I meant to add, I won't be spending that kind of cash on dedicated RAID hardware again. I'm not convinced its any faster, and it's much cheaper doing it all in the CPU and main RAM. Not to mention more configurable.
If you want to do RAID at the partition level (or logical volume level) rather than being restricted to having whole devices only as RAID group members, I don't know if any modern hardware RAID implementations can do that, but the Linux "md" RAID driver certainly can. I use it in conjunction with lvm (logical volume management) to get *complete* flexibility.
For example, try this, using two disks: First, format each of the two disk with three primary partitions: 1: LVM (just under half total capacity of the whole disk 2: swap partition 3: LVM (just under half again)
For each disk, a volume group is created out of its two large partitions. The two volume groups from the two disks are then configured as members of a RAID 1 group using the md driver (I could just as easily have set it up as "RAID 0" if my main concern had been write performance or throughput on large reads).
The benefits of this configuration are that I get all my RAIDed data space mounted as a single partition, therefore no worries about relative partition sizes etc; and I still get to have two independent (non-RAID) swap areas (thus parallelizing the swap load to some degree) each situated in the middle of its disk in order to minimize head movement during heavy swap usage.
It is actually possible to have even your root partition arranged like this if you know what you are doing. I had to do mine by hand because SuSE's YaST installer couldn't quite figure it out: essentially there needs to be a small, separate non-lvm, non-md boot partition; and the lvm and md drivers need to be available to the kernel at boot time before the system wants to remount root with the real root filesystem.
You can always compile these two modules into the kernel, but lvm needs access to some ancillary files as well anyway - so you still have to build the initrd (for the boot-time initial "ramdisk" root file system) to suit. Hence it's usual to keep lvm and md as loadable modules and just copy them into the initrd as well.
Ever since 1998 or so I've been running my two main Linux boxes on a couple of (now antiquated but still-good) DPT cacheing RAID controllers. But they were very expensive, costing about GBP900 back then even *without* the prohibitively expensive ECC cache SIMMs.
In "Eternity", the sequel to "Eon", Ser Olmy returns to 21st(?) century Earth having taken a *very* long round trip via the opposite end of the universe - which turns out incidentally to have been be horn shaped - in the traditional sense of a curved or rolled up tube with a wide bell-like flaring at one end.
The big difference between it and mail we have now is that only the notification of mail is sent, not the mail itself. The mail sits on the senders mailserver, waiting to be picked up, and if you want to retrieve it, your mail client does so from his server.
This sounds a bit like IMAP actually. But in practice its almost as much of a nuisance having to wade through dozens of headers to check if there is any mail at all from someone you know.
> Now that the hardware has caught up with what Linux really needs to breath
For your information, the hardware of five years ago was *more* than fast enough for the Linux of five years ago..
It might have been forgotten now, with all the bloat that has infested the kernel (yeah yeah recompile etc) and the desktop, but one of the main attractions of early Linux was that it did indeed run very quickly on modest hardware where Windows 95 or NT struggled.
I dunno, If it had all that interesting stuff popping out of it all over the event horizon it wouldn't much resemble the black holes of current theory in which (Hawking insisted) black holes "have no hair" which I took to mean that they cannot carry any distinguishing marks. An event horizon is an event horizon is an event horizon. The only fundamental property it has is its size, and all its other properties are derived from that.
You'd need a new theory altogether, pretty much throwing the book away.
Because the escapeing particle was never in the event horizon to begin with, it can contain no information from within the black hole.
Except that the pair of virtual particles are an entangled pair and if one catches the escaped one and measures its quantum state, one then knows the quantum state of the one that fell in. Catch enough of them and you know about an appreciable fraction of the black hole (in theory!)
Now, how the black hole doesn't gain mass from the anti-particle I'm not quite sure
The energy that was used to create the virtual pair came from the black hole's gravitational field, thus robbing the hole temporarily of mass. For each "virtual" particle that escapes as Hawking radiation, that mass is lost permanently so the mass of the hole goes down, over time. Now remember that this loss can only happen at the event horizon; if the black hole is very large, the tidal force (the gravity gradient) at the event horizon will be weak and thus the rate of particle loss will be very low. For very small black holes the tidal force at the event horizon will be enormous and almost all virtual pairs close to the boundary will separate in this way.
