Most importantly: what contract/guarantee do we have that our interests are being served?
You leave. If the personality of the site changes to something you don't like, then go somewhere that you do like.
It's really that simple. Slashdot owes you nothing. I'm picking on you, but I could say this to any number of these sort of posts. It's not your site. If they want to turn this into a fanboy site for Microsoft then they will, and they have the perfect right.
And just to head off the silly argument that "CmdrTaco owes me, because he got an IPO off the backs of his readers, wah wah wah", that's absolutely false. It was a two-way transaction. CmdrTaco and everyone worked hard to provide you a site for which you received value, and they received the aggregate fruits of that labor. You've already been paid.
I wish they would use Oracle, and then maybe they wouldn't have to flush all the past stories and comments (i.e., the whole database would be searchable). I would imagine the reason they do that is that MySQL is not known for scaling up to large databases.
Unfortunately, Slashdot probably feels they need to use Open Source no matter what, even if that means using inferior software. That's the big problem with embracing a religion, rather than embracing "the best".
I was all set to rip this guy (NOT a typical user, crediblity shot as soon as he said any WP under Linux is comparable to Office, etc, etc), but Slashdotters have beat me to it. Bravo!
And just when I was just about convinced that any positive article about Linux would get a huge pass by the majority of the Slashdot population.
You're not reading what I said. I didn't say there was anything wrong with cloning. I said there was something wrong with parents who would clone a dead child, as if that would cause that child to somehow be "reborn".
Put it this way... if you died, what would you think if your parents decided to take a DNA sample and just clone a new copy of you? And then just go on as if you had never existed, because after all, they have a brand-new one?
That is what they're saying with this "insurance" policy. If your child dies, heck, just create a new one! Good as new. Of course, the child isn't the same, and I find it creepy that there might be sick parents who would think the child was the same.
Stop "lying"? Look up lying in the dictionary. Apparently you don't know what it means.
Oh, OK. In my endless quest to educate the unwashed like yourself, let me help you out. "Lying" is the intentional attempt to mislead another person by putting forth knowingly untrue statements. In my case, assuming I don't know what I'm talking about, I would be "ignorant", since I'm arguing from an incorrect set of facts.
So the proper sentence would be "We'll be 'gracious' when you stop being ignorant". But even that has problems. First of all: "We'll". You look foolish when you presume to speak for others. It would be better to say "I'll be 'gracious' when you stop being ignorant."
Better. But still not correct. The statement is rather contradictory, since the time to be gracious is when educating somebody out of ignorance. But I don't think I can fix that personality problem of yours in this post. Maybe next time.
Other than if I grant that GC could utilize stack allocation in some cases in the second statement, what is wrong with my first statement? GC will be slower than manual allocation, because there is overhead to doing "something" (tracking references) rather than doing "nothing" (not having to track references).
The burden of proof is on you to demonstrate how a method of GC can be as fast or faster than manual allocation.
Incidently, I'll even grant to you the fact that the reason Smalltalk and Lisp are slower relative to C is not completely GC, but just the nature of the languages. Still, that does not prove that GC is as fast as manual memory management.
The other thing is that I feel I'm getting a bad rap here. I'm not against GC in languages, but there are cases where you don't want the unpredictability of GC and manual memory management makes sense. Are you trying to say that GC is appropriate for all problems?
This is what I love about many of you people: You're so classy and gracious.
Unlike you, I was not born knowing all human knowledge, and I even occasionally like to learn something. My ego is not so fragile that I can't admit when I learn something.
My "pontification" on the performance is based on actual experience. There are a lot more factors than just performance that come into play when using a GC language. I find it interesting that there are so potential solutions to the historical performance problems, but this does not mean that it's appropriate for all problems.
I might also point out that I'm am FAR from convinced that using Lisp or Smalltalk (which have notoriously bad performance compared to lightweight languages such as C) as an example is proof of anything.
The fact that you think these problems are so simple that you just "RTFM" instead of actually using your brain and experience tells me you have very little of either.
CLONAID® will also offer a service called INSURACLONE® which, for a $50,000 fee, will provide the sampling and safe storage of cells from a living child or from a beloved person in order to create a clone if the child dies of an incurable disease or through an accident. In the case of a genetic disease, the cells will be preserved until science can genetically repair it before recreating the child (or an adult).
Who would recreate a child who dies in an accident or incurable disease? Someone would have to have severe psychological problems to even consider it.
by "large" I mean conversational English, which is far less than 31,000.
Yes, I realize they claim they can handle more than that, but the rub is in the definition of "handle". My definition says that practical text to speech should be very close to the accuracy level of dictation to a human secretary (without requiring training, I might add). Their definition is -- far from that.
The reason I made the distinction of "large vocabulary" is that we actually have reasonable voice recognition of things like digits, which are very accurate and speaker independant, primarily because there is no context required.
"Common belief"? I've never heard it before, and to be honest, it doesn't really make much sense. Why would a clone universe be created? And even if a universe was "automatically" created from the death of a previous universe, why would it necessarily repeat time in exactly the same way?
I'll take the challenge, and I'll get 100% accuracy, guaranteed (assuming we're talking about normal speech, and not absurd levels of slurring).
It's always been the case that voice recognition works if you speak in a robotic, non-natural way. But I'm talking about speech to text for the masses, who will not (and have not in the past) be willing to learn to speak "the right way". Most are barely willing to go through training sessions. The reason is that for most people, they want to concentrate on what they're saying, rather than how they're saying it (which is not unreasonable).
Not to mention the environmental difficulties of having using a PDA. Will this work in a noisy airport? Doubt it.
If you saw five billion coins laying on the ground, 4 billion heads up and 1 billion heads down, would you conclude that flipping a coin gives 80% heads 20% tails?
No, but the reason is that there is a reasonable possibility of someone placing them in that configuration. In the case of observing life on other planets, I have no evidence that the probability was artificially manipulated, so it's valid to create a theory based on statistical survey.
Is it valid to try and create cosmological theories based on our observations that galaxies tend to "clump"? Of course, even though the theoretical possibility exists that someone went through and clumped them to fool us (ala coins on the ground).
I still have yet to use a voice recognition system that doesn't suck. I keep hearing that "we're almost there! Processors are almost fast enough!". Yeah right.
Here's an experiment that I would love to give any voice recognition company: I speak in my natural voice and read, say, 10 pages out of a novel. You can let your system process it for ONE WEEK. This is the equivalent using a processor 4000 times faster. After that, I would expect 100% accuracy if the limitation is cpu power!.
But you know 100% accuracy won't happen. The reason is that the problem with voice recognition is not lack of computer power, it's 1) lack of "world" knowledge, and 2) lack of understanding of how to apply world knowledge to language parsing (the AI problem, in other words). Even humans have trouble understanding speech, but we fill in the inaccuracies by knowing context.
I am doubtful that we will see true large vocabulary voice recognition (beyond the toy level) in the next 20, if not 50 years.
By the way, this is not to say that the toy level might not be useful for a device like this. But I have a feeling that they will oversell it like every other YAVRF and the public will get frustrated with its imperfections.
That's just foolish. Intel constantly has chips in the pipeline. Does anyone seriously think that the Intel engineers were sitting with their feet up on their desks and then suddenly there was the Transmeta announcement? "Good God! We better start designing a new chip!"
Not to mention that Intel constantly makes new roadmap announcements, so the FUD argument doesn't fly either.
I still might point out that Transmeta has not even shipped a product yet, so I would argue against throwing FUD in glass houses.
Because we don't sense the passage of time while we didn't exist. For all we know, it took the creation and destruction of 1e512 universes of time before intelligent life happened to arise and we were able to think about the fact.
Now, if we happen to discover life on other planets someway, then we would be able to say how statistically probable life is. Until then, "probablistic" arguments are complete nonsense.
Not to mention that it doesn't argue against Darwin anyway. Darwin only poses Natural Selection (I believe) which is an observed, provable fact of biology.
The stresses around the screw would be terrific; perhaps if you slowly ramped up the force the bone could adapt, but it is unlikely to be as strong as something which left the bone intact, like some kind of artificial tendon.
True; it might be better to wrap a strip of metal all the way around the bone and attach to that. It would probably be stronger and less invasive to boot.
You wouldn't have to attach to the bone the same way muscles attach... you could use a screw and drill it all the way in.
As for snapping them, I think bones are built to take a lot more stress than muscles can create. You could probably be significantly stronger before fractures became a serious risk.
After all, how often do body builders break bones? I mean, it happens, but it's pretty darn rare. In fact, muscle tears are far more common than broken bones, and that risk would go away.
A world-class power lifter can probably lift 4-5 times what the average man can lift. That would still be pretty cool, particularly if you could have the strength without all the extra mass.
No, I mean you don't have handles at the implementation level. And the advantage of putting that logic into the compiler is that you only have to get it right once, by the compiler authors. This is sometimes called code reuse.;)
Well, I suppose you could keep a list of all the references to the object, and update all the references when the object moves. The does mean you need all sorts of locking nonsense if you want preemptive multithreading.
I must admit you've given some food for thought about whether a practical GC could be done. I suppose the logical question is why hasn't it been done for a C++ compiler? If ever a language cried out for GC, it's C++.
Common Lisp is very much a practical language...
Well, I've never programmed a phone switch, so I can't say about that one. And I'm sure it may have snuck in as a scripting language in a game or two, and maybe some rogue programmer did something in an operating system.
But the question is whether it should have been. You can't deny that C is by far the most dominant language for doing high-performance system programming (leaving out numerical programming in FORTRAN), and there is a reason for it.
As for AI, don't cite that to me as a "practical" application, or I'll laugh at you. (AI is to Science as "1,2,3,Many" is to Mathematics).
Note that you don't need to use "handles", because in all of these languages there's no such thing as explicit pointers, so it's possible for the memory manager to know where all the pointers are and update them transparently.
Well, OK, maybe you're not using handles at the language level, but so what? There is still all that logic under the hood, and that is a big performance drain.
All the languages you have named (including (notoriously) Java) are not practical for any sort of large-scale application where performance is a concern. You haven't really given any evidence to support that GC is practical for a performance-critical major application.
In fact, I asked if this GC method had ever left the lab, and it could be argued that Lisp and Smalltalk have never left the lab.:)
First of all, if you are having to use reference counts, then you've probably screwed up the design of your memory management (there are cases where they're useful, but not many).
But moving on, it's trivially provable that garbage collection is slower, because you have the additional overhead of tracking references on EVERY allocation, versus manual memory management where you don't (in a properly design system, that is).
But that's not even the full story. The other problem with GC is that suddenly everything is allocated using dynamic allocation, rather than being able to story temporary objects on the stack. Dyamic allocation is WAY slower, and creates significant problems with memory fragmentation.
The other "human" factor I could bring up is that "allocation abuse" happens much more often in languages with GC, because it's so easy to create/destroy and throw around objects/memory willy-nilly. Unfortunately, there is such a buffer zone between the programmer and "what's really going on" that they don't the realize the inefficiency of what they're doing.
Now, this is not to say that I'm always against GC. In many cases, the inefficiency is worth it to gain the advantage of fewer memory allocation bugs. For example, in embedded systems this is a great thing.
If you haven't finished smoking whatever your smoking, perhaps you'd care to give an example of this "magic" garbage collection that is faster than straight malloc/free or new/delete?
Slashdot Mirror
Most importantly: what contract/guarantee do we have that our interests are being served?
You leave. If the personality of the site changes to something you don't like, then go somewhere that you do like.
It's really that simple. Slashdot owes you nothing. I'm picking on you, but I could say this to any number of these sort of posts. It's not your site. If they want to turn this into a fanboy site for Microsoft then they will, and they have the perfect right.
And just to head off the silly argument that "CmdrTaco owes me, because he got an IPO off the backs of his readers, wah wah wah", that's absolutely false. It was a two-way transaction. CmdrTaco and everyone worked hard to provide you a site for which you received value, and they received the aggregate fruits of that labor. You've already been paid.
--
I wish they would use Oracle, and then maybe they wouldn't have to flush all the past stories and comments (i.e., the whole database would be searchable). I would imagine the reason they do that is that MySQL is not known for scaling up to large databases.
Unfortunately, Slashdot probably feels they need to use Open Source no matter what, even if that means using inferior software. That's the big problem with embracing a religion, rather than embracing "the best".
--
I was all set to rip this guy (NOT a typical user, crediblity shot as soon as he said any WP under Linux is comparable to Office, etc, etc), but Slashdotters have beat me to it. Bravo!
And just when I was just about convinced that any positive article about Linux would get a huge pass by the majority of the Slashdot population.
My faith in geekdom is restored. :)
--
You're not reading what I said. I didn't say there was anything wrong with cloning. I said there was something wrong with parents who would clone a dead child, as if that would cause that child to somehow be "reborn".
Put it this way... if you died, what would you think if your parents decided to take a DNA sample and just clone a new copy of you? And then just go on as if you had never existed, because after all, they have a brand-new one?
That is what they're saying with this "insurance" policy. If your child dies, heck, just create a new one! Good as new. Of course, the child isn't the same, and I find it creepy that there might be sick parents who would think the child was the same.
--
Stop "lying"? Look up lying in the dictionary. Apparently you don't know what it means.
Oh, OK. In my endless quest to educate the unwashed like yourself, let me help you out. "Lying" is the intentional attempt to mislead another person by putting forth knowingly untrue statements. In my case, assuming I don't know what I'm talking about, I would be "ignorant", since I'm arguing from an incorrect set of facts.
So the proper sentence would be "We'll be 'gracious' when you stop being ignorant". But even that has problems. First of all: "We'll". You look foolish when you presume to speak for others. It would be better to say "I'll be 'gracious' when you stop being ignorant."
Better. But still not correct. The statement is rather contradictory, since the time to be gracious is when educating somebody out of ignorance. But I don't think I can fix that personality problem of yours in this post. Maybe next time.
--
Other than if I grant that GC could utilize stack allocation in some cases in the second statement, what is wrong with my first statement? GC will be slower than manual allocation, because there is overhead to doing "something" (tracking references) rather than doing "nothing" (not having to track references).
The burden of proof is on you to demonstrate how a method of GC can be as fast or faster than manual allocation.
Incidently, I'll even grant to you the fact that the reason Smalltalk and Lisp are slower relative to C is not completely GC, but just the nature of the languages. Still, that does not prove that GC is as fast as manual memory management.
The other thing is that I feel I'm getting a bad rap here. I'm not against GC in languages, but there are cases where you don't want the unpredictability of GC and manual memory management makes sense. Are you trying to say that GC is appropriate for all problems?
--
This is what I love about many of you people: You're so classy and gracious.
Unlike you, I was not born knowing all human knowledge, and I even occasionally like to learn something. My ego is not so fragile that I can't admit when I learn something.
My "pontification" on the performance is based on actual experience. There are a lot more factors than just performance that come into play when using a GC language. I find it interesting that there are so potential solutions to the historical performance problems, but this does not mean that it's appropriate for all problems.
I might also point out that I'm am FAR from convinced that using Lisp or Smalltalk (which have notoriously bad performance compared to lightweight languages such as C) as an example is proof of anything.
The fact that you think these problems are so simple that you just "RTFM" instead of actually using your brain and experience tells me you have very little of either.
--
Who would recreate a child who dies in an accident or incurable disease? Someone would have to have severe psychological problems to even consider it.
--
by "large" I mean conversational English, which is far less than 31,000.
