They might not have to change the licensing at all, as it seems like the amount of data and number of transactions per unit time on a typical SQL server will probably continue to scale faster than the ability of CPU makers to throw more cores on a chip. Especially once this downturn ends and web app development continues full steam again and starts to hit those database servers harder than ever in addition to increased user base.
It will make shot noise worse but may be used to improve other types of noise and distortion. Unfortunately, camera makers have already gotten much better at accommodating other flaws(at least, for domestic imaging) save shot noise which is a fundamental physical issue.
It's not strictly pointless. Oversampling has some uses, namely as a way to filter out random noise and other imperfections caused by the sensor itself. If you can model the random noise, and have many samples for the same atomic datum, you can find ways to still improve the SNR of the data. It still won't let you zoom in to see more detail, however, but still not totally useless for all applications.
A half inch jack wouldn't make any sense in practice. At that point two smaller connectors could have a more compact footprint with less noise. Even an XLR would be a competitive option. And of course in this era, you could run digital coax/optical at 192khz with a smaller footprint and see even less noise with both channels[or more] and having the ADC as close to the source amplifer on-board, as possible.
But that's exactly the way changing OS architectures and APIs can help. Right now the default behavior is to start a worker thread of some type that blocks on IO requests and then reports back. Most apps in the wild don't even bloody do this and just have a few threads do everything and some even have the main app loop block on IO.(let's all pretend we don't see our app windows grey out several times a day!)
We've argued for decades this was a programmer issue but that sort of pedantic criticism has accomplished nothing. Arguing that if all programmers were as smart as you we'd have no problems has been a bad attitude from the OS folks.
Maybe its time we abandoned the current basic Win32/stdlib type APIs for ones that were more IO friendly so that the easiest and most lazy-friendly way to write code is also the best practice for parallelism, and not the other way around which has failed miserably.
After Brown was devastated when he found out his DBZ and Sailor Moon SVGA wallpapers site with 2482 hits was closed down by the whims of the private sector after a good 14 year run, he decided that only the public sector could be trusted with running such a critical service.
Parent is totally inaccurate. Windows supports OpenGL fine.
The issue is that Microsoft usually ships Windows with subpar distributions of graphics drivers for 3D support in general(not just OpenGL) compared to actual vendor drivers in terms of feature support beyond what you need to use basic Windows programs. Probably in some attempt to provide marginally more stable drivers historically under the assumption most users wouldn't need anything more than basic API support and everyone else would have their apps somehow manage to install/warn them to get a better driver package. This doesn't really work for WebGL, however.
Google could just have their installer get you to update your GPU driver, but overall this is probably a cleaner solution for long term support since you have about +/- 10 year support expectations for Direct3D on a GPU in a way that is far better tested than OpenGL for "last years model".
By investing in foreign markets, you share in their profits, making the "trade" balance a non-issue. The key problem with achieving this in practice is there are really no guarantees that your rights will be respected by the foreign parties when they realize they no longer need your continued investment.
Its also worth noting that big corporate research may in some cases accomplish great innovation but it usually does little for the companies. PARC and Bell Labs did a lot for the world economy to this very day, but did little to help their parent companies out. Even now, look at what researchers at IBM and Microsoft do with absolutely massive budgets in terms of innovation in their key competency areas vs. smaller companies with incomparable or even no dedicated research.
Virtual memory isn't even close to the same thing. Harvard Architecture processors literally execute from ROM. Their program counter points to ROM and they literally cannot grab instructions from elsewhere in the basic architecture. The reason this is important to note is that in theory basic Harvard programs require ZERO ram to work(and indeed, a Hello World program for most Harvard microcontrollers would require none). x86 processors fetch instructions from RAM and therefore require something memory mapped at least to fetch instructions from. So saying the PIC has less resources is misleading as it uses it's resources in a different way so it's apples to oranges.
Virtual memory is just a system feature that allows for on-the-fly access of data in ROM as if it were already in memory. It is still eventually paged to RAM, so it makes no difference in this case and isn't really comparable at all.
Less than could go wrong with being fully dependent on sea lanes for your economy, especially when one of your political rivals completely dominates world sea power.
