No, it was not the star that "left" the galaxy, but the gas that later formed it. The gas compression period should last long enough for the matter to escape from vincinity of it's mother galaxy.
Even a hundred years ago many people were still living in hunger. These days anyone usually can eat as much as he likes (for some it's a question of how expensive food they choose). So it's not a surprise that now rich societies have both obesity and diabetes epidemic.
Genetical selection throughout history generally made much of poplation used to living without high sugar blood levels, and for those susceptible it now causes health problems and uncontrolled increase increase of body weight.
Medical solution would certainly change the world.
I liked 1.5 way of handling tabs. With 2.0 they somehow screwed it up, but I didn't take time to patiently search through all tabbing extensions to get that back, o got used to a new scheme. Otherwise I don't see much difference. UI feels snappier, while in 1.5 it was known to block for short moments. On my home PC I still used 1.5 few until a few days ago (because of extensions but I realized that now that all extensions which I can't live without have been made compatible with FF2.0+, so I moved to 2.0).
I guess it won't be too long before I move on 3.0 beta, if the performance is going that improved (yes, AJAX websites are not really fast on firefox currently).
I think it's still not impossible to develop C++ programs for it, if full stack will be open source (and if hardware won't block unsigned execution of core components). I guess a system daemon and C++ bindings could be written to allow pure C++ executables to bind over IPC mechanism to Java framework and cooperate with the Android system (same as Java app). It remains to be seen if Google even wants to allow that, though.
However I'm almost sure that vendors, maybe carriers will be putting closed source core components inside, to block certain kinds of system usage (e.g. blobs that do (from-userspace) programming of wifi/gsm/3D/DRM etc. hardware). Luckily GPLv2 prevents doing that to Linux kernel, so reverse engineering could be possible (unless devices would allow signed-only kernels to load).
Let's hope that we will be able to modify also non-Java Android components, like codec library (is that one even open source?).
Firefox' extension system is a key to it's success, not the Gecko engine (although Webkit didn't even exist when it started gaining market share). FF for desktop, Webkit for mobiles seems to be a future direction (with a mandatory IE).
Well, but Java is a good tool for bad coders. And naturally those are most common. With Java they are more likely to be prevented from doing bad stuff (like creating apps which gradually consume all memory).
The guy just can't be more right about carriers breaking progress in mobile market. Of course freezing current situation is what they want because of HUGE sums of money they earn this way. VoIP or IM availability might create a disaster and a big hole in their pockets.
Fortunately Apple got the stones rolling and now competition will pop up everywhere (or be pushed out of mobile shops). From carriers to big phone brands. Yes, Apple sc*ewed them badly with iPhone feature breakthrough and now there is no way back. It's finally happening, and everyone still has a chance to be in this market (as there is no real "Windows" this time around).
Vendors have been pushed to port their apps to Linux and yet many ignored that plea. So why are they reluctant?:
-tiny desktop ("market") share of Linux -non-mature and often changing of environment and API's/ABI's (mostly userspace) -multiple desktop environments to support -binary compatibility nightmare accross distributions -open source mentality - many people dislike non-OSS applications on Linux -windows-tie in: apps written in VS, running on a native Windows API's.
We have seen some porting though. NeroLinux, Picasa, GoogleEarth among them. This is also a point where Wine can be helpful, but native app, of course, is the best solution. Mono could become really important porting tool when/if companies move their big Windows products to NET.
They also push it via Xbox 360. I mean, it's not hard to support either OGL or DX on Xbox, as ATI produces also OpenGL driver, but they obviously chose to do that to promote their API (why don't MS f*ckers submit it to ISO as well?).
Fortunately Wii and PS3 are OpenGL/OpenGL-ES so noone can actually base on ONE graphical API, except MS-owned developers.
Regarding sound, there has been something called "dmix" for ages. As far as more recent development is concerned, there's PulseAudio which is supposed to bring some order to Linux sound handling (which PA author also calls balkanized).
As for ABI's, there is a good reason (and purely techincal) why there isn't one. It's to keep ability to improve kernel, and not be held hostage by proprietary drivers for hardware without public specs (for which vendor doesn't care much anymore). YOU still have a choice to use a proprietary driver, as long as vendor cares about updating it for newer kernels.
(and even binary blobs can be automated to re-install with a new kernel as interface is often open source component, so people patch it)
PM indeed is one of biggest problems for Linux currently. I hope that they will finally pick one of available suspend/hibernate solutions and adopt into the kernel. Indeed, tickless changes are a start and even userspace is getting fixed lately. Such state is not because of server having a priority, but because power management until recently wasn't a problem for desktop machines and laptops had far bigger problems (ACPI, wireless, graphic cards) anyway, mostly because laptops were (are) built for Windows.
Yep, there could be a "linux-compatible" trademark as a reward for companies which have drivers in the vanilla kernel.