So large black holes will simmer coldly, shrinking only with glacial slowness if at all, and small ones will be hot and shrink very rapidly indeed - finally disappearing altogether in an brief, intense burst of radiation, according to Hawking's theory.
they said that it was more expensive than keeping what they had
Ballmer knows exactly what they said. But he also knows that many people out there, particularly those who harboured doubts about the move, will not bother to check up on the facts. Instead they will go around repeating Ballmer's words and they will be believed by many.
It's called the "Big Lie" and it is a spin technique that was pioneered, and openly praised, by Hitler's Propaganda minister Josep Goebbels,
I have whole freezers full of mammalian, bacteria, and yeast cells that I can though out and revive with a usual 80% efficiency.Freezing cells for later use is VERY common in the research world.
Yeah I have meat, yoghurt and bread dough in my freezer as well. Its very common in the "breakfast lunch and dinner" world as well.
Actually some of these processes are fairly well understood. eg. warmer waters plus increased dissolved CO2 in the ocean causes massive algal blooms which wipe out all the fish.
However, your so-called "theory" is just uneducated armchair bullshit because - as anyone who has done high school chemistry would know - the solubility of gases in water decreases as temperature increases. So as the ice melts and the oceans get warmer, the dissolved CO2 would in fact be released into the atmosphere, accelerating the process.
I went to see this show in London, and I was very impressed. Impressed enough that I had vague plans to take my mother to see it. But this news has entirely put me off, simply because it's hard to imagine that this could be done without cheapening the overall experience. I've no wish to travel to the piss-sodden shit-hole that is London's West End these days, and pay megabucks for a ticket to a musical that doesn't have real musicians. Cameron Mackintosh can fuck off.
I find it difficult to believe that someone went from being a perfectly normal person to a killer sjust he viewed some internet porn.
Not very well acquainted with the darker side of human nature, are you?
It's not as simple as being instantly corruped by a single image, but it is true that impulses to sex, violence (and particularly sexual violence) which are simple consequences of human nature, can be encouraged to find expression when fed with suitably inflammatory images.
Such images don't deliver your mind to any particular destination, but they do help to turn it in a particular direction and give it a little shove.
Unfortunately the public consciousness of this subject that would be required to illuminate the truth so that everybody could understand this, is unlikely to emerge soon; the subject is far too sensitive. Anyone who attempts to discuss crimes like paedophilia in an objective way, i.e. as an outgrowth and inevitable consequence of human nature) is likely to be shouted down.
If that were true, half of Slashdot readership could turn into killers!
Yes indeed, fortunately there are more factors in play, such as moral education, law, social taboo etc. which are reasonably effective at discouraging all but the weakest minds from allowing themselves to be sucked into dangerous mental habits.
For what it's worth, I am generally anti-censorship. And ultimately if we have to choose between politically motivated censorship and no censorship at all then I choose the latter. But a better way would be to acknowledge the weaknesses in human nature and keep the most negative sexual material out of harm's way.
How the hell was this determined? Particularly the wobbly Uranium nucleus. Is it just a theory based on mathematical predictions, or is it actually based on direct observations like X-ray, neutron or electron diffraction studies?
Since this is clearly a very personal issue for you, I assume that you are also either a heavy abuser of chemical stimulation, or have sustained some type of brain lesion. Or possibly both.
Anyway I was just agreeing with Galvatron. Why pick on me?
This would have been a good thing if it had happened five years ago. But the shortsightedness of the standards bodies concerned (what fscking use is a standard that nobody is allowed to read?) kept it out of our hands so all the Linux software we know and love may have diverged from the original standard somewhat.
So why the belated change of heart? They had no choice, that's why. Let me illustrate:
As open source adoption in the enterprise has spread, it's increasingly the case that the ONLY exposure most people have to POSIX is through the bastardized interpretations available through open source software. For example, it seems that most Solaris deployments (in the investment banking industry at least) are still, even today, set up with the C shell as the login default. It seems to be a universal that frustrated programmers and admins then smuggle in bash, zsh, whatever.
Soon enough these OSS shells become the shell of choice throughout the system and people then get tempted to use them for scripting as well because of the POSIX feature set originally borrowed from the Korn shell.
Now sometimes these scripts are specific to the adopted shell, but conscientious scripters (who often don't know if these OSS shells have been installed on any given system) will instead leverage their POSIX knowledge from these shells to write Korn shell scripts instead, which they know can be migrated to any Solaris system without hassle.
Thus these techs become familiar with a POSIX standard, and refer to it as POSIX, possibly without realizing that it isn't necessarily authentic POSIX.
The standards bodies must surely have realized that they were therefore effectively losing control of the standard altogether. The de facto owners of a standard are the maintainers of the most commonly used software which purports to follow that standard (q.v. Microsoft's erstwhile plans to "embrace and extend" Java).
Personally I think it is by this point mostly irrelevant. Modern implementations - even commercial implementations - do diverge somewhat because of holes and historical errors in the standards themselves, where the policy of vendors differs in how to deal with these ambiguities. The OSS approach has been to always make the software do, by default, what the user most likely intended it to do. Where idiocy is required to make it follow the standard strictly, this is available by setting some environment cvariable or whatever. Sun Microsystem's approach has been to slavishly obey every crappy aspect of the standard to the letter. Thus, for example, some Sun versions of egrep (on Solaris 8 I think) will fail to work properly at all if you provide an pattern like '^thing(foo|bar|)$' containing an empty alternate - because this is undefined in the relevant POSIX standard. Whereas the GNU version instead matches the null branch correctly to a null string - it will match lines consisting of "thingfoo", "thingbar" and "thing". This makes a whole lot more sense than the POSIX standard does.
So in my opinion this is too little too late. GNU should have had full representation on the POSIX committee a long time ago. At this stage in the game it's just far too late. What should happen now (and probably will) is that the POSIX standards will slowly be brought into line with existing practice as in the most popular implementations. If not, these committees are only hastening their own end by making themselves irrelevant.
According to the "Vickrey Auction" article, another name for this "uniform second-price auction" is "Dutch Auction". These do indeed take place on eBay but it is not the default option and is normally only used when the seller has multiple identical items to sell.
The principle of the Dutch Auction is that where there are n items to be sold, at close the top n bids win but the transaction is completed at the price of the lowest successful bid (which only happens to be the second highest bid in auctions where n=2).
Slashdot Mirror
I meant to add, I won't be spending that kind of cash on dedicated RAID hardware again. I'm not convinced its any faster, and it's much cheaper doing it all in the CPU and main RAM. Not to mention more configurable.
If you want to do RAID at the partition level (or logical volume level) rather than being restricted to having whole devices only as RAID group members, I don't know if any modern hardware RAID implementations can do that, but the Linux "md" RAID driver certainly can. I use it in conjunction with lvm (logical volume management) to get *complete* flexibility.
For example, try this, using two disks: First, format each of the two disk with three primary partitions:
1: LVM (just under half total capacity of the whole disk
2: swap partition
3: LVM (just under half again)
For each disk, a volume group is created out of its two large partitions. The two volume groups from the two disks are then configured as members of a RAID 1 group using the md driver (I could just as easily have set it up as "RAID 0" if my main concern had been write performance or throughput on large reads).
The benefits of this configuration are that I get all my RAIDed data space mounted as a single partition, therefore no worries about relative partition sizes etc; and I still get to have two independent (non-RAID) swap areas (thus parallelizing the swap load to some degree) each situated in the middle of its disk in order to minimize head movement during heavy swap usage.
It is actually possible to have even your root partition arranged like this if you know what you are doing. I had to do mine by hand because SuSE's YaST installer couldn't quite figure it out: essentially there needs to be a small, separate non-lvm, non-md boot partition; and the lvm and md drivers need to be available to the kernel at boot time before the system wants to remount root with the real root filesystem.
You can always compile these two modules into the kernel, but lvm needs access to some ancillary files as well anyway - so you still have to build the initrd (for the boot-time initial "ramdisk" root file system) to suit. Hence it's usual to keep lvm and md as loadable modules and just copy them into the initrd as well.
Ever since 1998 or so I've been running my two main Linux boxes on a couple of (now antiquated but still-good) DPT cacheing RAID controllers. But they were very expensive, costing about GBP900 back then even *without* the prohibitively expensive ECC cache SIMMs.
Because they do it on their own low-end home theatre kit as well, probably.
In "Eternity", the sequel to "Eon", Ser Olmy returns to 21st(?) century Earth having taken a *very* long round trip via the opposite end of the universe - which turns out incidentally to have been be horn shaped - in the traditional sense of a curved or rolled up tube with a wide bell-like flaring at one end.
So there.
> [...] so I can connect the laptop and make alterations on the fly.
That's fine if you have Geordi LaForge in the co-pilot...engineer...er...passenger seat.