But who can say what is best to have? Someone with a horrible genetic disease might be made immune to some future superbug by that condition, and therefore this "diseased" line would be more fit than any other. Likewise, they may require far less food energy to live than the average human, making them fit for some sort of future global famine an Olympic medalist with a Ph.D. in Physics might not make it through. Traits modern humans seem to value like excess intelligence may not be beneficial in survival scenarios beyond the already high intelligence of pre-history homo sapiens sapiens, but other unpredictable factors may be.
How exactly? Using exploits in non-deprecated wireless security is far more technically involved than running some script kiddie application which will list all wi-fi networks in range, SSID broadcasting or not, and also the mac address of clients on those networks automagically, as well as crack obsolete security like WEP.
So really, anyone who could even think about cracking a WPA or RADIUS network, which would take quite a bit of time and effort and probably days of information gathering to achieve in practice, would find such measures trivial to break.
However, these measures still lower the supportability of your network, which means they would be very costly for something useless. And even worse, because users who had issues with say, your MAC address filter, might not know how to fix them, they might do something stupid to their machine which actually has the net effect of making your network LESS secure. Fun.
Using WPA or MAC address filters would be like arguing that putting a thumbtack on the floor outside a fortress enhances it's security. Objectively undeniable, but still laughable. Sure it will help keep stupid little kids out of your fortress, but those are not the type of people who could never get past the giant walls, moats, archers, etc your actual fortress security employs. On the other hand, this tack, not being in the fortress standards, might actually manage to make miserable the life of a well intentioned, if stupid, servant, guard, etc.
1) If Microsoft provided any software feature for allowing you to boot other operating systems on their console, they'd have to support this function, which would be a relatively expensive thing to do for something that could bring them no real profit since supporting a bootloader is far from trivial.
2) Why should Microsoft have to make it easy to run 3rd party software/operating systems on their hardware? I have hundreds of devices like graphic calculators, AV boxen, monitors, PC peripherals, cars, etc that have reprogrammable computers embedded with them with no clear support for running 3rd party software when it is clearly technically feasible. Just because YOU WANT to be able to easily run other software on this particular device doesn't mean they have to consider that at all. The only truly wrong thing would be if they used lame legal resources to prevent you from reusing your property through your own labor and ingenuity, like "protections" offered via the DMCA, but otherwise they can ethically sell whatever hardware build they want.
Actually, tweets would be challenging to archive compared to traditional texts. The hardest part of archiving is creating a good index and cross-references to the archived material. Most even short traditional documents have some sort of natural outlining and some sort of built in summary with keywords in it, like an abstract or executive summary.
A tweet, however, might be best described by keywords that do not even exist in it but rather only by the material it links to or even just the events surrounding it. It may therefore require far more metadata, in words, to describe the significance of a short statement than the statement itself.
Most Chinese/asian made components are metric, but these sort of high tech/super high quality American made components are usually only Imperial. "Scientists" can do their maths in whatever units they like, but at some point they have to use components that actually exist. And even many of these metric parts have fun values like "2.54 mm" as standards...
A big part of the reason high quality components are Imperial is because that is what all the shop equipment used to make them are designed for, and replacing them all would be impractical since they are very expensive and now too low volume for a re-investment to ever see returns.
My guess is this service is not marketed toward those interested in truly fault tolerant systems. Rather, it is for those who are running your standard "one beige box server with some sort of regular backup" who would like to squeeze a little bit more average uptime out of it for cheap and no effort. Many many small and midsize businesses fall into this sort of infrastructure category.
As a professional electronics engineer, my assessment is that there is no PROPER way to build a 14 TeV particle accelerator. Point me to the application note for it if I happened to overlook it.
The years and dollars aren't gone. Most of the effort of Ares from what I can tell has been relearning how to do what we did in the past, slightly better, with modern technology and the team at NASA now. Its not like that team will magically forget everything they learned with that time and money if the White House and Congress want a different rocket. They will only lose the marginal differences between the old design and the new design requirements.Overall, it could actually save years and dollars if the new design winds up superior to the last, in a certain way of thinking.
I doubt it. The pricing of Windows has changed very over the years. So unless they planned for PCs in the late 200x's to be around the same cost as a windows license in 1995, It is mostly just incidental.
Porting Firefox to the iPhone/iPad under even modified appstore rules that would allow another browser would be VERY hard, due to the absurd language restriction.
Slashdot Mirror
They might not have to change the licensing at all, as it seems like the amount of data and number of transactions per unit time on a typical SQL server will probably continue to scale faster than the ability of CPU makers to throw more cores on a chip. Especially once this downturn ends and web app development continues full steam again and starts to hit those database servers harder than ever in addition to increased user base.
