Anyone who has worked in the finance industry on the tech side of things has probably seen eye-searing levels of problems like this. It's clusterfucks all the way down. It always surprised me that something that seems like such a natural fit for software was always, without fail, so riddled with glaring bugs that it's almost unfathomable that you are the first person to notice them. At a lot of shops, the bugs are so ingrained in the process that they can't even be fixed. Working in the finance industry certainly doesn't inspire confidence in the finance industry.
As cheap manufacturing countries start directly competing against their customers, the cost of using those countries for manufacturing will increase tremendously. At some point, knowing that you are likely to be competing against your own product (but cheaper and possibly built with slightly substandard parts) will make it more cost effective to build your product locally. It's kind of surprising that the governments of these countries aren't bending over backwards to try and prevent these counterfeiting/knockoff operations because it seems like the economic impacts for those countries could be devastating.
These blatant stereotypes about public jobs as a land of lavish benefits are long out of date. Cities vary enormously. Very many cops are *very* poorly paid in America.
Cops are not extravagantly paid but, any profession with a "20 years and I'm out with a pension" kind of system ends up being an extremely well paid profession in the long term. My father was an enlisted military man and, at the age of 40, retired with a pension and disability pay. Never saw even a glimpse of combat and was not injured in the line of duty. "Disability pay" actually means, "I filed the paperwork to indicate that my shoulder hurts now and didn't hurt 20 years ago". So, after working a desk job for 20 years, he retired with a pension that more than covered the mortgage on a very nice house and also had a resume that landed him a crazy good job. As an ex-enlisted man. Not an officer.
If you want to gamble and try to be rich in your 20s or 30s, go into the technology sector. If you want to be laughably well off in your 40s, go into the public sector/military, put in your 20 years and then do whatever you want. I did the former and, holy shit do I wish I'd done the latter.
An interesting stance. It kind of implies that my blame on "The War on Drugs" isn't the right origin. And, maybe it's not. But, I would say that a "War on Poverty" and "A War on the Poor" aren't the same. A "War on Poverty" is effectively socialism. However, I think the war on drugs is very much a war on the poor. And that war is insanely profitable. And, you are wrong that we didn't enslave those poor: The vast number of the casualties in our war end up in for-profit prisons. So, rather than import black people to work as slaves in cotton fields, we've rounded them up and let people profit by the simple headcount.
It's almost too horrible to think about it without getting sick to your stomach. It's literally just a less obvious form of slavery.
The problem with "just say no" (and the resulting DARE programs) is that it's effectively the same as taking a policy of teaching "abstinence only" in schools. It would have been a lot more useful for me to have a drug education program in school that, yes, told me to not take drugs but, at the same time educated you on the real effects of drugs: "If you drink, don't drive or do anything else that requires precise motor coordination", "If you take LSD, do it with other people and, if possible, have a chaperon", "Hydration is vitally important if you are taking ecstasy", etc, etc.
People are going to take drugs. An "abstinence only" policy towards drugs is going to result in unnecessary deaths.
I kind of assumed I'd have some responses along these lines and, you make a valid point. But, I think there is a genuine difference between "The War on Drugs" or "The War on Terror" and "The War on Poverty". In the former cases, you are declaring war on something where there are definable enemy combatants (drug dealers and terrorists). In the latter case, it's more a symbolic gesture to acknowledge that the government understands that their is a problem and wants to address it with the same seriousness that they'd address a war.
It's maybe a pedantic difference, sure. But, the need to wage war on poverty didn't arise from the fact that the poor were enemies of the state. On the other hand, the wars on drugs and terror were very much inspired by the governments desire to incarcerate or kill people they deemed enemies of the state. In those two cases, we now have literal wars and, we are starting to see that maybe a war mindset wasn't the best solution to the problems.
You are correct. It was Nixon who coined the term. I apologize for the oversight. I just remember the term becoming very popular in the 80s and so assumed it was Reagan (considering how crazy Nancy was with her anti-drug campaigns).
"I mean you call something a war, and pretty soon everyone is going to be running around acting like warriors." -- Major Colvin
Regan declared The War On Drugs and, unsurprisingly, people started acting like warriors. We now have a militarized police force that, in many areas, is effectively an occupying military. Guess what happens when an occupying military starts killing civilians? Insurgents are created.
I have a feeling this situation is going to spiral out of control pretty quickly.
What's "legitimate"? Is it "legitimate" to want to be able to handle 50% more simultaneous TCP-connections to your server or from your desktop, for example? Look at my earlier post for differences in sizes of several vital data-structures.
If pointer size is your primary concern, why not run a 16-bit OS? I say that in jest but, come on... If you are running a 32-bit machine (without PAE), moving from 4GB (well, more likely 3.5-ish gigabytes) to 8GB on 64-bits is going to cost you about $30. I can't even fathom a mission critical system unable to absorb a $30 upgrade.
