can't we?
Look at it this way. 600 staff means that Canonical needs to raise 600 salaries. Unless Shuttleworth has enough cash to pay them indefinitely he needs to find investors. Now investors want a return on their investment, i.e. profits.
To be fair to him the editor he was using was probably really hard to use. So he's lucky he didn't end up with an script file that was all fubar with ^H^H^H characters all over it.
Where's the revenue stream from their core clientèle? It seems like unless they can get it used on tablets by companies who pay a licence fee they don't have one. Or they could sell support. But slashdotters aren't going to pay for that. They need to get it onto corporate servers as a sort of software as a service
* Support services (mostly to business) alongside which they sell Landscape * Contracting services to businesses (for instance working with OEMs such as Dell, or helping Google with Chrome OS). As Ubuntu makes its way onto mobile phones and TVs then this will grow. * Ubuntu One (online file storage and syncronisation service) * Ubuntu One Music Store (selling music from within Ubuntu) * Ubuntu Software Centre's paid section (Canonical takes a cut of purchases) * The Canonical Store (selling physical Ubuntu branded items) * Closed-source projects wishing to use Launchpad.net can purchase a license
I think you need to be careful talking about 'core clientèle' when the clientèle is not the one that will make the company turn a profit.
Now admittedly I'm not sure how this development will get them any closer to making a profit. If they're going for tablets they are screwed because Google/Apple and even Microsoft have a competitive advantage their. And if they're going for servers who really cares about UIs when everyone is sshing in anyway.
Maybe ATI is paying them to torment NVidia? ATI have bad open source drivers and NVidia have good closed source ones.
open-source Linux graphics drivers, and they're pressuring hardware vendors with commercial closed-source drivers to support it too.
I think a lot of open source is actually a very elaborate practical joke at NVidia's expense possibly coordinated by an evil genius with an irrational fear of NVidia like Charlie Demerjian. For example Linus himself flipped them off over Optimus support.
Other people pointed out that they should use DMA-BUF to do to the copies between the two GPUs in an Optimus system. So they assigned an intern to do that
But that intern was told that they couldn't use it unless they GPL'd their drivers.
Hahaha. I bet his boss went apeshit when he saw that email.
I used to joke that the reason Microsoft had ISV and IHVs supporting them is because they didn't do stuff like this. However with Windows Phone they have adopted Linux style trolling of their ISVs. Windows Phone 7 didn't run C/C++ Win32 applications so everyone had to rewrite in C#. Then on Windows Phone 8 they went back to C++ but a new API, WinRT. C# applications still worked but they wouldn't get access to new APIs.
What happened? Well Skype for example announced they wouldn't support anything after Windows Mobile. So MS bought them. Actually that will save them rewriting too because MS are allowed to run Win32 application on Windows Phone.
They told all the browser makers - Opera, Mozilla etc that even if they rewrote they wouldn't allow the browsers in the store. So they all dropped support for any Microsoft mobile OSs. All the other Windows Mobile ISVs seem to have jumped ship to Android.
What about the IHVs? Well they got Samsung and HTC onboard initially. They told HTC they couldn't use Sense or any of their old Windows Mobile software. Then they promoted Nokia as the saviour of the platform and grumbled that the pre-Nokia IHVs hadn't done a good enough job.
Then again since Windows Phone 7 MS has had a Linux like share of the mobile OS market i.e. 1-2%. By contrast Windows Mobile had 10%+
So it does seem that being a complete dick to your IHVs and ISVs leads to poor support.
So what does it all mean.
I think HTC and Samsung should drop Windows Phone and let Microsoft buy up Nokia as their in house manufacturer. Putting a cellphone into production that doesn't sell is very expensive. They should only support the platform if Microsoft offers act as a buyer of last resort - i.e. if they have n unsold handsets MS should pay n times the street price.
I think NVidia should leave their Linux drivers to interns instead of dropping it for political reasons however. One intern-year per year of effort is basically nothing and it's probably easy to find interns who want to work with Linux.
The United States Department of Energy estimates that if algae fuel replaced all the petroleum fuel in the United States, it would require 15,000 square miles (39,000 km2) which is only 0.42% of the U.S. map,[10] or about half of the land area of Maine. This is less than 1â7 the area of corn harvested in the United States in 2000.[11] However, these claims remain unrealized commercially. According to the head of the Algal Biomass Organization algae fuel can reach price parity with oil in 2018 if granted production tax credits.
It's just that there are currently cheaper sources than that.
Also farmers prefer to grow corn because it is heavily subsidised.