It will make shot noise worse but may be used to improve other types of noise and distortion. Unfortunately, camera makers have already gotten much better at accommodating other flaws(at least, for domestic imaging) save shot noise which is a fundamental physical issue.
It's not strictly pointless. Oversampling has some uses, namely as a way to filter out random noise and other imperfections caused by the sensor itself. If you can model the random noise, and have many samples for the same atomic datum, you can find ways to still improve the SNR of the data. It still won't let you zoom in to see more detail, however, but still not totally useless for all applications.
A half inch jack wouldn't make any sense in practice. At that point two smaller connectors could have a more compact footprint with less noise. Even an XLR would be a competitive option. And of course in this era, you could run digital coax/optical at 192khz with a smaller footprint and see even less noise with both channels[or more] and having the ADC as close to the source amplifer on-board, as possible.
But that's exactly the way changing OS architectures and APIs can help. Right now the default behavior is to start a worker thread of some type that blocks on IO requests and then reports back. Most apps in the wild don't even bloody do this and just have a few threads do everything and some even have the main app loop block on IO.(let's all pretend we don't see our app windows grey out several times a day!)
We've argued for decades this was a programmer issue but that sort of pedantic criticism has accomplished nothing. Arguing that if all programmers were as smart as you we'd have no problems has been a bad attitude from the OS folks.
Maybe its time we abandoned the current basic Win32/stdlib type APIs for ones that were more IO friendly so that the easiest and most lazy-friendly way to write code is also the best practice for parallelism, and not the other way around which has failed miserably.
After Brown was devastated when he found out his DBZ and Sailor Moon SVGA wallpapers site with 2482 hits was closed down by the whims of the private sector after a good 14 year run, he decided that only the public sector could be trusted with running such a critical service.
Parent is totally inaccurate. Windows supports OpenGL fine.
The issue is that Microsoft usually ships Windows with subpar distributions of graphics drivers for 3D support in general(not just OpenGL) compared to actual vendor drivers in terms of feature support beyond what you need to use basic Windows programs. Probably in some attempt to provide marginally more stable drivers historically under the assumption most users wouldn't need anything more than basic API support and everyone else would have their apps somehow manage to install/warn them to get a better driver package. This doesn't really work for WebGL, however.
Google could just have their installer get you to update your GPU driver, but overall this is probably a cleaner solution for long term support since you have about +/- 10 year support expectations for Direct3D on a GPU in a way that is far better tested than OpenGL for "last years model".
People love biased reporting so long as the bias is the same as theirs. Better yet when the bias just happens to be in their best interest as well.
By investing in foreign markets, you share in their profits, making the "trade" balance a non-issue. The key problem with achieving this in practice is there are really no guarantees that your rights will be respected by the foreign parties when they realize they no longer need your continued investment.
Its also worth noting that big corporate research may in some cases accomplish great innovation but it usually does little for the companies. PARC and Bell Labs did a lot for the world economy to this very day, but did little to help their parent companies out. Even now, look at what researchers at IBM and Microsoft do with absolutely massive budgets in terms of innovation in their key competency areas vs. smaller companies with incomparable or even no dedicated research.
Virtual memory isn't even close to the same thing. Harvard Architecture processors literally execute from ROM. Their program counter points to ROM and they literally cannot grab instructions from elsewhere in the basic architecture. The reason this is important to note is that in theory basic Harvard programs require ZERO ram to work(and indeed, a Hello World program for most Harvard microcontrollers would require none). x86 processors fetch instructions from RAM and therefore require something memory mapped at least to fetch instructions from. So saying the PIC has less resources is misleading as it uses it's resources in a different way so it's apples to oranges.
Virtual memory is just a system feature that allows for on-the-fly access of data in ROM as if it were already in memory. It is still eventually paged to RAM, so it makes no difference in this case and isn't really comparable at all.
But PICs use a Harvard architecture so they are not comparable in terms of RAM to an x86 which must work completely from RAM.
Less than could go wrong with being fully dependent on sea lanes for your economy, especially when one of your political rivals completely dominates world sea power.
But who can say what is best to have? Someone with a horrible genetic disease might be made immune to some future superbug by that condition, and therefore this "diseased" line would be more fit than any other. Likewise, they may require far less food energy to live than the average human, making them fit for some sort of future global famine an Olympic medalist with a Ph.D. in Physics might not make it through.
