You're not the first person to think of such a thing. Problem is: it's pretty risky.
Most high end RAID controllers do this already, if you set to write-back. But they also have big batteries attached to them, and even then, you have something like 24 hours to power back on or total system corruption can occur. This means that mentioned systems must be affirmatively managed.
Can you imagine what a hassle this would be for the HD makers, particularly in the notebook use case? It would be a never ending chain of angry users blaming the HD maker for their data loss...
I think the right place to do this is way up in the OS, with a file system that is aware of the issues of small page commits to these devices, and therefore doing some kind of page-coalescence thing. Sun's ZFS can do this. Now we just need something over in consumer space.
Except that it often doesn't work that way. Sometimes it's a notice posted up on their website (often in a not-obvious area). Sometimes it's a hidden clause buried in a mound of other useless paperwork. And in any of the above cases, there is no verification process using written or even verbal acceptance.
I am aware.
Frankly, in most contract of adhesion, the buyer isn't aware of the terms during the early rounds, no less the later ones. I'm against the whole concept. If the thing being contracted isn't worth having a sales representative go over every term of the contract with you, having you initialing you understanding, THEN THERE SHOULD BE NO CONTRACT.
Sadly, it's because the bank has much more resources than the individual.
Hmm. I don't think so. I recall reading at one time that if an unknown third party successfully impersonates you at a bank, the law is on the side of the bank. That's the bad news.
The good news is that as a matter of reduction to practice, banks will quietly fix these problems for you just to avoid the bad publicity.
C//
p.s. these laws are one of the reasons its safer to use credit cards and not debit cards on the net...
Next to be tested in court, the clause stating "we reserve the right to change the terms of this contract at any time without notice".
They sort of do have the legal right to do that. They give a new agreement out, and you, should you continue with service and pay, have given them consideration and voila. New contract of adhesion.
It's just another line item in why I disagree with the very concept of the contract of adhesion. No contract should be permissable that doesn't obey the rules of ordinary contracts ("meeting of the minds," etc).
Business wants the ability to enter into contracts without going through the due dilligence. This is, and always has been, ethically lopsided. For the consumers, there is no real fairness.
We don't really know that. Granted, their past performance is not encouraging... however, have you seen those stock market waivers they like new investors to sign? "Past performance is not an indicator of future outcomes." Words for the wise.
The thing one really ought to pay attention to, however, is staff movements. The Larrabee team: how many talented engineers have defected from Nvidia or ATI to work it? Lots, and look out. Too few, and it will have a long row to ho.
It implements very simple vector primitives. They're really quite easy, insofar as SIMD is for you. Of course, SIMD is the general use case for these things, so hopefully that's why you'd use them. If you'd like to do the MIMD thing ala CUDA (which is of course possible), you will indeed find it to be quite a pain.
Just so you know, I think the perception is real. And I can very much see what you say happening, but also believe that anyone with an H1B no matter their seniority is quite likely to get shafted due to the captive audience effect.
It's just the usual fear and uncertainty about how things outside one's scope of control can change one's life. There is both truth and exaggeration to this, and as usual the truth lies in the middle.
To imagine the exaggeration: suppose that there were 1,000,000 IT workers in the US and one day they decided to let in 1,000,000 H1-B visa holders (and further suppose there were that many to let in). This would cause massive perturbation to the labor market for IT.
Anyway, to understand the stresses more pragmatically, the current system has rules put into effect by agreement with the People that "all H1-B visa holders are to be paid the prevailing wage". There is perception--real or not--that the government here is not enforcing this requirement.
I.e., since the government has a perceived lack of interest in protecting the worker, there is suspicion amongst those who have that perception.
Mod "+1, Right and Righteous" to the AC, who really should have posted as a real person, because being known for a sensible opinion even if others may punish you for it is no weakness.
And the elephant in the room here is that visas are irrelevant in this case. I can't think of a job that can be more easily offshored than computer programming. If you tightly restrict immigration of programmers into the U.S., they'll all set up shop in their home countries, where they can charge even less due to lower cost of living.