Corn is the top crop for subsidy payments. The Energy Policy Act of 2005 mandates that billions of gallons of ethanol be blended into vehicle fuel each year, guaranteeing demand, but US corn ethanol subsidies are between $5.5 billion and $7.3 billion per year. Producers also benefitted from a federal subsidy of 51 cents per gallon, additional state subsidies, and federal crop subsidies that can bring the total to 85 cents per gallon or more. However, the federal ethanol subsidy expired December 31, 2011.[16] (US corn-ethanol producers were shielded from competition from cheaper Brazilian sugarcane-ethanol by a 54-cent-per-gallon tariff, however that tariff also expired December 31, 2011
Of course corn has the benefit that you can grow it on fields. Algae biodiesel would require covering 15,000 square miles of the US with huge tanks of water. Which, US agriculture being what it is, would need to be paid for by the US tax payer to keep the farmers happy.
What needs to happen is for subsidies to be phased out and then let the market decide what is worth growing. But that will never happen - it would be political suicide to even suggest it.
I've got a relative on a Non Invasive Ventilator and the doctors that tweak it were very averse to giving him pure O2 because your brain monitors CO2 levels and uses it to control breathing rate. So if you breath an atmosphere with no CO2 at all your breathing reflex is suppressed.
It's actually funny how much of hack evolved organisms are - put them in a novel situation like pure O2 and they will crash and burn because they monitor CO2 levels instead of O2 ones - presumably because sensing CO2 levels in easier to do because you can look at blood pH or something.
But then again I suppose it's unreasonable to expect evolved organisms to not crash and burn when put in a situation that never occurs in nature. Probably if you put a lot of organisms into pure O2 a few of them would survive because of some other strange metabolic misfeature (it's not like there is a shortage of CO2 when you're burning sugar and O2 to make it constantly, so you could just use that to keep the sensors in your brain stem happy) and their offspring would end up dominating things.
B.S. in Forestry Management from the University of Idaho
I like the way they call it a B.S. It's like a good con man dares his marks to object, secure in the knowledge he can convince them they're terrible people when they do.
Slashdot Mirror
But what happens when you run into people like me on slashdot talking about <nonsense> how Obama is great even though he's an African socialist Muslim because he's killed 6x more people with drones than Bush?
As Deng Xiao Ping said "Who cares if the cat is black or white so long as it catches mice?". Obama may be black but goddamn if that boy ain't a good mouser. Maybe we need a Muslim President because he knows how them Muslims think because he's one of 'em and that's what makes him a better killer.
It's like Deckard in Blade Runner. "He does a man's job". Maybe we shouldn't worry if he's a man or not.
Yeah it's no problem. Once you buy and "Extended Lifetime Pass" online from Seagate (pricing TBA) the drive will be unlocked for "HD only mode".
Its like cars (or motorcycles)...
When your younger, you don't mind the beater car that you have to repair all the time. You dream of the day when its perfectly restored, but you never get there. One day you just realize, you have other things you want to do, so you buy a new car that just works. If you're lucky you can now afford one because you stopped f'ing with linux and started f'ing with something else that you can make a good living at. And if you're really lucky, you pick up some pile of junk to work on solely as a hobby and without the stress of wondering if the f'ing thing is going to get you to work on time.
That explains the popularity of Gentoo
Now, I'd rather just pop in a DVD or download a binary blob and drag it to /Applications. My family time is limited and I'd rather be spending it with them.
That's what the Germans probably said. "I don't care about elections and a free press and all that stuff. I'd rather just have the NAZIS sort things out so I can spend time with my Kinder, Kuche, Kirche"
Now you may say the comparison between Apple and the NAZIS is a bit hyperbolic. But is it really? Both Apple fans and the SS wore mostly black clothes and are almost entirely Caucasian. Sure there are some Asians in there, but then NAZIS were quite keen on the Japanese.
The more you think of it, the more you realise that buying an Apple produced device is exactly the same as voting for the NAZIS.
Still it's better than using Linux or bloody Windows 8.
objdump -d exefile
Is far more complete than
man exefile
once you get past the learning curve.
I'm pretty sure the USA constitution has words to the effect that all people are equal.
It says they are convertible by it doesn't specify a 1:1 exchange rate. A member of group A might be only worth 3/5 of a member of group B.
x86 has done great things for them, with both patents and it's general insane difficulty to implement creating huge barriers to entry for others helping them maintain their monopoly.