I was a holdout on 64-bit stuff as well. With PAE, you get 36-bit pointers in the kernel and that's pretty cool. But, it comes at a performance cost. And these days, the performance cost is a lot worse because it means you are missing out on the performance *benefits* you'd get by moving to 64-bits.
Now, is it "legitimate" for developers to "steal" the hardware gains for their own benefit?
I'm not even sure what this means. Who stole anything? And who benefited by this theft? As far as I can tell, distro maintainers have decided that the maintenance of an almost unused flavor of their distro is no longer beneficial. And, it makes sense: Almost all Intel/AMD hardware has been 64-bit for many years. But, again, as I said, even Ubuntu will support 32-bit hardware until 2021. More than a decade after 32-bit Intel/AMD hardware was completely fucking obsolete.
I agree with you to a certain point but, how many people are running a 32-bit OS for legitimate reasons? Some, without a doubt. But probably not enough to justify the effort of keeping it around. I understand the argument of 64-bit pointer size increasing memory usage but, when running in 64-bit mode, the CPU has access to WAY more registers, things like SSE are implied, etc. It's a performance to memory usage trade off. Memory has been abundant and cheap for years and even lowly Atom chips have been 64-bit for quite some time.
Now, if the kernel decided to drop support for 32-bit builds, that would be pretty crazy. But, having the major distros drop it isn't that big of a deal. If you have hardware/software that legitimately *needs* a 32-bit distro, there will be options available for quite some time. In fact, Ubuntu 16.04 is supported until 2021. That's a pretty large runway to get your poop in a group.
Exactly this. It was a silly show but, you couldn't help but watch it even if your didn't give a shit about cars. It was *fun*. And had the added benefit of seeing insanely expensive cars that you'll never drive. Being driven by a mute gimp in a helmet. Brilliant.
But, that's not how it works at all. It probably costs less than $100 to build the ASICs and their planned obsolescence is measured in months. So, they will sell you a mining rig that *might* pay for itself in a year. But, they are taking the money you paid and building more efficient mining rigs. The more powerful rigs, once in use, increase the complexity of computing the bitcoins and, pretty quickly, the cost of running your mining rig doesn't cover the electricity costs of running it.
It really is the ultimate scam and I found the entire experience to be fascinating. Costly but fascinating.
If people don't pump money into mining, what will happen to the currency? I think it's pretty obvious that the "currency" will crash because it's largely propped up by Chinese miners and miner hype. It really is genuinely as scam. Sure, a handful of people have gotten rich from bitcoins but, unless you were in on the ground floor, trying to mine bitcoins is going to be a money loss. Trying to triage them on a meta-market to make some money is pure insanity. Unless you own a silicon fabrication plant, in China, your ability to make money with bitcoin is just a flip of a coin.
For a pure user of bitcoin, the Chinese control of the mining is not really an issue, sure. But, when you pull back the curtains, it's very, very obvious that the currency exists solely for the benefit of the companies that produce the ASICs that allow the computational complexity of the algorithms to grow. I understand that the difficulty of computing a bitcoin grows with the CPU power that is utilized to compute it. That's *exactly* what the Chinese manufacturers are banking on you *not* understanding.
When they sell you powerful ASICs that can compute bitcoins, they are banking on the fact that, in 6 months, when they've produce a chip that is 10x as fast, they'll be able to sell it to miners *again* at an absurd premium. And, for a short time, those miners might make a small amount of money. But, in the end, the profits will mostly lie with the company that sells you the ASIC. Because they control the computational price of the currency.
No, what happened is that the cheap bitcoins went away. A cursory examination of the dynamics of the thing would have revealed that mining bit coins increases in computational/electricity cost over time. So as long as value of bit coins exceeds the electricity cost of mining them, then there will be massive computational resources thrown at them.
This is wrong. You have to take into account the cost of the mining rig as well. If a $1000 mining rig is generating enough bitcoin to pay the electricity bill (which is non-trivial) then you have to look at how long it will take to pay back that $1000. Bitcoins are highly volatile so, today it might seem like it will repay itself in a few weeks. Tomorrow, it might literally look like centuries.
I stand by my statement that the only people that make money from bitcoin mining are the people that make the mining equipment.
No, by "scam" I mean exactly what I said: Why would someone sell you a magic money printing machine instead of running it themselves? The answer is: Because they can make more money selling it to you than they would make by running it themselves. Which kind of implies that buying a mining rig is a losing proposition. A scam.
Agreed. But, it's not something you realize until you've bought your shovels. I did it as an experiment so, I didn't mind losing some money to learn about the industry but, I'd like to educate potential miners: The magic money printing machine you bought was sold to you because the people who made it thought it was more profitable to sell it to you than to run it themselves.