Yes. There's truth to this. There's also truth to the notion that there is a cost of outsourcing above and beyond what is paid, purely, in wages for the operation. There is a differential in both quality and strongly managed outcomes (the latter being largely the direct cause of the former). And while the places where quality has been an issue have and are continuing to remedy them rapidly, when they do this the result is that their wages improve remarkably, granting them a more equal basis in raw cost. Since the "they're a remote and therefore disconnected part of our management chain" effect will always be there, these phenomenon can be significant indeed.
As for opening up the borders fully, you must understand that any economy needs some predictability to the outcome of its "capital" investments, particularly technical knowledge capital. If workers cannot predictably choose careers without some level of assurance that market volatilty won't decimate their degrees even as they earn them, they're will be consequences to pay.
These workers may become reluctant to invest their precious "time capital" just after the same fashion that capitalists can be reluctant to invest heavily in actual markets where there is uncertainty associated with ROI.
One could argue that perhaps workers should invest in more general degrees that have plural types of use. That's fine, but do understand that this can discourage heavy investment into specialties.
Actually its quite a common policy in MegaCorps to reject software that require machine specific or expiring license keys for use in "Mission Critical" applications.
--------------
Yes. I usually tell our enterprise software sales people: 1) how much we intend to pay for their software, 2) the manner in which we will accept licensing.
If they won't play ball on either front, usually I'm looking somewhere else.
Doesn't always work, though. Take VMWare, for instance. _Bastards_.:-)
What we should do is put a floor on the price of a barrel of oil. Say $90/barrel.
That floor is less than the current price of oil.
So why do it, you ask?
As it turns out there have been great fluctuations in the past. Alternative energy takes a signfiicant amount of tooling and capitalization, and past investors have done this, only to be destroyed by precipitous drops in the price of oil. By simply guaranteeing a minimum, risk is removed, and investors can do what makes sense to do right now, but are simply not doing for fear alone.
Hmmmmm. Well. Any pure form of libertarianism ought to, based on fundamental tenants, permit workers to "unionize," however no pure form of libertarianism would require employers to do business with said unions. That I can think of.
That said, I'm a sort of Libertarian of the Cato Institute variety. I.e., if your belief that we ought to have various kinds of "libertarian" policies is one based on your interpretation of facts that you think would make the economy better, there is certainly room for some government nannyism here and there, if you think the facts are that said nannyisms make the economy better sometimes, even if less nannyisms make the economy better most of the time.
This isn't very pure Libertarianism at all, however. Pure Libertarianism is structured around moral positioning, is mostly a creed of black and white, and brooks little by way of negotiation or compromise.
They are also exempt if they are paid a certain (large) minimum amount. That's going to happen with a call center worker in a pig's eye, though. I think it's like $75K and up or some such.
And now that both VMware Server and ESXi are available free, seems like VMware Server just became the red-headed stepchild.
Not really. VMWare Server runs with a host OS already installed. So if you have Windows on your desktop, you can install VMWare Server as an ordinary program. Not true of ESX, or ESXi.
Alas, am aware of all that. It's ridiculous on its face, though, as it was with inarguable obviousness not even an inkling of the intent of those who put together our government to begin with.
It would be as if to say that our then-founders, presently suspicious of the sorts of powers that unrestricted governments wield... English Monarchy dontcha know... would insert words into the Constitution meaning that "well, we just wrote all that limiting stuff, but if you can squeeze your idea even thinly into 'commerce' or 'general welfare' you can do just anything."
Preposterous. The very idea boggles.
Sadly, we don't really have a Constitutional government today in any real sense.
Federal law on this point preempts (trumps, overrules) state law on this point, thanks to the federal constitution.
What part of "enumerated powers" as well as the 10th amendment (sort of a "we really meant it!" amendment) do you not understand? Why not actually try reading your Constitution. Powers are only assigned to the Federal government by enumeration in the Constitution (expressly), otherwise powers are held by the States or the People. See the 10th.
Declare and initialize objects as close as possible to where they are used....which has virtual, as does putting them all in one place, where one knows to look for them. Or, thinking of it another way:
This part of the method is for allocating resources, this part for logic.
Draw your line from the closing brace up to the first line with any text on it, that line is the start of your block.
I have a mundanely mechanistic complaint with that.