AMD invented x64 and have licensed it to Via and Transmeta and have apparently said they will license it to anyone. The bits of x86 you need to implement user mode x64 (aka "run third party applications") are getting a bit old now it seems like in the long run x64 will be an architecture that anyone can license from AMD or Intel (who get access to any AMD patents for free) for a fee.
The hard bit is SSE. x64 originally includes SSE which was released in 2000. In fact an x64 chip must include SSE because SSE is baked into the ABI.
Still 17 years from 2000 that means that patents expire in 2017 - i.e. about 4 years from now.
So it seems like we're only four years away from NVidia being able to pay a license fee to AMD and build an x64 chip. In fact if they pay a fee to Intel for SSE2 they could probably build one now.
I.e. in the long run x64 will be a licensable architecture just like ARM is now. The older bits of x86 are already old enough to be out of patent.
What about the kernel mode stuff? There I think it is not as critical. For a fee Microsoft will write a hardware abstraction layer for pretty much any processor. But you need to implement all the registers and instructions in user mode so you can run third party applications. It's not even good enough to trap some as invalid instructions and execute them in software if you look at Mips v Lexra.
I suspect that part of the reason Intel invented Itanium is that a totally new and very weird architecture means a new set of patents for the core stuff that you need to make a processor which can run user mode code written to the standard ABI.
I think you could argue that Risc was a very good fit for the technological limitations of the 1990's - i.e. single issue, single cycle machines with a short pipeline. Also memory latency was low.
It's not at all clear that it is much of a win when you have a wide super scalar machine with a humongous pipeline with loads of memory latency.
That was the inspiration behind EPIC and Itanium. Intel wanted to out-Risc the Risc chips by doing all the wide multi issue scheduling in software instead of hardware. They had predication to try to hide memory latency.
Now as is usual for radical Intel designs it was all let down by compilers.
Still look at the state of the art - Intel and AMD have some excellent server grade 64 chips. No one has done anything like that for ARM. We don't even know what the performance of AArch64 will be.
x86 has some benefits if you care about single thread performance. The code density is surprisingly good so it is cache efficient (though Arm now has Thumb2 which has good code density too - but not for 64 bit mode). You can transform it into UOPS which actually get executed, and each microarchitecture can define its own UOPS scheme. This all takes a lot of silicon to do quickly of course, but on a server grade chip the die is mostly cache. So the look up table that does the transformation is small compared to that.
Modern x86 or particularly x64 manages to be reasonably dense whilst having a decent number of registers and being 64 bit.
And I suspect being able to do a transformation from instructions to UOPS and also having free range to design the micro architecture that executes the UOPS - what Hennessy and Patterson called a 'decoupled architecture' - allows for a lot of flexibility that ARM implementers do not have. Perhaps you can get all the parallelism out of typical programs with this scheme and so EPIC style explicit parallelism isn't really needed.
In any case it's noticeable that since Intel and AMD started to build decoupled x86 and x64 chips those chips have been leaders in single threaded performance. And most of the time the fastest ARM implementation has barely outperformed the slowest x86 one - the Intel Atom.
ARM is nice and all - Qualcomm have some excellent lower power chips for your cellphone. I just don't see it being competitive in single threaded performance with x64 in servers.
Now don't get me wrong - that doesn't mean people aren't going to use ARM in servers. There are probably a lot of places where you don't need an x86 or x64. But for the heaving lifting stuff I think x64 is here to stay.
I worked on an system that was a more or less standard x86 PC with a bunch of PCI cards with i960's running vxWorks and FPGAs that did the critical stuff. It was set up so the PC could crash and reboot while the system kept doing its mission critical stuff because all the intelligence was in the PCI cards.
It was a very interesting project if you liked hacking around vxWorks and Windows and the i960 was used because it came in IO accelerator chips with PCI-PCI bridges. We set things up so the PC saw a memory window it used to communicate with the system but everything downstream of that was hidden. The reason for that is that Windows will reconfigure everything on the PCI bus when it boots. Still if you set things up just right everything was hidden from Windows and could go on running.
Incidentally this is very similar to what happens on a mobile phone. You have an application processor running Android or iOS. Then you have an modem which runs a realtime OS. They typically have some shared memory to communicate but the network stack is isolated from all the crap and chaos on the application side.
The reason for that is - as someone put it - you don't want a bug report saying "When I play this specific Britney Spears song I lose connectivity".
So you can have a real time domain where people write in C with very strict rules on disabling interrupts and not fragmenting the heap and a craplication domain where anything goes, even Java.