Bitcoin as an idea is very interesting but, in actual function, it's a scam. I wanted to learn more about it a year ago and so bought some mining ASICs. It doesn't take long before you realize that your magic money printing device was sold to you at a cost that means you will invariably lose money. Which makes sense. If I have a magic money printing device, why the hell would I sell it to you instead of running it myself?
At the time when I bought my ASICs, the Chinese companies that were making them started changing directions and were moving towards a model of letting you rent ASICs that they would host. It's kind of a brilliant plan because it means they can reduce the costs of ASICs by creating WAY more than people will ever buy and then just renting people as many as they want. The rent more than pays for the electricity and manufacturing costs so, no matter what happens, the ASIC manufacturers always win. Unfortunately, they are the *only* people that win.
It's not surprising at all that China controls the bitcoin market. They *are* the market. All the ASICs come from China and if they don't sell the ASICs, they make a profit by running them. If they do sell the ASICs, they make more profit by selling them at an inflated price that speculators will pay. It's a hell of a racket.
Well, I'm sad to say that it was actually my fault. I have a habit when I'm debugging particularly tricky MPI code (we do massively parallel scientific simulation) to use "sleep(pid)" where "pid" is the "processor ID" or "rank" of the current MPI process.
See, your problem was that you should have used "sleep(pid % 10)". That way when you release your debug code, performance is just "not quite right" instead of "broken in an unfathomable way". Then, when you realize you've released your debug code, you can quietly go in, take it out, bless a new build and humbly accept your accolades for your engineering prowess.
Slack was the first distro I ever loaded back in the 90s.
I imagine that a large portion of slashdot readers cut their teeth on slack. Slack was probably the most popular user distro when slashdot was founded. And I imagine most people installed it by getting a CDROM out of the back of their first linux book. Downloading slack over a 2400bps connection was practically a sisyphean task.
They might be opposed to systemd on a moral level but, they may not be able to avoid it on a technical level. As systemd absorbs more and more things, and more things start to depend on systemd, it's a simple problem of manpower: Unless your distro is run by a multi-billion dollar company that has the resources to undo the damage caused by systemd, you may have no choice but to adopt it.
It's interesting to think of systemd in terms of The Cathedral and Bazaar. RedHat built a really nice stall in the middle of the bazaar. Initially everyone loved it and it was very healthy for the bazaar. Then one day they looked around and realized that they owned considerable interests in all the surrounding stalls in the bazaar and decided, "You know what? We should just build a cathedral right here in the middle of the bazaar and bring all these stalls inside".
Eudev is more or less a fork of udev before it was absorbed by systemd. In my dealings with it, it works exactly like how udev worked before the systemd developers hijacked the project so, I would imagine it was a trivial but necessary change for Slackware.
It will be interesting to see how long Slackware can resist systemd. Even venerable projects like LFS (Linux From Scratch) seem to be leaning towards its adoption. They still provide the non-systemd book as the default but, looking at the mailing lists, I'm not sure how long it will remain the default.
Slashdot Mirror
Anyone who has worked in the finance industry on the tech side of things has probably seen eye-searing levels of problems like this. It's clusterfucks all the way down. It always surprised me that something that seems like such a natural fit for software was always, without fail, so riddled with glaring bugs that it's almost unfathomable that you are the first person to notice them. At a lot of shops, the bugs are so ingrained in the process that they can't even be fixed. Working in the finance industry certainly doesn't inspire confidence in the finance industry.
As cheap manufacturing countries start directly competing against their customers, the cost of using those countries for manufacturing will increase tremendously. At some point, knowing that you are likely to be competing against your own product (but cheaper and possibly built with slightly substandard parts) will make it more cost effective to build your product locally. It's kind of surprising that the governments of these countries aren't bending over backwards to try and prevent these counterfeiting/knockoff operations because it seems like the economic impacts for those countries could be devastating.
Presumably it would only rain if you washed your spaceship that morning.
These blatant stereotypes about public jobs as a land of lavish benefits are long out of date. Cities vary enormously. Very many cops are *very* poorly paid in America.
Cops are not extravagantly paid but, any profession with a "20 years and I'm out with a pension" kind of system ends up being an extremely well paid profession in the long term. My father was an enlisted military man and, at the age of 40, retired with a pension and disability pay. Never saw even a glimpse of combat and was not injured in the line of duty. "Disability pay" actually means, "I filed the paperwork to indicate that my shoulder hurts now and didn't hurt 20 years ago". So, after working a desk job for 20 years, he retired with a pension that more than covered the mortgage on a very nice house and also had a resume that landed him a crazy good job. As an ex-enlisted man. