Yes, flying is better understood with math, it's a physical problem. I wasn't implying that no explanation at all is better than any model, I was refering to verbal explanation, like when you talk with a best friend, and if he is one of those guys that have a clear understanding of life and you, he'll be able to tell the REAL problem based on the behaveour.
Though the model may be more interesting or elegant...
You are missing something, to have an econometric model the right way, you must first have a teory. if you do not have a theory you can't use econometrics to prove or disprove your theory. Steps for an econometric model:
- Present your theory or thesis - Present the stochastic econmetric model you'll use to test it's validity - Estimate coefficients - Accept or reject your hipotesis (or refine your theory and go pack to the beggining) - Predict
It has been proved that you make relly great but meaningless econometric models. Like average lenght of women dresses and (can't remember well, but it was something funny) the stock market. I think the correlation was significnt at 99% confidence level...:)
Nobody cares about averages here, you are supposed to predict the future, as in the order in which the coins will land as time passes. If you miss the order the results will vary to a great degree.
In other words, you could easily predict the NEXT coin flip (i keep on using the coin flip, but i am thinking on the physchohistory of humankind, so this experiments are NOT radom as in a normal coin flip) with near 100% accuracy. But you cannot predict the 10000 coin flip, because to predict that coin flip you'll need to know the exact flips that came before and in what order.
I think I made a short explanation long, but the point is the original poster does know that statistics work, but NOT so well when future predictions depend on the accuracy of all past predictions. Errors do get accumulated as any time series analist (or any econometrician) will tell you... prediction does lose accuracy as you get further away from "present time"...
I'd say phychohistory would be possible if and only if ALL individual acctions and all physical fenomens could be traced, and we'd be talking just about physical on a grand scale. We'dd be just attoms in deterministic paths...
For any conceibable behaveour there is a mathematical way of fitting the behaveour with a certain degree of probability. If something is not pure noise, then there must be some way to formalize it, though language itself or in mathematical notation.
This works, of course, don't add much value because they never explain how or why things are like that. With physics you don't have to explain the basic laws, they "just are", but with everything else, you better have some explanation of some sort because, in reallity, they are nothing more than constructs based on physical constraints.
On the other side, it might be funny to see how some people could see these formalizations as expressing more or being more accurate than "plain verb" explanation. "If it's hard to understand then it's real science!!" (wrong!)...
Just my thoughts so (I am biased yes, I've seen to many quantitative economics to believe equations express more just because the math is hard...they usually don't).
I think it's a great solution, in the variants you described, or being removable (a connector problem, not an HD problem) or being for buffering or standalone if you need extra storage (say 60 GB). In any case, you could have all three options (one internal fixed, a slot for removable variant and a firewire port & mode that activates buffering.
A plain non-removable big HD consumes a lot and is ugly, and solid state fash cards are really unnecesarily expensive. Of course, who cares, some people will still think they need the solid state solution:)
Tape needs to be located somewher also, and there's no requirement that the HDD array be at the same location that the backed up systems.
Of course, HDD could be hacked more easily because there sure exist a network link to them. Unless it's a read only network (as in nothing can get deleted nor read unless you are inside certain room). This latter can also be enforced.
Better yet, why not use a laptop HD. Price/performance/size/consumption wise it's the best compromise. $150 or less for a 40 GB mini IDE drive. A standard adaptor for USB/PCMCIA/Firewire transfer and we are all set.
Unless you are buying bandwidth. In that case, the analogy would be a cable company selling you some signal. As long as you don't draw more current than needed or unless you are not messing the signal, you are ok.
Re:I'm going to have to agree with him on this one
on
Linus on DRM
·
· Score: 1
I think that's not a nice argument. There are many arguments for allowing DRM in the kernel, but not this one. For instance, Nukes are ok if rhe US or china has access to it, and bad if Saddam has it. Or is it that nukes are just plain bad? You can under for yourself...