Let me tell you a story. A bunch of Swedish guys stay in a hotel in the US. Their manager speaks Spanish and chats to the staff. The staff complain the Swedes don't tip. So the manager talks to them and explains they should all put a dollar bill on the table each day. Some of them leave change and the cleaners tell the manager this is unacceptable. Eventually all but one of them do the crisp $1 per day thing. The one that doesn't claims that tipping is feudal and turns the cleaners into supplicants, the hotel should pay the staff a decent wage like in Sweden, the US should have a social democratic party like in Sweden to stick up for the workers and so on and so on and refuses to do it.
When he checks out he finds out the cleaners have put on the porn channel every day after he left the room and turned it off just before he got back.
I think we can all learn a lesson from that story, can't we?
Look at it this way. 600 staff means that Canonical needs to raise 600 salaries. Unless Shuttleworth has enough cash to pay them indefinitely he needs to find investors. Now investors want a return on their investment, i.e. profits.
I was going to argue but actually this is exactly the right way to respond to people misquoting Nietzsche.
Have you read any Ayn Rand by the way?
Sorry, just kidding.
That which doesn't kill you makes you evolve a bit faster.
Canonical employs 600 people. Those people need to be paid. And the people paying them need to get a return on their investment.
To be fair to him the editor he was using was probably really hard to use. So he's lucky he didn't end up with an script file that was all fubar with ^H^H^H characters all over it.
Where's the revenue stream from their core clientèle? It seems like unless they can get it used on tablets by companies who pay a licence fee they don't have one. Or they could sell support. But slashdotters aren't going to pay for that. They need to get it onto corporate servers as a sort of software as a service
http://askubuntu.com/questions/21730/how-does-ubuntu-make-money
* Support services (mostly to business) alongside which they sell Landscape
* Contracting services to businesses (for instance working with OEMs such as Dell, or helping Google with Chrome OS). As Ubuntu makes its way onto mobile phones and TVs then this will grow.
* Ubuntu One (online file storage and syncronisation service)
* Ubuntu One Music Store (selling music from within Ubuntu)
* Ubuntu Software Centre's paid section (Canonical takes a cut of purchases)
* The Canonical Store (selling physical Ubuntu branded items)
* Closed-source projects wishing to use Launchpad.net can purchase a license
I think you need to be careful talking about 'core clientèle' when the clientèle is not the one that will make the company turn a profit.
Now admittedly I'm not sure how this development will get them any closer to making a profit. If they're going for tablets they are screwed because Google/Apple and even Microsoft have a competitive advantage their. And if they're going for servers who really cares about UIs when everyone is sshing in anyway.
It doesn't pray though, Cats are sceptical of deities. Well unless those deities are them, obviously. But you don't pray to yourself.
Well unless you're Larry Ellison.
Maybe ATI is paying them to torment NVidia? ATI have bad open source drivers and NVidia have good closed source ones.
open-source Linux graphics drivers, and they're pressuring hardware vendors with commercial closed-source drivers to support it too.
I think a lot of open source is actually a very elaborate practical joke at NVidia's expense possibly coordinated by an evil genius with an irrational fear of NVidia like Charlie Demerjian. For example Linus himself flipped them off over Optimus support.
http://www.engadget.com/2012/06/18/linus-torvalds-nvidia-linux/
Other people pointed out that they should use DMA-BUF to do to the copies between the two GPUs in an Optimus system. So they assigned an intern to do that
But that intern was told that they couldn't use it unless they GPL'd their drivers.
http://linux.slashdot.org/story/12/10/11/1918251/alan-cox-to-nvidia-you-cant-use-dma-buf
Hahaha. I bet his boss went apeshit when he saw that email.
I used to joke that the reason Microsoft had ISV and IHVs supporting them is because they didn't do stuff like this. However with Windows Phone they have adopted Linux style trolling of their ISVs. Windows Phone 7 didn't run C/C++ Win32 applications so everyone had to rewrite in C#. Then on Windows Phone 8 they went back to C++ but a new API, WinRT. C# applications still worked but they wouldn't get access to new APIs.
What happened? Well Skype for example announced they wouldn't support anything after Windows Mobile. So MS bought them. Actually that will save them rewriting too because MS are allowed to run Win32 application on Windows Phone.
They told all the browser makers - Opera, Mozilla etc that even if they rewrote they wouldn't allow the browsers in the store. So they all dropped support for any Microsoft mobile OSs. All the other Windows Mobile ISVs seem to have jumped ship to Android.
What about the IHVs? Well they got Samsung and HTC onboard initially. They told HTC they couldn't use Sense or any of their old Windows Mobile software. Then they promoted Nokia as the saviour of the platform and grumbled that the pre-Nokia IHVs hadn't done a good enough job.
Then again since Windows Phone 7 MS has had a Linux like share of the mobile OS market i.e. 1-2%. By contrast Windows Mobile had 10%+
So it does seem that being a complete dick to your IHVs and ISVs leads to poor support.
So what does it all mean.
I think HTC and Samsung should drop Windows Phone and let Microsoft buy up Nokia as their in house manufacturer. Putting a cellphone into production that doesn't sell is very expensive. They should only support the platform if Microsoft offers act as a buyer of last resort - i.e. if they have n unsold handsets MS should pay n times the street price.
I think NVidia should leave their Linux drivers to interns instead of dropping it for political reasons however. One intern-year per year of effort is basically nothing and it's probably easy to find interns who want to work with Linux.
s/do/alternet said they do/
Funrobot is MSI's somewhat Heath Robinson robotics brand. They've got high end ones with ultrasonic sensors
This is the M800
http://www.youtube.com/watch?v=RzV76Zjru5A
It works pretty well and can find the docking station to recharge. It is somewhat expensive (US$400 - above most people's impulse buy threshold)
The R500 has bump sensors
http://www.youtube.com/watch?v=6lFwhqcLnzo
It's cheap (about US$120) but also very irritating. Bump sensors make a lot of noise and it will also get stuck on cables, curtains and so on.
If you look on the net it is now very hard to buy an R500 but the M800 are [still being sold](http://ecshweb.pchome.com.tw/search/v2/?q=icleaner)
This makes me think that bump sensors are not a very good idea, even for vacuum cleaners.
They all fall ... FORE!!! ... the delicious mistakes.
Broke that moar for you.
So you want to use "CO2 emissions permits" to kill the people who say things you don't like?
That's just super.
You could make it from algae if you really needed to.
http://en.wikipedia.org/wiki/Algae_fuel
The United States Department of Energy estimates that if algae fuel replaced all the petroleum fuel in the United States, it would require 15,000 square miles (39,000 km2) which is only 0.42% of the U.S. map,[10] or about half of the land area of Maine. This is less than 1â7 the area of corn harvested in the United States in 2000.[11] However, these claims remain unrealized commercially. According to the head of the Algal Biomass Organization algae fuel can reach price parity with oil in 2018 if granted production tax credits.
It's just that there are currently cheaper sources than that.
Also farmers prefer to grow corn because it is heavily subsidised.
http://en.wikipedia.org/wiki/Agricultural_subsidy#United_States
Corn is the top crop for subsidy payments. The Energy Policy Act of 2005 mandates that billions of gallons of ethanol be blended into vehicle fuel each year, guaranteeing demand, but US corn ethanol subsidies are between $5.5 billion and $7.3 billion per year. Producers also benefitted from a federal subsidy of 51 cents per gallon, additional state subsidies, and federal crop subsidies that can bring the total to 85 cents per gallon or more. However, the federal ethanol subsidy expired December 31, 2011.[16] (US corn-ethanol producers were shielded from competition from cheaper Brazilian sugarcane-ethanol by a 54-cent-per-gallon tariff, however that tariff also expired December 31, 2011
Of course corn has the benefit that you can grow it on fields. Algae biodiesel would require covering 15,000 square miles of the US with huge tanks of water. Which, US agriculture being what it is, would need to be paid for by the US tax payer to keep the farmers happy.
What needs to happen is for subsidies to be phased out and then let the market decide what is worth growing. But that will never happen - it would be political suicide to even suggest it.
I've got a relative on a Non Invasive Ventilator and the doctors that tweak it were very averse to giving him pure O2 because your brain monitors CO2 levels and uses it to control breathing rate. So if you breath an atmosphere with no CO2 at all your breathing reflex is suppressed.
It's actually funny how much of hack evolved organisms are - put them in a novel situation like pure O2 and they will crash and burn because they monitor CO2 levels instead of O2 ones - presumably because sensing CO2 levels in easier to do because you can look at blood pH or something.
But then again I suppose it's unreasonable to expect evolved organisms to not crash and burn when put in a situation that never occurs in nature. Probably if you put a lot of organisms into pure O2 a few of them would survive because of some other strange metabolic misfeature (it's not like there is a shortage of CO2 when you're burning sugar and O2 to make it constantly, so you could just use that to keep the sensors in your brain stem happy) and their offspring would end up dominating things.
B.S. in Forestry Management from the University of Idaho
I like the way they call it a B.S. It's like a good con man dares his marks to object, secure in the knowledge he can convince them they're terrible people when they do.