This all lies on the assumption that he would, if denied ultimate ownership of the copyright. Recall that Namesys was making a bit of money using that copyright, according to their site.
We invent assumptions, aka axioms, and discover interesting consequences of those assumptions. Those assumptions are often born of our own interaction with the world, such as the idea that the shortest distance between two points is a line, or that two parallel lines never intersect. Whole new mathematical realms are discovered when we relax these rules, but they're not necessarily out there in nature. And there's certainly no rule that nature must follow these rules, everywhere or anywhere.
Like I said, you don't have to trust me, but if you don't trust the people on the committee that is planning to fix it, they've written papers about what's wrong. They do represent the rest of the computer industry, after all. But hey, it's a lot easier to sell what you've got working today in your office and ignore the random odd bugs that you can't duplicate than it is to own up to a minor breakage in assumptions about multithreaded viability, so I can see where you're coming from.
Well yea you could have skipped it, there wasn't enough time to actually change anything. They just tried to shove compiz by default really hard, and introduce upstart. A few updates and gnome, but no time to do anything remarkable.
Yeah , because adding the "volatile" keyword is such a chore. Oh my dear god. You've raised the stakes far too high to be this wrong. This quote is a steaming turd of falsehood. For the benefit of anyone reading this who isn't a troll, volatile only works the way you think it does in Java. In C and C++ the only uses for it is when the hardware itself might write to the variable, common in memory mapped devices. Hans Boehm, co-author of a world class C++ garbage collector runtime, and current member of the C++0x committee gave a talk to Google about how concurrency can be trifled with, from the hardware level to the language specification to optimizations, and what the committee is doing to fix it. I actually had that one sitting on my hard drive but hadn't watched it yet, so I didn't bother trotting it out as well.
This isn't something I've just looked up to spite you. It's come up recently on the websites I read, as GCC recently bit the Linux kernel. I've studied proof techniques for semaphores, locks, monitors and so on. This is the sort of stuff that interests me. If memory ordering and non-atomic writes don't make sense to you, then please watch these and maybe read some of the papers. You don't have to understand me, but if you can't understand them then maybe it's time to stop defending the viability of multithreading today. And probably, you should stop using it.
No, it is fundamentally broken in places. It's not a matter of library support. It's a matter of the compiler taking those libraries and totally fucking with their intended purpose because the language spec lets them and it can improve optimizations.
But I'm glad idiots have opinions. It gives people like this guy (video) something to do, by educating you. What ends up happening is that compilers end up presenting non-standard extensions. I've used an embedded language nesC that would provide atomic blocks and implement them by disabling interrupts (not a great idea when avoidable, but passable if you know the drawbacks).
Well, his complaint is that the language isn't a good low level one, on the grounds that bits matter and until C99 was approved and implemented, was a reasonable assumption. Reasons you might want to know the size of datatypes are easy to come up with: you're defining a network protocol or a file storage format. If you use int and compile for x86 and amd64, for example, you could wind up with two incompatible memory layouts. This is the sort of reason we came up with stdint.h in the first place.
Ah, well there's no use trying to use C++ as an embedded language / assembler. One look at template inheritance and you shall melt like the Nazi's in Indiana Jones.
At the moment, the C language standard is simply broken for multithreading. There's good reasons for it too. Imaging arithmetic on a 32bit int on a platform of 16bit sized words. There's apparently plans to deal with this in C++, and then C will follow on the assumption that C++ is harder to solve, and C will basically try to be as compatible as possible with their complex needs.
Java is actually in a place to handle this problem quite perfectly. They recognized from the start that the memory model needed work and went about defining a suitable language for multithreading. It's built into the language itself, like garbage collection is. This is why it's popular among enterprise developments. With smart people writing multithreaded code (which we lack in any langauge), you can get some pretty nice results and be fairly efficient, despite Java's renown for being slow.
The difference is that people are inventing new hammers daily, with no sign of slowing down, and using them correctly and efficiently can take years of investment. Carpentry is a very old profession, so there's been plenty of time to identify the right tools. But today, with a historically large population, new tools are invented all the time in this profession in it's infancy. These sorts of popularity measurements can justify an institution using a new language over an older one. "See, plenty of people use it. It's not something we'll have to hire a PhD to use and fix!"
Because the law doesn't prohibit genetic testing -- it prohibits companies from acting on such information. You as a consumer of medical goods (and perhaps medical insurance) can still take such tests and buy disproportionate amounts of insurance.