Crossover is best server chilled when MSOffice is a requirement. You know, some people use other OS and have everything there. I can't even access my home dir under windows (can't say the opossite), and that is just sad... not to say that vendor support is shitty sometimes (for example, my IBM Thinkpad i-series 1460 freezes every now and then under Windows 2000 for unknown issues).
With Linux I was just lucky or something, because it never EVER hangs, and never gives me trouble (witched to gentoo recently because became fored of baby sitting Linux, so i am even happier now).
The point is it is good to be able to separate OS and applications, and specially if the only thing that's preventing you from trying them is Office.
It is not supported under Linux, so they should have no problems at all. In fact, there are many products that are not supported under several variants of Windows, but there are how-to's for the brave that want to try to make it work despite they will be on their own.
You can get 20 items at $5 that would bring home back $40 rebates. The $100 item will only get you $2. That's some kind of normalization. Or look at is as a percentage...
It comes at 22, and at other places too. The Dell cluster uses a mix of servers, and it's not really a supercomputer (or I got confused at their page?) but a cluster of computer. It's nice to note that the IBM system at stop #5 is dated in 2000 while the rest of the systems are all 2002. So it's still at #5 after two years...nice uh?
If I where in a position to choose who to blame, in one side a 100 billion company, and in the other side, two kids asking for some hardware because they only FTP server went dead, then I'd choose Microsoft.
Large companies don't want the long, sensible answer, they want to quickly be able to point the finger at someone (Microsoft, or whoever allowed OSS to be used) and be gone. That's how they usually deal with problems, and this is nothing that may come to you as new. The details can be worked out afterwards.
I am not saying that line of reasoning is right, but that it's usual, so some people feeding kids think they are more secure beign able to blame a large company (not to be able to get the lost productivity back). This is where IBM, Oracle and some others come to play. OSS image must improve to the point where one can simply state "we are using what the most sucessfull companies use" and carry on to apply the patch that by then is surelly available...
You can't buy specific libraries from Red Hat. You ass is somewhat covered in general, but when you are making specific bets at specific libraries I guess the panic factor (or blamability) grows stronger. Also, there's a lot of people that do not OSS is not about turning the software industry into alstruism or religion, but about something more reasonable in term of efficiency.
It's supposed to be a temporary solution. It being temporary depends on Linux taking off in, among many other markets, the games market. So you need to WineX for linux to take off, and you need Linux to take off before ports start to make sense. It's a double blade knife yes, it could well be the case that games never get ported to Linux, but that is not intrinsic to your reasoning, it depends on other things. For instance, if we get to a point where it's easier to develop under Linux, then having a Linux port will make sense economically. If it never makes sense quility wise and specially, economically, all hope is lost.
If other companies see that opportunity and it starts working for them, your company will start to see trouble ahead. This is how everything seems to work in most liberal countries.
You may say "why not us??"...well, you copmany probably thinks that if they start using open source libraries then your clients will take your products less seriously (if they are internal, then whoever decides how good and "propietary" to your company you IT department is), and also, they don't even want to think about the possibility of any legal trouble. They don't care spending more money for things that are "supposed" to be profesionally built and supported.
Of course, the problem is one of non-innovation and superficial analysis. But hey, not many people got fired by not using open source, and surelly a lof of people did got fired because of that. People have prejudices and a carreer to take care, not to mention kids to feed. That's part of the equation too...
But things gradually change. As long as OSS becomes more and more usefull, early adopters that are wise enough to have a bussines plan that focuses on a real need will reap the rewards. Ofter a while, it will be everyone, but I think we'll have to way a decade or more...
Your bullets are fine, but you are missing the ones where you clearly understand where the company want to go and where they would like to be, and how you can help them. I've seen these neglected in a lot of situations. If you've got a great staff your three points are fine, but sometimes it's the other way arround. Clueless IT department doing all kind of useless crap because that's all they want to do...or or because that the only thing they can do...and nobody can tell them otherwise cos "they ain't have a clue"...