This sounds a bit like IMAP actually. But in practice its almost as much of a nuisance having to wade through dozens of headers to check if there is any mail at all from someone you know.
> Now that the hardware has caught up with what Linux really needs to breath
For your information, the hardware of five years ago was *more* than fast enough for the Linux of five years ago..
It might have been forgotten now, with all the bloat that has infested the kernel (yeah yeah recompile etc) and the desktop, but one of the main attractions of early Linux was that it did indeed run very quickly on modest hardware where Windows 95 or NT struggled.
I dunno, If it had all that interesting stuff popping out of it all over the event horizon it wouldn't much resemble the black holes of current theory in which (Hawking insisted) black holes "have no hair" which I took to mean that they cannot carry any distinguishing marks. An event horizon is an event horizon is an event horizon. The only fundamental property it has is its size, and all its other properties are derived from that.
You'd need a new theory altogether, pretty much throwing the book away.
Except that the pair of virtual particles are an entangled pair and if one catches the escaped one and measures its quantum state, one then knows the quantum state of the one that fell in. Catch enough of them and you know about an appreciable fraction of the black hole (in theory!)
Now, how the black hole doesn't gain mass from the anti-particle I'm not quite sure
The energy that was used to create the virtual pair came from the black hole's gravitational field, thus robbing the hole temporarily of mass. For each "virtual" particle that escapes as Hawking radiation, that mass is lost permanently so the mass of the hole goes down, over time. Now remember that this loss can only happen at the event horizon; if the black hole is very large, the tidal force (the gravity gradient) at the event horizon will be weak and thus the rate of particle loss will be very low. For very small black holes the tidal force at the event horizon will be enormous and almost all virtual pairs close to the boundary will separate in this way.
So large black holes will simmer coldly, shrinking only with glacial slowness if at all, and small ones will be hot and shrink very rapidly indeed - finally disappearing altogether in an brief, intense burst of radiation, according to Hawking's theory.
Ballmer knows exactly what they said. But he also knows that many people out there, particularly those who harboured doubts about the move, will not bother to check up on the facts. Instead they will go around repeating Ballmer's words and they will be believed by many.
It's called the "Big Lie" and it is a spin technique that was pioneered, and openly praised, by Hitler's Propaganda minister Josep Goebbels,
AND is implied in any list of barewords; OR is interpreted correctly as selecting either of two alternatives. Thus:
Foo Bar OR Baz Quux
will return matches for both:
Foo AND Bar AND Quux
and:
Foo AND Baz AND Quux
Yeah I have meat, yoghurt and bread dough in my freezer as well. Its very common in the "breakfast lunch and dinner" world as well.
> for anything other than truly unique situations, it's not worth it.
Such as surfing the web while sitting on the toilet. So if tablets take off, expect sales of haemorrhoid cream to soar.
Actually some of these processes are fairly well understood. eg. warmer waters plus increased dissolved CO2 in the ocean causes massive algal blooms which wipe out all the fish.
However, your so-called "theory" is just uneducated armchair bullshit because - as anyone who has done high school chemistry would know - the solubility of gases in water decreases as temperature increases. So as the ice melts and the oceans get warmer, the dissolved CO2 would in fact be released into the atmosphere, accelerating the process.
Can anyone say "Venus"?
Say what you like.
I went to see this show in London, and I was very impressed. Impressed enough that I had vague plans to take my mother to see it. But this news has entirely put me off, simply because it's hard to imagine that this could be done without cheapening the overall experience. I've no wish to travel to the piss-sodden shit-hole that is London's West End these days, and pay megabucks for a ticket to a musical that doesn't have real musicians. Cameron Mackintosh can fuck off.
Christ on a bike... if that's the way you do business in the USA, then I'm glad I'm on the other side of the pond. Who needs that kind of shit.
Not very well acquainted with the darker side of human nature, are you?
It's not as simple as being instantly corruped by a single image, but it is true that impulses to sex, violence (and particularly sexual violence) which are simple consequences of human nature, can be encouraged to find expression when fed with suitably inflammatory images.
Such images don't deliver your mind to any particular destination, but they do help to turn it in a particular direction and give it a little shove.
Unfortunately the public consciousness of this subject that would be required to illuminate the truth so that everybody could understand this, is unlikely to emerge soon; the subject is far too sensitive. Anyone who attempts to discuss crimes like paedophilia in an objective way, i.e. as an outgrowth and inevitable consequence of human nature) is likely to be shouted down.