Yes, I realize they claim they can handle more than that, but the rub is in the definition of "handle". My definition says that practical text to speech should be very close to the accuracy level of dictation to a human secretary (without requiring training, I might add). Their definition is -- far from that.
The reason I made the distinction of "large vocabulary" is that we actually have reasonable voice recognition of things like digits, which are very accurate and speaker independant, primarily because there is no context required.
--
"Common belief"? I've never heard it before, and to be honest, it doesn't really make much sense. Why would a clone universe be created? And even if a universe was "automatically" created from the death of a previous universe, why would it necessarily repeat time in exactly the same way?
--
I'll take the challenge, and I'll get 100% accuracy, guaranteed (assuming we're talking about normal speech, and not absurd levels of slurring).
It's always been the case that voice recognition works if you speak in a robotic, non-natural way. But I'm talking about speech to text for the masses, who will not (and have not in the past) be willing to learn to speak "the right way". Most are barely willing to go through training sessions. The reason is that for most people, they want to concentrate on what they're saying, rather than how they're saying it (which is not unreasonable).
Not to mention the environmental difficulties of having using a PDA. Will this work in a noisy airport? Doubt it.
--
If you saw five billion coins laying on the ground, 4 billion heads up and 1 billion heads down, would you conclude that flipping a coin gives 80% heads 20% tails?
No, but the reason is that there is a reasonable possibility of someone placing them in that configuration. In the case of observing life on other planets, I have no evidence that the probability was artificially manipulated, so it's valid to create a theory based on statistical survey.
Is it valid to try and create cosmological theories based on our observations that galaxies tend to "clump"? Of course, even though the theoretical possibility exists that someone went through and clumped them to fool us (ala coins on the ground).
--
New acronym I just invented: YAVRF.
I still have yet to use a voice recognition system that doesn't suck. I keep hearing that "we're almost there! Processors are almost fast enough!". Yeah right.
Here's an experiment that I would love to give any voice recognition company: I speak in my natural voice and read, say, 10 pages out of a novel. You can let your system process it for ONE WEEK. This is the equivalent using a processor 4000 times faster. After that, I would expect 100% accuracy if the limitation is cpu power!.
But you know 100% accuracy won't happen. The reason is that the problem with voice recognition is not lack of computer power, it's 1) lack of "world" knowledge, and 2) lack of understanding of how to apply world knowledge to language parsing (the AI problem, in other words). Even humans have trouble understanding speech, but we fill in the inaccuracies by knowing context.
I am doubtful that we will see true large vocabulary voice recognition (beyond the toy level) in the next 20, if not 50 years.
By the way, this is not to say that the toy level might not be useful for a device like this. But I have a feeling that they will oversell it like every other YAVRF and the public will get frustrated with its imperfections.
--
That's just foolish. Intel constantly has chips in the pipeline. Does anyone seriously think that the Intel engineers were sitting with their feet up on their desks and then suddenly there was the Transmeta announcement? "Good God! We better start designing a new chip!"
Not to mention that Intel constantly makes new roadmap announcements, so the FUD argument doesn't fly either.
I still might point out that Transmeta has not even shipped a product yet, so I would argue against throwing FUD in glass houses.
--
Uh, that's what I meant. Even then, the grammar kind of sucks.
How about: "The probability of life is irrelevent". Much better.
--
Because we don't sense the passage of time while we didn't exist. For all we know, it took the creation and destruction of 1e512 universes of time before intelligent life happened to arise and we were able to think about the fact.
Now, if we happen to discover life on other planets someway, then we would be able to say how statistically probable life is. Until then, "probablistic" arguments are complete nonsense.
Not to mention that it doesn't argue against Darwin anyway. Darwin only poses Natural Selection (I believe) which is an observed, provable fact of biology.
--
The stresses around the screw would be terrific; perhaps if you slowly ramped up the force the bone could adapt, but it is unlikely to be as strong as something which left the bone intact, like some kind of artificial tendon.
True; it might be better to wrap a strip of metal all the way around the bone and attach to that. It would probably be stronger and less invasive to boot.
--
You wouldn't have to attach to the bone the same way muscles attach... you could use a screw and drill it all the way in.
As for snapping them, I think bones are built to take a lot more stress than muscles can create. You could probably be significantly stronger before fractures became a serious risk.
After all, how often do body builders break bones? I mean, it happens, but it's pretty darn rare. In fact, muscle tears are far more common than broken bones, and that risk would go away.
A world-class power lifter can probably lift 4-5 times what the average man can lift. That would still be pretty cool, particularly if you could have the strength without all the extra mass.
--
Not to mention that you'll still look like a geek.
Except for those rippling muscles of plastic! Yeah, the babes will swoon over those.
--
No, I mean you don't have handles at the implementation level. And the advantage of putting that logic into the compiler is that you only have to get it right once, by the compiler authors. This is sometimes called code reuse. ;)
Well, I suppose you could keep a list of all the references to the object, and update all the references when the object moves. The does mean you need all sorts of locking nonsense if you want preemptive multithreading.
I must admit you've given some food for thought about whether a practical GC could be done. I suppose the logical question is why hasn't it been done for a C++ compiler? If ever a language cried out for GC, it's C++.
Common Lisp is very much a practical language...
Well, I've never programmed a phone switch, so I can't say about that one. And I'm sure it may have snuck in as a scripting language in a game or two, and maybe some rogue programmer did something in an operating system.
But the question is whether it should have been. You can't deny that C is by far the most dominant language for doing high-performance system programming (leaving out numerical programming in FORTRAN), and there is a reason for it.
As for AI, don't cite that to me as a "practical" application, or I'll laugh at you. (AI is to Science as "1,2,3,Many" is to Mathematics).
--
Just out of curiosity, which side are you on? :)
--
Note that you don't need to use "handles", because in all of these languages there's no such thing as explicit pointers, so it's possible for the memory manager to know where all the pointers are and update them transparently.
Well, OK, maybe you're not using handles at the language level, but so what? There is still all that logic under the hood, and that is a big performance drain.
All the languages you have named (including (notoriously) Java) are not practical for any sort of large-scale application where performance is a concern. You haven't really given any evidence to support that GC is practical for a performance-critical major application.
In fact, I asked if this GC method had ever left the lab, and it could be argued that Lisp and Smalltalk have never left the lab. :)
--
It sounds interesting, but where has this been demonstrated in real world implementations?
Since objects can be moved around by the memory manager, you have to have "handles" with double pointers, locking, all kinds of nonsense.
It sounds good in theory, but I would be interested to know if it's ever been used in a real application outside the lab.
--
Samrobb has already said it well, but I'll just add that you don't really define "do worse" in your statement.
If you are talking bug-free code, then clearly GC is a big win. If you're talking fastest code and smallest memory footprint, then GC is a big lose.
--
Equally frank: Bullshit.
First of all, if you are having to use reference counts, then you've probably screwed up the design of your memory management (there are cases where they're useful, but not many).
But moving on, it's trivially provable that garbage collection is slower, because you have the additional overhead of tracking references on EVERY allocation, versus manual memory management where you don't (in a properly design system, that is).
But that's not even the full story. The other problem with GC is that suddenly everything is allocated using dynamic allocation, rather than being able to story temporary objects on the stack. Dyamic allocation is WAY slower, and creates significant problems with memory fragmentation.
The other "human" factor I could bring up is that "allocation abuse" happens much more often in languages with GC, because it's so easy to create/destroy and throw around objects/memory willy-nilly. Unfortunately, there is such a buffer zone between the programmer and "what's really going on" that they don't the realize the inefficiency of what they're doing.
Now, this is not to say that I'm always against GC. In many cases, the inefficiency is worth it to gain the advantage of fewer memory allocation bugs. For example, in embedded systems this is a great thing.
If you haven't finished smoking whatever your smoking, perhaps you'd care to give an example of this "magic" garbage collection that is faster than straight malloc/free or new/delete?
--