We don't know the programming language much, so now we investigating it. Like trying to reverse engineering how C looks by looking at the source and playing arround with what we now is a compiled binary. Of course, C is a little a bit simpler than this, so we'll have to way a bit.
This news kind of look like we've figured out how many subroutines are there. The CPU rules are not very novel as it's physics itself.
The cure to cancer will come with a huge cost to human race. We'll know we are beautifull constructs and nothing more, and a lot of people will be "tweaking" us or trying to destroy us. This will much worse than the nanotech gray goo predictions. I want to answer, but I know many people will commit suicide even before they fully understand the implications. The "cure to cancer" looks like a joke compared to the change understanding us will bring along.
We consider superior to any other being right now, so the assertion is quite correct. It doesn't matter if you are more fit than a cow, we'll still try to save you while we'll eat the cow.
The only rule that applies that who's in charge makes the rules. Morality is whatever the ones in charge want or believe it to be.
Personally, I think that they should simply sell the printers without ink (and that it should be mandated). The printer manufacturers would probably love this, as it would let their revenue stream be the ink.
No, what we need is somewhat the opposite. If their revenue stream is the ink we are fucked, because then all their strategies comes down to trying to change catridge format and circuitry so that no other ink runs reliably with them. And it allows each munufacturer to NOT compete in the "already sold" market (the ink market is somewhat static given an initial market share).
So, in brief, what we need is to SEPARATE the ink and printer market. If we do that, well be charged what printers cost (expensive printer always better than cheapo ones) and what ink costs (expensive inks and cartridges). Standard sizes and connetors for ink cartridges would be the goal.
In that fashion, printers manufacturers should me mandated to follow certain rules for allowing third parties to produce geniune legal and quality catridges.
I was not talking about this case, but that for each case you need to take into account the hole path of consecuences. Of course in this case, reverse engineering is AOQ.
Slashdot Mirror
ln -s `which badname` /usr/bin/nicename
Reanse, repeat, and you have it your way.
I think the original author of the study was samuelson.
Yes, flying is better understood with math, it's a physical problem. I wasn't implying that no explanation at all is better than any model, I was refering to verbal explanation, like when you talk with a best friend, and if he is one of those guys that have a clear understanding of life and you, he'll be able to tell the REAL problem based on the behaveour.
Though the model may be more interesting or elegant...
You are missing something, to have an econometric model the right way, you must first have a teory. if you do not have a theory you can't use econometrics to prove or disprove your theory. Steps for an econometric model:
:)
- Present your theory or thesis
- Present the stochastic econmetric model you'll use to test it's validity
- Estimate coefficients
- Accept or reject your hipotesis (or refine your theory and go pack to the beggining)
- Predict
It has been proved that you make relly great but meaningless econometric models. Like average lenght of women dresses and (can't remember well, but it was something funny) the stock market. I think the correlation was significnt at 99% confidence level...
Nobody cares about averages here, you are supposed to predict the future, as in the order in which the coins will land as time passes. If you miss the order the results will vary to a great degree.
In other words, you could easily predict the NEXT coin flip (i keep on using the coin flip, but i am thinking on the physchohistory of humankind, so this experiments are NOT radom as in a normal coin flip) with near 100% accuracy. But you cannot predict the 10000 coin flip, because to predict that coin flip you'll need to know the exact flips that came before and in what order.
I think I made a short explanation long, but the point is the original poster does know that statistics work, but NOT so well when future predictions depend on the accuracy of all past predictions. Errors do get accumulated as any time series analist (or any econometrician) will tell you... prediction does lose accuracy as you get further away from "present time"...
I'd say phychohistory would be possible if and only if ALL individual acctions and all physical fenomens could be traced, and we'd be talking just about physical on a grand scale. We'dd be just attoms in deterministic paths...
For any conceibable behaveour there is a mathematical way of fitting the behaveour with a certain degree of probability. If something is not pure noise, then there must be some way to formalize it, though language itself or in mathematical notation.