Linux is scalable. It tries to run well on "big iron", but it is engineered carefully not to harm desktop. Heck, Linux is even popular on mobile phones.
Reason that Linux still bit lacks in some desktop bits (e.g. power management) is because main Linux devs don't tend to put immature stuff in the kernel, nor will they do it if "solution" cripples performance on either high-end or low-end hardware.
So it's probably limited by synchronisation; thread-switch overhead which increases with framerate. That is much less present with multicore as threads do their jobs independently (and even HT helps here)
Good example of multicore performance increase is PCSX2 emulator - 2 cores bring average 100% FPS increase over single.
16-bit colorspace (and graphic card mode) which is probably something Nvidia nor ATI don't care as much as before (these days all games are true color, i.e. 32-bit framebuffers and often even higher color transformation precision)
Fortunately, it is irrelevant where Tor server actually runs:) It seems that idea of such directives is to prevent common case of communication from becoming really secure, so that anyone can be a suspect just if he/she ever used that method way of communication. For that reason we won't soon (or ever) see secure authentication and exchange of decryption keys in e.g. mobile-phones: so that police can tune in and listen whenever they want. Although we already see this "problem" with VoIP which is widely used as replacement for a fixed telephony.
It is not collected by authorised and trusted institution (like police), instead it is performed by some company and this is no proof in a court (at least not in Europe contries)
Atlas had a luck. As the triplet is near to the CMS, there is a (small) possibility that gas shockwave damaged some of it's on-detector hardware (people are discussing it here at CERN, but not much is known yet about the schedule impact).
The news is unfortunate though. It might mean that there could be only a new cosmic run this year run instead of real beam if the repair of this takes time.
True, Descent was ahead of it's time in many ways. AI was very clever and thanks to that game was challenging. Of course, simple geometry certainly made AI algorithms simpler to implement, these days gameplay scenery is much more complex and making AI aware of all that can be bug prone.
I have no problem using windows-style behavior independently to linux style copy-paste (example with firefox can work with window shortcuts as long as it's between more recent toolkit apps). GTK and Qt programs seem to happily cooperate here this. Of course there are more traditional apps (Vim, Emacs, Console) and their communities find it hard to interact with rest of OSS realm, but for those applications unix-style universally works (as well as for the consoleDE cases and it isn't bd once you get used to it).
Maybe Linux is not at the level of interaction like in OSX where you can even drag'n'drop picture from firefox to powerpoint, but in fact at freedesktop.org there are descriptions which try to define such interfaces and they are even implemented in most popular DE's (although applications like Openoffice or firefox maybe don't follow enough of it). Portland is another project that tries to deal with this kind of stuff and bring in coherence, so there is ongoing progress in this area.
Don't know about Windows efficiency, but with dual core chips all this seems possible in software and without adding much overhead. Clearly this NIC targets gamers, so if Windows gaming performace stays as described in the article, the card might improve latency and FPS (although, as we see, by few percent only).
They seem to have a callback which sets a bit ready in userspace (i.e. in game process memory), so game doesn't have to synchronously poll for data. This is also possible to do with on-board NIC's, provided that driver framework allows it, as it is still the kernel that sets this bit. Recently lot of attention in Linux community was given to async interfaces to userspace, among them kevents, threadlets (lightweight threads) etc. Such interfaces could allow drivers to easily send data "ready" events to userspace callback, which sets e.g. a data ready bit, so main game thread doesn't need to poll and wait for it.
There are also similar interfaces in Windows, only the on-board drivers should probably be enabled to use it (and btw. there is a software QoS as well).
Another point is that multicore CPU's are now mainstream and a tasks, either kernel or userspace based, which e.g. processes network data and talks to hardware can be run at no visible cost to game performance, and in some cases this can even yield better latency as CPU's are faster than a NIC processor (for example with wireless chips when en(de)cryption is used).
If you have a gaming machine, you probably don't even care about electricity (especiall as low-end machine with integrated graphics will need a fraction of power used by a monster with an overclocked Athlon-FX X2 and a SLI nvidia 8800 setup).
And gamer's parents typically don't know how much power a PC can consume:)
So you think that if you eat a meat from young cow that you will be growing as well? Not really. Hormones break down in digestive system and don't come into bloadstream.
So it doesn't matter if the cow was genetically engineered or not. For you it is just amino acids, sugar and fat. (yes, you can eat snake poison and won't get poisoned)
Of course as a grownup you certainly burn more calories when doing something intensive. In children it is possible that similar amount energy is always spent, just differently balanced (metabolism & growth vs. just mechanical movement). Otherwise you also can't explain how some children (but also even adults) can't gain weight no matter how much they eat.
Slashdot Mirror
No, it was not the star that "left" the galaxy, but the gas that later formed it.