When reviewing someone else's overly long function, I will sometimes hover over the trailing brace in vi, then hit the "%" key. This will take me to the opening brace, but pan to the right. It jars, and doesn't help.
Supposing I knew a keystroke that took me to the beginning of the statement that opened the curly brace......but I don't.
On the subject of editors, I've often thought that an editor that knew all my blocks and did, on the far left, in a very very light color, a nesting diagram would be very useful. Alas... old and trapped in my ways, I still use vi(m).
I think there are issues with the AMD-Intel x86 license agreement that make it difficult to impossible for a buyer of AMD to also have those licenses. Well, so I've heard it said.
The real question is how likely this will be repeated for the next product.
Intel was almost ignoring their x86 product line in terms of strategic development. They have subsequently focused in on this with laser intensity. You can count on them not misexecuting for a while. The entire "tick, tock" strategy could put AMD out of business entirely if Intel ratchets it down hard enough.
They will probably deliberately let up if it looks like it will really do that, though.
Honestly, I think it's several small, but strategic problems that very quickly led to a very bad place.
First, Intel was/misexecuting/, throwing good money after bad on Itanium. Now they're not.
You could end the story there, almost.
But let's look at AMD's errors:
1. Dumb error not creating a strap-on 4 core to tide over their business. They could have done this faster and better than Intel did, but insisted on theoretical purism over market pragmatics. Dumb. Dumb, dumb, dumb.
1a. But they're showing signs that they are more driven by pragmatism now. Question is, how hard is Intel really driving their own efforts? "Tick, tock" could almost put AMD out of business, and it won't really be AMD's fault.
2. Indigestion and distraction over the ATI acquisition.
Really, I think this superficial analysis is about as deep as one needs to go.
Slashdot Mirror
You're not the first person to think of such a thing. Problem is: it's pretty risky.
Most high end RAID controllers do this already, if you set to write-back. But they also have big batteries attached to them, and even then, you have something like 24 hours to power back on or total system corruption can occur. This means that mentioned systems must be affirmatively managed.
Can you imagine what a hassle this would be for the HD makers, particularly in the notebook use case? It would be a never ending chain of angry users blaming the HD maker for their data loss...
I think the right place to do this is way up in the OS, with a file system that is aware of the issues of small page commits to these devices, and therefore doing some kind of page-coalescence thing. Sun's ZFS can do this. Now we just need something over in consumer space.
C//
Except that it often doesn't work that way. Sometimes it's a notice posted up on their website (often in a not-obvious area). Sometimes it's a hidden clause buried in a mound of other useless paperwork. And in any of the above cases, there is no verification process using written or even verbal acceptance.
I am aware.
Frankly, in most contract of adhesion, the buyer isn't aware of the terms during the early rounds, no less the later ones. I'm against the whole concept. If the thing being contracted isn't worth having a sales representative go over every term of the contract with you, having you initialing you understanding, THEN THERE SHOULD BE NO CONTRACT.
C//
Sadly, it's because the bank has much more resources than the individual.
Hmm. I don't think so. I recall reading at one time that if an unknown third party successfully impersonates you at a bank, the law is on the side of the bank. That's the bad news.
The good news is that as a matter of reduction to practice, banks will quietly fix these problems for you just to avoid the bad publicity.
C//
p.s. these laws are one of the reasons its safer to use credit cards and not debit cards on the net...
Next to be tested in court, the clause stating "we reserve the right to change the terms of this contract at any time without notice".
They sort of do have the legal right to do that. They give a new agreement out, and you, should you continue with service and pay, have given them consideration and voila. New contract of adhesion.
It's just another line item in why I disagree with the very concept of the contract of adhesion. No contract should be permissable that doesn't obey the rules of ordinary contracts ("meeting of the minds," etc).
Business wants the ability to enter into contracts without going through the due dilligence. This is, and always has been, ethically lopsided. For the consumers, there is no real fairness.
C//
Intel's GPU tech is terrible.
We don't really know that. Granted, their past performance is not encouraging... however, have you seen those stock market waivers they like new investors to sign? "Past performance is not an indicator of future outcomes." Words for the wise.