This works, of course, don't add much value because they never explain how or why things are like that. With physics you don't have to explain the basic laws, they "just are", but with everything else, you better have some explanation of some sort because, in reallity, they are nothing more than constructs based on physical constraints.
On the other side, it might be funny to see how some people could see these formalizations as expressing more or being more accurate than "plain verb" explanation. "If it's hard to understand then it's real science!!" (wrong!)...
Just my thoughts so (I am biased yes, I've seen to many quantitative economics to believe equations express more just because the math is hard...they usually don't).
I think it's a great solution, in the variants you described, or being removable (a connector problem, not an HD problem) or being for buffering or standalone if you need extra storage (say 60 GB). In any case, you could have all three options (one internal fixed, a slot for removable variant and a firewire port & mode that activates buffering.
:)
A plain non-removable big HD consumes a lot and is ugly, and solid state fash cards are really unnecesarily expensive. Of course, who cares, some people will still think they need the solid state solution
Tape needs to be located somewher also, and there's no requirement that the HDD array be at the same location that the backed up systems.
Of course, HDD could be hacked more easily because there sure exist a network link to them. Unless it's a read only network (as in nothing can get deleted nor read unless you are inside certain room). This latter can also be enforced.
Better yet, why not use a laptop HD. Price/performance/size/consumption wise it's the best compromise. $150 or less for a 40 GB mini IDE drive. A standard adaptor for USB/PCMCIA/Firewire transfer and we are all set.
Opinion anyone?
Unless you are buying bandwidth. In that case, the analogy would be a cable company selling you some signal. As long as you don't draw more current than needed or unless you are not messing the signal, you are ok.
I think that's not a nice argument. There are many arguments for allowing DRM in the kernel, but not this one. For instance, Nukes are ok if rhe US or china has access to it, and bad if Saddam has it. Or is it that nukes are just plain bad? You can under for yourself...
Crossover is best server chilled when MSOffice is a requirement. You know, some people use other OS and have everything there. I can't even access my home dir under windows (can't say the opossite), and that is just sad... not to say that vendor support is shitty sometimes (for example, my IBM Thinkpad i-series 1460 freezes every now and then under Windows 2000 for unknown issues).
With Linux I was just lucky or something, because it never EVER hangs, and never gives me trouble (witched to gentoo recently because became fored of baby sitting Linux, so i am even happier now).
The point is it is good to be able to separate OS and applications, and specially if the only thing that's preventing you from trying them is Office.
It is not supported under Linux, so they should have no problems at all. In fact, there are many products that are not supported under several variants of Windows, but there are how-to's for the brave that want to try to make it work despite they will be on their own.
Could run != Is supported
You can get 20 items at $5 that would bring home back $40 rebates. The $100 item will only get you $2. That's some kind of normalization. Or look at is as a percentage...
It comes at 22, and at other places too. The Dell cluster uses a mix of servers, and it's not really a supercomputer (or I got confused at their page?) but a cluster of computer. It's nice to note that the IBM system at stop #5 is dated in 2000 while the rest of the systems are all 2002. So it's still at #5 after two years...nice uh?
If I where in a position to choose who to blame, in one side a 100 billion company, and in the other side, two kids asking for some hardware because they only FTP server went dead, then I'd choose Microsoft.
Large companies don't want the long, sensible answer, they want to quickly be able to point the finger at someone (Microsoft, or whoever allowed OSS to be used) and be gone. That's how they usually deal with problems, and this is nothing that may come to you as new. The details can be worked out afterwards.
I am not saying that line of reasoning is right, but that it's usual, so some people feeding kids think they are more secure beign able to blame a large company (not to be able to get the lost productivity back). This is where IBM, Oracle and some others come to play. OSS image must improve to the point where one can simply state "we are using what the most sucessfull companies use" and carry on to apply the patch that by then is surelly available...
You can't buy specific libraries from Red Hat. You ass is somewhat covered in general, but when you are making specific bets at specific libraries I guess the panic factor (or blamability) grows stronger. Also, there's a lot of people that do not OSS is not about turning the software industry into alstruism or religion, but about something more reasonable in term of efficiency.
It's supposed to be a temporary solution. It being temporary depends on Linux taking off in, among many other markets, the games market. So you need to WineX for linux to take off, and you need Linux to take off before ports start to make sense. It's a double blade knife yes, it could well be the case that games never get ported to Linux, but that is not intrinsic to your reasoning, it depends on other things. For instance, if we get to a point where it's easier to develop under Linux, then having a Linux port will make sense economically. If it never makes sense quility wise and specially, economically, all hope is lost.
If other companies see that opportunity and it starts working for them, your company will start to see trouble ahead. This is how everything seems to work in most liberal countries.
You may say "why not us??"...well, you copmany probably thinks that if they start using open source libraries then your clients will take your products less seriously (if they are internal, then whoever decides how good and "propietary" to your company you IT department is), and also, they don't even want to think about the possibility of any legal trouble. They don't care spending more money for things that are "supposed" to be profesionally built and supported.
Of course, the problem is one of non-innovation and superficial analysis. But hey, not many people got fired by not using open source, and surelly a lof of people did got fired because of that. People have prejudices and a carreer to take care, not to mention kids to feed. That's part of the equation too...
But things gradually change. As long as OSS becomes more and more usefull, early adopters that are wise enough to have a bussines plan that focuses on a real need will reap the rewards. Ofter a while, it will be everyone, but I think we'll have to way a decade or more...
Your bullets are fine, but you are missing the ones where you clearly understand where the company want to go and where they would like to be, and how you can help them. I've seen these neglected in a lot of situations. If you've got a great staff your three points are fine, but sometimes it's the other way arround. Clueless IT department doing all kind of useless crap because that's all they want to do...or or because that the only thing they can do...and nobody can tell them otherwise cos "they ain't have a clue"...
Don't give them ideas, they might as well try them and find out Mom likes it, or doesn't notice it, shrinking our heritage even further...
We don't know the programming language much, so now we investigating it. Like trying to reverse engineering how C looks by looking at the source and playing arround with what we now is a compiled binary. Of course, C is a little a bit simpler than this, so we'll have to way a bit.
This news kind of look like we've figured out how many subroutines are there. The CPU rules are not very novel as it's physics itself.
The cure to cancer will come with a huge cost to human race. We'll know we are beautifull constructs and nothing more, and a lot of people will be "tweaking" us or trying to destroy us. This will much worse than the nanotech gray goo predictions. I want to answer, but I know many people will commit suicide even before they fully understand the implications. The "cure to cancer" looks like a joke compared to the change understanding us will bring along.
We consider superior to any other being right now, so the assertion is quite correct. It doesn't matter if you are more fit than a cow, we'll still try to save you while we'll eat the cow.
The only rule that applies that who's in charge makes the rules. Morality is whatever the ones in charge want or believe it to be.
Personally, I think that they should simply sell the printers without ink (and that it should be mandated). The printer manufacturers would probably love this, as it would let their revenue stream be the ink.
No, what we need is somewhat the opposite. If their revenue stream is the ink we are fucked, because then all their strategies comes down to trying to change catridge format and circuitry so that no other ink runs reliably with them. And it allows each munufacturer to NOT compete in the "already sold" market (the ink market is somewhat static given an initial market share).
So, in brief, what we need is to SEPARATE the ink and printer market. If we do that, well be charged what printers cost (expensive printer always better than cheapo ones) and what ink costs (expensive inks and cartridges). Standard sizes and connetors for ink cartridges would be the goal.
In that fashion, printers manufacturers should me mandated to follow certain rules for allowing third parties to produce geniune legal and quality catridges.
I was not talking about this case, but that for each case you need to take into account the hole path of consecuences. Of course in this case, reverse engineering is AOQ.