The far bigger problem that I see in insurance today is the near impossibility of comparing two coverages or even understanding the one you might have. The emergence of standard contracts is great for free markets -- it means that you can price shop around easily because the goods are interchangable, and making information available to consumers can reduce prices. (Apparently term life insurance rates dropped magically around the time companies started hosting websites capable of advertising their rates). But neither you nor I know whether insurance will accept your medical claims. When my brother's leg was terribly broken during a football game, the insurance company denied the ambulance ride as it was not "pre-approved". Apparently this is a common practice, though terrible and the insured are almost never aware of this.
Re:Great release unfortunately no Abiword 2.6
on
Ubuntu 8.04 Released
·
· Score: 1
A backport may be likely. The situation for abiword is odd. Their last release was some time ago, and their last major release was in 2005. Additionally, you linked to the 2.6.0 release notes, when 2.6.2 is already out twenty days later. Debian still doesn't have 2.6.x in unstable, so that's an added source of tension. All this is slowly adding up to the typical Open Source poor release engineering. Which isn't bad; just not necessarily LTS material. A long term support distribution has to basically do it's own release engineering in the absence of it upstream.
Ubuntu does work closely with some software groups to track beta / development releases if they have faith in the upstream schedule and so on. If you want software released days or hours before release in a specific version of Ubuntu, it's best to email ubuntu-devel@lists.ubuntu.com and start a conversation, so they know you're a generally reliable upstream committed to quality. It also requires that you have a plan to release far sooner than two weeks notice.
As far as backports go, it depends on how well an enterprising Ryan Pavlik gets with his work. If you're concerned with this problem at all, you should probably be testing his ppa software and make sure it gets a proper backport request.
Actually, 6.06 was an abnormality a result of some terrible asian languages support, and being the first Ubuntu LTS ever. Canonical has made a vow to stick by the launch date schedule ever since. Instead of releasing whenever a feature is done, they include a feature in a given release depending on whether it's done or not. Obviously this is a challenge when bugs arise, and it often frustrates people to find out that upstream has fixed the bug in a new release, but Ubuntu won't carry it as it's too late.
On the other hand, several people are paid to make Linux efficient for very large scale computers. It doesn't take many customers of million dollar computers to justify one kernel hacker's salary. If you have a thousand core system running linux, being able to upgrade the kernel without a reboot means a lot of saved CPU time.
I listened to a lecture given by Eben Moglen on the subject of Patents, with emphasis on software. When he came to the subject of undoing the current system, he mentioned what I thought to be an interesting market approach. Professional engineers are currently expected to develop patents on their creations, and are rewarded for their efforts with bonuses, promotions, etc. The suggestion is to ask companies to reward engineers for every patent they overturn in the process of bringing a product to market. With that, there's now equal market incentive to overturn patents as there is to create new ones, and a company is forced to recognize that the patent system that gives them IP is as a whole a barrier to their future.
I'm sure there's holes in that simple proposal, but I'd be interested to see if that idea can be made workable.
His regime was certainly pretty good at what it did, if you view war as a legitimate means. Not exactly a friendly or moral group, but fairly successful. It took some pretty heavy casualties to stop the Nazi war machine. Under the full weight of time, he's probably harmed Germany in irreparable ways by refusing to surrender a lost war, etc.
If you want to refute the point that Hitler was a patriot, simply recall that had he surrendered earlier, it would have saved Germany a lot of casualties and perhaps even land. Since a surrender would have likely placed him out of power, he placed his own power over the interests of Germany, an unpatriotic move.
That's because initrd stores a number of scripts related to startup. Obviously there's no bug to report here as you've deployed your own software here. In the past the scripts have fucked up the UUIDs, in my case totally breaking handy things like swap. And resume. I think I've been able to ferret out a least a couple of those.
And a GUI for configuring all the features supported by Synaptic touchpad drivers (that already are in kernel)? The mouse preferences Dialog in GNOME has a few touchpad options: tap to click, and vertical and horizontal scrolling. It doesn't handle super 10 button mice, or tablets. But I'm reasonably comfortable with the defaults. gSynaptics might have more touchpad stuff, but it's in universe and not installed by default like the mouse preferences dialog is. You'll probably have to write the Super button configurator yourself or perhaps bribe others with mice to get the other half done. =(
And does it have a GUI for configuring xrandr defaults on X startup, so that users (with compatible drivers,of course) can easily set multi-monitor setups (that have full 3D acceleration support, unlike with Xinerama)? If I understand the question, then yes.
Good news to everyone then -- Xorg in hardy comes without an xorg.conf by default. I still need it for some crazy configuration options, but I'm significantly removed from the mainstream Ubuntu user. I expect it will only get better over time, as xrandr1.2 matures and the output and input drivers finish up their ends.
A video codec is different than a programmable canvas plugin to a web browser. I'm not sure it's easier, and even if you did have an open flash competitor, have you really won anything significant? Dumb programmers can still make CPU hungry modules, and ads are still in your video. DOH!
Well it depends on how committed the general counsel is to the survival of Monster, and how intelligent Monster's owner is. Monster's a privately held company it seems, so you might be able to provoke a single influential person to commit to a courtroom date. But either way, the next guy's less likely to cave now, and Monster should be less likely to send these letters in response. A courtroom victory against Blue Jeans would vindicate Monster's entire patent threat campaign, but Monster would probably choose to hit someone a bit less prepared to handle the suit instead and may muddy the waters suitably for the next letter writing campaign.
Slashdot Mirror
I'm pretty sure the Wal-Mart greeter was his idea.
This all lies on the assumption that he would, if denied ultimate ownership of the copyright. Recall that Namesys was making a bit of money using that copyright, according to their site.
We invent assumptions, aka axioms, and discover interesting consequences of those assumptions. Those assumptions are often born of our own interaction with the world, such as the idea that the shortest distance between two points is a line, or that two parallel lines never intersect. Whole new mathematical realms are discovered when we relax these rules, but they're not necessarily out there in nature. And there's certainly no rule that nature must follow these rules, everywhere or anywhere.
Like I said, you don't have to trust me, but if you don't trust the people on the committee that is planning to fix it, they've written papers about what's wrong. They do represent the rest of the computer industry, after all. But hey, it's a lot easier to sell what you've got working today in your office and ignore the random odd bugs that you can't duplicate than it is to own up to a minor breakage in assumptions about multithreaded viability, so I can see where you're coming from.
Well yea you could have skipped it, there wasn't enough time to actually change anything. They just tried to shove compiz by default really hard, and introduce upstart. A few updates and gnome, but no time to do anything remarkable.
This isn't something I've just looked up to spite you. It's come up recently on the websites I read, as GCC recently bit the Linux kernel. I've studied proof techniques for semaphores, locks, monitors and so on. This is the sort of stuff that interests me. If memory ordering and non-atomic writes don't make sense to you, then please watch these and maybe read some of the papers. You don't have to understand me, but if you can't understand them then maybe it's time to stop defending the viability of multithreading today. And probably, you should stop using it.
No, it is fundamentally broken in places. It's not a matter of library support. It's a matter of the compiler taking those libraries and totally fucking with their intended purpose because the language spec lets them and it can improve optimizations.
But I'm glad idiots have opinions. It gives people like this guy (video) something to do, by educating you. What ends up happening is that compilers end up presenting non-standard extensions. I've used an embedded language nesC that would provide atomic blocks and implement them by disabling interrupts (not a great idea when avoidable, but passable if you know the drawbacks).
Well, his complaint is that the language isn't a good low level one, on the grounds that bits matter and until C99 was approved and implemented, was a reasonable assumption. Reasons you might want to know the size of datatypes are easy to come up with: you're defining a network protocol or a file storage format. If you use int and compile for x86 and amd64, for example, you could wind up with two incompatible memory layouts. This is the sort of reason we came up with stdint.h in the first place.
typedef unsigned int uint32_t; Limits describes various limits on the system and types.
Ah, well there's no use trying to use C++ as an embedded language / assembler. One look at template inheritance and you shall melt like the Nazi's in Indiana Jones.
At the moment, the C language standard is simply broken for multithreading. There's good reasons for it too. Imaging arithmetic on a 32bit int on a platform of 16bit sized words. There's apparently plans to deal with this in C++, and then C will follow on the assumption that C++ is harder to solve, and C will basically try to be as compatible as possible with their complex needs.
Java is actually in a place to handle this problem quite perfectly. They recognized from the start that the memory model needed work and went about defining a suitable language for multithreading. It's built into the language itself, like garbage collection is. This is why it's popular among enterprise developments. With smart people writing multithreaded code (which we lack in any langauge), you can get some pretty nice results and be fairly efficient, despite Java's renown for being slow.
The difference is that people are inventing new hammers daily, with no sign of slowing down, and using them correctly and efficiently can take years of investment. Carpentry is a very old profession, so there's been plenty of time to identify the right tools. But today, with a historically large population, new tools are invented all the time in this profession in it's infancy. These sorts of popularity measurements can justify an institution using a new language over an older one. "See, plenty of people use it. It's not something we'll have to hire a PhD to use and fix!"
Because the law doesn't prohibit genetic testing -- it prohibits companies from acting on such information. You as a consumer of medical goods (and perhaps medical insurance) can still take such tests and buy disproportionate amounts of insurance.
The far bigger problem that I see in insurance today is the near impossibility of comparing two coverages or even understanding the one you might have. The emergence of standard contracts is great for free markets -- it means that you can price shop around easily because the goods are interchangable, and making information available to consumers can reduce prices. (Apparently term life insurance rates dropped magically around the time companies started hosting websites capable of advertising their rates). But neither you nor I know whether insurance will accept your medical claims. When my brother's leg was terribly broken during a football game, the insurance company denied the ambulance ride as it was not "pre-approved". Apparently this is a common practice, though terrible and the insured are almost never aware of this.
A backport may be likely. The situation for abiword is odd. Their last release was some time ago, and their last major release was in 2005. Additionally, you linked to the 2.6.0 release notes, when 2.6.2 is already out twenty days later. Debian still doesn't have 2.6.x in unstable, so that's an added source of tension. All this is slowly adding up to the typical Open Source poor release engineering. Which isn't bad; just not necessarily LTS material. A long term support distribution has to basically do it's own release engineering in the absence of it upstream.
Ubuntu does work closely with some software groups to track beta / development releases if they have faith in the upstream schedule and so on. If you want software released days or hours before release in a specific version of Ubuntu, it's best to email ubuntu-devel@lists.ubuntu.com and start a conversation, so they know you're a generally reliable upstream committed to quality. It also requires that you have a plan to release far sooner than two weeks notice.
As far as backports go, it depends on how well an enterprising Ryan Pavlik gets with his work. If you're concerned with this problem at all, you should probably be testing his ppa software and make sure it gets a proper backport request.
Actually, 6.06 was an abnormality a result of some terrible asian languages support, and being the first Ubuntu LTS ever. Canonical has made a vow to stick by the launch date schedule ever since. Instead of releasing whenever a feature is done, they include a feature in a given release depending on whether it's done or not. Obviously this is a challenge when bugs arise, and it often frustrates people to find out that upstream has fixed the bug in a new release, but Ubuntu won't carry it as it's too late.
On the other hand, several people are paid to make Linux efficient for very large scale computers. It doesn't take many customers of million dollar computers to justify one kernel hacker's salary. If you have a thousand core system running linux, being able to upgrade the kernel without a reboot means a lot of saved CPU time.
I listened to a lecture given by Eben Moglen on the subject of Patents, with emphasis on software. When he came to the subject of undoing the current system, he mentioned what I thought to be an interesting market approach. Professional engineers are currently expected to develop patents on their creations, and are rewarded for their efforts with bonuses, promotions, etc. The suggestion is to ask companies to reward engineers for every patent they overturn in the process of bringing a product to market. With that, there's now equal market incentive to overturn patents as there is to create new ones, and a company is forced to recognize that the patent system that gives them IP is as a whole a barrier to their future.
I'm sure there's holes in that simple proposal, but I'd be interested to see if that idea can be made workable.
Does this mean that an ISP that strips virus's from websites can be stopped by copyright?
His regime was certainly pretty good at what it did, if you view war as a legitimate means. Not exactly a friendly or moral group, but fairly successful. It took some pretty heavy casualties to stop the Nazi war machine. Under the full weight of time, he's probably harmed Germany in irreparable ways by refusing to surrender a lost war, etc.
If you want to refute the point that Hitler was a patriot, simply recall that had he surrendered earlier, it would have saved Germany a lot of casualties and perhaps even land. Since a surrender would have likely placed him out of power, he placed his own power over the interests of Germany, an unpatriotic move.
That's because initrd stores a number of scripts related to startup. Obviously there's no bug to report here as you've deployed your own software here. In the past the scripts have fucked up the UUIDs, in my case totally breaking handy things like swap. And resume. I think I've been able to ferret out a least a couple of those.
Good news to everyone then -- Xorg in hardy comes without an xorg.conf by default. I still need it for some crazy configuration options, but I'm significantly removed from the mainstream Ubuntu user. I expect it will only get better over time, as xrandr1.2 matures and the output and input drivers finish up their ends.
UUIDs were supposed to fix this. Is this a different problem? Is there a bug report where I can read more about this?
A video codec is different than a programmable canvas plugin to a web browser. I'm not sure it's easier, and even if you did have an open flash competitor, have you really won anything significant? Dumb programmers can still make CPU hungry modules, and ads are still in your video. DOH!
Well it depends on how committed the general counsel is to the survival of Monster, and how intelligent Monster's owner is. Monster's a privately held company it seems, so you might be able to provoke a single influential person to commit to a courtroom date. But either way, the next guy's less likely to cave now, and Monster should be less likely to send these letters in response. A courtroom victory against Blue Jeans would vindicate Monster's entire patent threat campaign, but Monster would probably choose to hit someone a bit less prepared to handle the suit instead and may muddy the waters suitably for the next letter writing campaign.