The gas compression period should last long enough for the matter to escape from vincinity of it's mother galaxy.
Even a hundred years ago many people were still living in hunger. These days anyone usually can eat as much as he likes (for some it's a question of how expensive food they choose). So it's not a surprise that now rich societies have both obesity and diabetes epidemic.
Genetical selection throughout history generally made much of poplation used to living without high sugar blood levels,
and for those susceptible it now causes health problems and uncontrolled increase increase of body weight.
Medical solution would certainly change the world.
I liked 1.5 way of handling tabs. With 2.0 they somehow screwed it up, but I didn't take time to patiently search through all tabbing extensions to get that back, o got used to a new scheme. Otherwise I don't see much difference. UI feels snappier, while in 1.5 it was known to block for short moments. On my home PC I still used 1.5 few until a few days ago (because of extensions but I realized that now that all extensions which I can't live without have been made compatible with FF2.0+, so I moved to 2.0).
I guess it won't be too long before I move on 3.0 beta, if the performance is going that improved (yes, AJAX websites are not really fast on firefox currently).
That would violate FCC regulations.
I think it's still not impossible to develop C++ programs for it, if full stack will be open source (and if hardware won't block unsigned execution of core components). I guess a system daemon and C++ bindings could be written to allow pure C++ executables to bind over IPC mechanism to Java framework and cooperate with the Android system (same as Java app). It remains to be seen if Google even wants to allow that, though.
However I'm almost sure that vendors, maybe carriers will be putting closed source core components inside, to block certain kinds of system usage (e.g. blobs that do (from-userspace) programming of wifi/gsm/3D/DRM etc. hardware). Luckily GPLv2 prevents doing that to Linux kernel, so reverse engineering could be possible (unless devices would allow signed-only kernels to load).
Let's hope that we will be able to modify also non-Java Android components, like codec library (is that one even open source?).
Firefox' extension system is a key to it's success, not the Gecko engine (although Webkit didn't even exist when it started gaining market share). FF for desktop, Webkit for mobiles seems to be a future direction (with a mandatory IE).
Well, but Java is a good tool for bad coders. And naturally those are most common.
With Java they are more likely to be prevented from doing bad stuff (like creating apps which gradually consume all memory).
The guy just can't be more right about carriers breaking progress in mobile market. Of course freezing current situation is what they want because of HUGE sums of money they earn this way. VoIP or IM availability might create a disaster and a big hole in their pockets.
Fortunately Apple got the stones rolling and now competition will pop up everywhere (or be pushed out of mobile shops). From carriers to big phone brands. Yes, Apple sc*ewed them badly with iPhone feature breakthrough and now there is no way back. It's finally happening, and everyone still has a chance to be in this market (as there is no real "Windows" this time around).
Vendors have been pushed to port their apps to Linux and yet many ignored that plea. So why are they reluctant?:
-tiny desktop ("market") share of Linux
-non-mature and often changing of environment and API's/ABI's (mostly userspace)
-multiple desktop environments to support
-binary compatibility nightmare accross distributions
-open source mentality - many people dislike non-OSS applications on Linux
-windows-tie in: apps written in VS, running on a native Windows API's.
We have seen some porting though. NeroLinux, Picasa, GoogleEarth among them. This is also a point where Wine can be helpful, but native app, of course, is the best solution. Mono could become really important porting tool when/if companies move their big Windows products to NET.
They also push it via Xbox 360. I mean, it's not hard to support either OGL or DX on Xbox, as ATI produces also OpenGL driver, but they obviously chose to do that to promote their API (why don't MS f*ckers submit it to ISO as well?).
Fortunately Wii and PS3 are OpenGL/OpenGL-ES so noone can actually base on ONE graphical API, except MS-owned developers.
Regarding sound, there has been something called "dmix" for ages. As far as more recent development is concerned, there's PulseAudio which is supposed to bring some order to Linux sound handling (which PA author also calls balkanized).
As for ABI's, there is a good reason (and purely techincal) why there isn't one. It's to keep ability to improve kernel, and not be held hostage by proprietary drivers for hardware without public specs (for which vendor doesn't care much anymore). YOU still have a choice to use a proprietary driver, as long as vendor cares about updating it for newer kernels.
(and even binary blobs can be automated to re-install with a new kernel as interface is often open source component, so people patch it)
PM indeed is one of biggest problems for Linux currently. I hope that they will finally pick one of available suspend/hibernate solutions and adopt into the kernel. Indeed, tickless changes are a start and even userspace is getting fixed lately. Such state is not because of server having a priority, but because power management until recently wasn't a problem for desktop machines and laptops had far bigger problems (ACPI, wireless, graphic cards) anyway, mostly because laptops were (are) built for Windows.
Yep, there could be a "linux-compatible" trademark as a reward for companies which have drivers in the vanilla kernel.
Linux is scalable. It tries to run well on "big iron", but it is engineered carefully not to harm desktop. Heck, Linux is even popular on mobile phones.
Reason that Linux still bit lacks in some desktop bits (e.g. power management) is because main Linux devs don't tend to put immature stuff in the kernel, nor will they do it if "solution" cripples performance on either high-end or low-end hardware.
So it's probably limited by synchronisation; thread-switch overhead which increases with framerate. That is much less present with multicore as threads do their jobs independently (and even HT helps here)
Good example of multicore performance increase is PCSX2 emulator - 2 cores bring average 100% FPS increase over single.
16-bit colorspace (and graphic card mode) which is probably something Nvidia nor ATI don't care as much as before (these days all games are true color, i.e. 32-bit framebuffers and often even higher color transformation precision)
Fortunately, it is irrelevant where Tor server actually runs :)
It seems that idea of such directives is to prevent common case of communication from becoming really secure, so that anyone can be a suspect just if he/she ever used that method way of communication.
For that reason we won't soon (or ever) see secure authentication and exchange of decryption keys in e.g. mobile-phones: so that police can tune in and listen whenever they want. Although we already see this "problem" with VoIP which is widely used as replacement for a fixed telephony.
It is not collected by authorised and trusted institution (like police),
instead it is performed by some company and this is no proof in a court (at least not in Europe contries)
Atlas had a luck. As the triplet is near to the CMS, there is a (small) possibility that gas shockwave damaged some of it's on-detector hardware (people are discussing it here at CERN, but not much is known yet about the schedule impact).
The news is unfortunate though. It might mean that there could be only a new cosmic run this year run instead of real beam if the repair of this takes time.
Well, the Fermilab produced those quadrupole triplet magnets, you should read more carefully. Not that it's purely their fault, though.
True, Descent was ahead of it's time in many ways. AI was very clever and thanks to that game was challenging. Of course, simple geometry certainly made AI algorithms simpler to implement, these days gameplay scenery is much more complex and making AI aware of all that can be bug prone.
I have no problem using windows-style behavior independently to linux style copy-paste (example with firefox can work with window shortcuts as long as it's between more recent toolkit apps). GTK and Qt programs seem to happily cooperate here this. Of course there are more traditional apps (Vim, Emacs, Console) and their communities find it hard to interact with rest of OSS realm, but for those applications unix-style universally works (as well as for the consoleDE cases and it isn't bd once you get used to it).
Maybe Linux is not at the level of interaction like in OSX where you can even drag'n'drop picture from firefox to powerpoint, but in fact at freedesktop.org there are descriptions which try to define such interfaces and they are even implemented in most popular DE's (although applications like Openoffice or firefox maybe don't follow enough of it). Portland is another project that tries to deal with this kind of stuff and bring in coherence, so there is ongoing progress in this area.
Don't know about Windows efficiency, but with dual core chips all this seems possible in software and without adding much overhead. Clearly this NIC targets gamers, so if Windows gaming performace stays as described in the article, the card might improve latency and FPS (although, as we see, by few percent only).
They seem to have a callback which sets a bit ready in userspace (i.e. in game process memory), so game doesn't have to synchronously poll for data. This is also possible to do with on-board NIC's, provided that driver framework allows it, as it is still the kernel that sets this bit. Recently lot of attention in Linux community was given to async interfaces to userspace, among them kevents, threadlets (lightweight threads) etc. Such interfaces could allow drivers to easily send data "ready" events to userspace callback, which sets e.g. a data ready bit, so main game thread doesn't need to poll and wait for it.
There are also similar interfaces in Windows, only the on-board drivers should probably be enabled to use it (and btw. there is a software QoS as well).
Another point is that multicore CPU's are now mainstream and a tasks, either kernel or userspace based, which e.g. processes network data and talks to hardware can be run at no visible cost to game performance, and in some cases this can even yield better latency as CPU's are faster than a NIC processor (for example with wireless chips when en(de)cryption is used).
If you have a gaming machine, you probably don't even care about electricity (especiall as low-end machine with integrated graphics will need a fraction of power used by a monster with an overclocked Athlon-FX X2 and a SLI nvidia 8800 setup).
:)
And gamer's parents typically don't know how much power a PC can consume
So you think that if you eat a meat from young cow that you will be growing as well?
Not really. Hormones break down in digestive system and don't come into bloadstream.
So it doesn't matter if the cow was genetically engineered or not. For you it is just amino acids, sugar and fat.
(yes, you can eat snake poison and won't get poisoned)
Of course as a grownup you certainly burn more calories when doing something intensive. In children it is possible that similar amount energy is always spent, just differently balanced (metabolism & growth vs. just mechanical movement). Otherwise you also can't explain how some children (but also even adults) can't gain weight no matter how much they eat.
Not that it is any kind of credible science anyway.