The thing one really ought to pay attention to, however, is staff movements. The Larrabee team: how many talented engineers have defected from Nvidia or ATI to work it? Lots, and look out. Too few, and it will have a long row to ho.
C//
It implements very simple vector primitives. They're really quite easy, insofar as SIMD is for you. Of course, SIMD is the general use case for these things, so hopefully that's why you'd use them. If you'd like to do the MIMD thing ala CUDA (which is of course possible), you will indeed find it to be quite a pain.
C//
If you think GPGPU is hard to program, you are using the wrong tools. Check out RapidMind. It's very easy.
C//
Just so you know, I think the perception is real. And I can very much see what you say happening, but also believe that anyone with an H1B no matter their seniority is quite likely to get shafted due to the captive audience effect.
C//
It's just the usual fear and uncertainty about how things outside one's scope of control can change one's life. There is both truth and exaggeration to this, and as usual the truth lies in the middle.
To imagine the exaggeration: suppose that there were 1,000,000 IT workers in the US and one day they decided to let in 1,000,000 H1-B visa holders (and further suppose there were that many to let in). This would cause massive perturbation to the labor market for IT.
Anyway, to understand the stresses more pragmatically, the current system has rules put into effect by agreement with the People that "all H1-B visa holders are to be paid the prevailing wage". There is perception--real or not--that the government here is not enforcing this requirement.
I.e., since the government has a perceived lack of interest in protecting the worker, there is suspicion amongst those who have that perception.
C//
Mod "+1, Right and Righteous" to the AC, who really should have posted as a real person, because being known for a sensible opinion even if others may punish you for it is no weakness.
C//
And the elephant in the room here is that visas are irrelevant in this case. I can't think of a job that can be more easily offshored than computer programming. If you tightly restrict immigration of programmers into the U.S., they'll all set up shop in their home countries, where they can charge even less due to lower cost of living.
Yes. There's truth to this. There's also truth to the notion that there is a cost of outsourcing above and beyond what is paid, purely, in wages for the operation. There is a differential in both quality and strongly managed outcomes (the latter being largely the direct cause of the former). And while the places where quality has been an issue have and are continuing to remedy them rapidly, when they do this the result is that their wages improve remarkably, granting them a more equal basis in raw cost. Since the "they're a remote and therefore disconnected part of our management chain" effect will always be there, these phenomenon can be significant indeed.
As for opening up the borders fully, you must understand that any economy needs some predictability to the outcome of its "capital" investments, particularly technical knowledge capital. If workers cannot predictably choose careers without some level of assurance that market volatilty won't decimate their degrees even as they earn them, they're will be consequences to pay.
These workers may become reluctant to invest their precious "time capital" just after the same fashion that capitalists can be reluctant to invest heavily in actual markets where there is uncertainty associated with ROI.
One could argue that perhaps workers should invest in more general degrees that have plural types of use. That's fine, but do understand that this can discourage heavy investment into specialties.
C//
Actually its quite a common policy in MegaCorps to reject software that require machine specific or expiring license keys for use in "Mission Critical" applications.
--------------
Yes. I usually tell our enterprise software sales people: 1) how much we intend to pay for their software, 2) the manner in which we will accept licensing.
If they won't play ball on either front, usually I'm looking somewhere else.
Doesn't always work, though. Take VMWare, for instance. _Bastards_. :-)
What we should do is put a floor on the price of a barrel of oil. Say $90/barrel.
That floor is less than the current price of oil.
So why do it, you ask?
As it turns out there have been great fluctuations in the past. Alternative energy takes a signfiicant amount of tooling and capitalization, and past investors have done this, only to be destroyed by precipitous drops in the price of oil. By simply guaranteeing a minimum, risk is removed, and investors can do what makes sense to do right now, but are simply not doing for fear alone.
C//
Hmmmmm. Well. Any pure form of libertarianism ought to, based on fundamental tenants, permit workers to "unionize," however no pure form of libertarianism would require employers to do business with said unions. That I can think of.