If that were true, half of Slashdot readership could turn into killers!
Yes indeed, fortunately there are more factors in play, such as moral education, law, social taboo etc. which are reasonably effective at discouraging all but the weakest minds from allowing themselves to be sucked into dangerous mental habits.
For what it's worth, I am generally anti-censorship. And ultimately if we have to choose between politically motivated censorship and no censorship at all then I choose the latter. But a better way would be to acknowledge the weaknesses in human nature and keep the most negative sexual material out of harm's way.
How the hell was this determined? Particularly the wobbly Uranium nucleus. Is it just a theory based on mathematical predictions, or is it actually based on direct observations like X-ray, neutron or electron diffraction studies?
Since this is clearly a very personal issue for you, I assume that you are also either a heavy abuser of chemical stimulation, or have sustained some type of brain lesion. Or possibly both.
Anyway I was just agreeing with Galvatron. Why pick on me?
I wasn't joking. The theory I offered seems like the most likely explanation given the frequency and type of errors involved.
Not everything is an issue of political correctness. Deal.
That piece is so screwed up it can only be drug abuse or brain injury that's behind it, IMHO.
This would have been a good thing if it had happened five years ago. But the shortsightedness of the standards bodies concerned (what fscking use is a standard that nobody is allowed to read?) kept it out of our hands so all the Linux software we know and love may have diverged from the original standard somewhat.
So why the belated change of heart? They had no choice, that's why. Let me illustrate:
As open source adoption in the enterprise has spread, it's increasingly the case that the ONLY exposure most people have to POSIX is through the bastardized interpretations available through open source software. For example, it seems that most Solaris deployments (in the investment banking industry at least) are still, even today, set up with the C shell as the login default. It seems to be a universal that frustrated programmers and admins then smuggle in bash, zsh, whatever.
Soon enough these OSS shells become the shell of choice throughout the system and people then get tempted to use them for scripting as well because of the POSIX feature set originally borrowed from the Korn shell.
Now sometimes these scripts are specific to the adopted shell, but conscientious scripters (who often don't know if these OSS shells have been installed on any given system) will instead leverage their POSIX knowledge from these shells to write Korn shell scripts instead, which they know can be migrated to any Solaris system without hassle.
Thus these techs become familiar with a POSIX standard, and refer to it as POSIX, possibly without realizing that it isn't necessarily authentic POSIX.
The standards bodies must surely have realized that they were therefore effectively losing control of the standard altogether. The de facto owners of a standard are the maintainers of the most commonly used software which purports to follow that standard (q.v. Microsoft's erstwhile plans to "embrace and extend" Java).
Personally I think it is by this point mostly irrelevant. Modern implementations - even commercial implementations - do diverge somewhat because of holes and historical errors in the standards themselves, where the policy of vendors differs in how to deal with these ambiguities. The OSS approach has been to always make the software do, by default, what the user most likely intended it to do. Where idiocy is required to make it follow the standard strictly, this is available by setting some environment cvariable or whatever. Sun Microsystem's approach has been to slavishly obey every crappy aspect of the standard to the letter. Thus, for example, some Sun versions of egrep (on Solaris 8 I think) will fail to work properly at all if you provide an pattern like '^thing(foo|bar|)$' containing an empty alternate - because this is undefined in the relevant POSIX standard. Whereas the GNU version instead matches the null branch correctly to a null string - it will match lines consisting of "thingfoo", "thingbar" and "thing". This makes a whole lot more sense than the POSIX standard does.
So in my opinion this is too little too late. GNU should have had full representation on the POSIX committee a long time ago. At this stage in the game it's just far too late. What should happen now (and probably will) is that the POSIX standards will slowly be brought into line with existing practice as in the most popular implementations. If not, these committees are only hastening their own end by making themselves irrelevant.
According to the "Vickrey Auction" article, another name for this "uniform second-price auction" is "Dutch Auction". These do indeed take place on eBay but it is not the default option and is normally only used when the seller has multiple identical items to sell.
The principle of the Dutch Auction is that where there are n items to be sold, at close the top n bids win but the transaction is completed at the price of the lowest successful bid (which only happens to be the second highest bid in auctions where n=2).
I found exactly the same thing. Actually I was astounded that this 600-year old version of my native tongue was so readily comprehensible.
It's a far cry from Anglo Saxon - I can hardly understand Boewulf at all when it's read aloud.
That's the funniest thing I have seen so far this year!