Traits modern humans seem to value like excess intelligence may not be beneficial in survival scenarios beyond the already high intelligence of pre-history homo sapiens sapiens, but other unpredictable factors may be.
How exactly? Using exploits in non-deprecated wireless security is far more technically involved than running some script kiddie application which will list all wi-fi networks in range, SSID broadcasting or not, and also the mac address of clients on those networks automagically, as well as crack obsolete security like WEP.
So really, anyone who could even think about cracking a WPA or RADIUS network, which would take quite a bit of time and effort and probably days of information gathering to achieve in practice, would find such measures trivial to break.
However, these measures still lower the supportability of your network, which means they would be very costly for something useless. And even worse, because users who had issues with say, your MAC address filter, might not know how to fix them, they might do something stupid to their machine which actually has the net effect of making your network LESS secure. Fun.
Using WPA or MAC address filters would be like arguing that putting a thumbtack on the floor outside a fortress enhances it's security. Objectively undeniable, but still laughable. Sure it will help keep stupid little kids out of your fortress, but those are not the type of people who could never get past the giant walls, moats, archers, etc your actual fortress security employs. On the other hand, this tack, not being in the fortress standards, might actually manage to make miserable the life of a well intentioned, if stupid, servant, guard, etc.
1) If Microsoft provided any software feature for allowing you to boot other operating systems on their console, they'd have to support this function, which would be a relatively expensive thing to do for something that could bring them no real profit since supporting a bootloader is far from trivial.
2) Why should Microsoft have to make it easy to run 3rd party software/operating systems on their hardware? I have hundreds of devices like graphic calculators, AV boxen, monitors, PC peripherals, cars, etc that have reprogrammable computers embedded with them with no clear support for running 3rd party software when it is clearly technically feasible. Just because YOU WANT to be able to easily run other software on this particular device doesn't mean they have to consider that at all.
The only truly wrong thing would be if they used lame legal resources to prevent you from reusing your property through your own labor and ingenuity, like "protections" offered via the DMCA, but otherwise they can ethically sell whatever hardware build they want.
Actually, tweets would be challenging to archive compared to traditional texts. The hardest part of archiving is creating a good index and cross-references to the archived material. Most even short traditional documents have some sort of natural outlining and some sort of built in summary with keywords in it, like an abstract or executive summary.
A tweet, however, might be best described by keywords that do not even exist in it but rather only by the material it links to or even just the events surrounding it. It may therefore require far more metadata, in words, to describe the significance of a short statement than the statement itself.
That's exactly my point. Even the most popular standard "metric" parts in design use today are really Imperial.
Most Chinese/asian made components are metric, but these sort of high tech/super high quality American made components are usually only Imperial. "Scientists" can do their maths in whatever units they like, but at some point they have to use components that actually exist. And even many of these metric parts have fun values like "2.54 mm" as standards... A big part of the reason high quality components are Imperial is because that is what all the shop equipment used to make them are designed for, and replacing them all would be impractical since they are very expensive and now too low volume for a re-investment to ever see returns.
My guess is this service is not marketed toward those interested in truly fault tolerant systems. Rather, it is for those who are running your standard "one beige box server with some sort of regular backup" who would like to squeeze a little bit more average uptime out of it for cheap and no effort. Many many small and midsize businesses fall into this sort of infrastructure category.
As a professional electronics engineer, my assessment is that there is no PROPER way to build a 14 TeV particle accelerator. Point me to the application note for it if I happened to overlook it.
The years and dollars aren't gone. Most of the effort of Ares from what I can tell has been relearning how to do what we did in the past, slightly better, with modern technology and the team at NASA now. Its not like that team will magically forget everything they learned with that time and money if the White House and Congress want a different rocket. They will only lose the marginal differences between the old design and the new design requirements.Overall, it could actually save years and dollars if the new design winds up superior to the last, in a certain way of thinking.
I doubt it. The pricing of Windows has changed very over the years. So unless they planned for PCs in the late 200x's to be around the same cost as a windows license in 1995, It is mostly just incidental.
Porting Firefox to the iPhone/iPad under even modified appstore rules that would allow another browser would be VERY hard, due to the absurd language restriction.
...because in Soviet Russia, you don't own a Saab...oh wait, I guess things are the same. :(