That said, I'm a sort of Libertarian of the Cato Institute variety. I.e., if your belief that we ought to have various kinds of "libertarian" policies is one based on your interpretation of facts that you think would make the economy better, there is certainly room for some government nannyism here and there, if you think the facts are that said nannyisms make the economy better sometimes, even if less nannyisms make the economy better most of the time.
This isn't very pure Libertarianism at all, however. Pure Libertarianism is structured around moral positioning, is mostly a creed of black and white, and brooks little by way of negotiation or compromise.
C//
I thought we were supposed to pick it up and talk to it:
*holding mouse to mouth*
"Computer...."
"Computer...."
*moment of realization*
"Oh, keyboard! How quaint!"
They are also exempt if they are paid a certain (large) minimum amount. That's going to happen with a call center worker in a pig's eye, though. I think it's like $75K and up or some such.
C//
And now that both VMware Server and ESXi are available free, seems like VMware Server just became the red-headed stepchild.
Not really. VMWare Server runs with a host OS already installed. So if you have Windows on your desktop, you can install VMWare Server as an ordinary program. Not true of ESX, or ESXi.
C//
Alas, am aware of all that. It's ridiculous on its face, though, as it was with inarguable obviousness not even an inkling of the intent of those who put together our government to begin with.
It would be as if to say that our then-founders, presently suspicious of the sorts of powers that unrestricted governments wield... English Monarchy dontcha know... would insert words into the Constitution meaning that "well, we just wrote all that limiting stuff, but if you can squeeze your idea even thinly into 'commerce' or 'general welfare' you can do just anything."
Preposterous. The very idea boggles.
Sadly, we don't really have a Constitutional government today in any real sense.
Amendments are no longer needed.
C//
Federal law on this point preempts (trumps, overrules) state law on this point, thanks to the federal constitution.
What part of "enumerated powers" as well as the 10th amendment (sort of a "we really meant it!" amendment) do you not understand? Why not actually try reading your Constitution. Powers are only assigned to the Federal government by enumeration in the Constitution (expressly), otherwise powers are held by the States or the People. See the 10th.
C//
I can think of a few where most of the game logic was even TCL,...
I think I want to commit suicide now.
C//
Declare and initialize objects as close as possible to where they are used. ...which has virtual, as does putting them all in one place, where one knows to look for them. Or, thinking of it another way:
This part of the method is for allocating resources, this part for logic.
C//
Draw your line from the closing brace up to the first line with any text on it, that line is the start of your block.
I have a mundanely mechanistic complaint with that.
When reviewing someone else's overly long function, I will sometimes hover over the trailing brace in vi, then hit the "%" key. This will take me to the opening brace, but pan to the right. It jars, and doesn't help.
Supposing I knew a keystroke that took me to the beginning of the statement that opened the curly brace... ...but I don't.
On the subject of editors, I've often thought that an editor that knew all my blocks and did, on the far left, in a very very light color, a nesting diagram would be very useful. Alas... old and trapped in my ways, I still use vi(m).
C//
I think there are issues with the AMD-Intel x86 license agreement that make it difficult to impossible for a buyer of AMD to also have those licenses. Well, so I've heard it said.
C//
The real question is how likely this will be repeated for the next product.
Intel was almost ignoring their x86 product line in terms of strategic development. They have subsequently focused in on this with laser intensity. You can count on them not misexecuting for a while. The entire "tick, tock" strategy could put AMD out of business entirely if Intel ratchets it down hard enough.
They will probably deliberately let up if it looks like it will really do that, though.
C//
Honestly, I think it's several small, but strategic problems that very quickly led to a very bad place.
First, Intel was /misexecuting/, throwing good money after bad on Itanium. Now they're not.
You could end the story there, almost.
But let's look at AMD's errors:
1. Dumb error not creating a strap-on 4 core to tide over their business. They could have done this faster and better than Intel did, but insisted on theoretical purism over market pragmatics. Dumb. Dumb, dumb, dumb.
1a. But they're showing signs that they are more driven by pragmatism now. Question is, how hard is Intel really driving their own efforts? "Tick, tock" could almost put AMD out of business, and it won't really be AMD's fault.
2. Indigestion and distraction over the ATI acquisition.
Really, I think this superficial analysis is about as deep as one needs to go.
Feel free to comment. :-)
C//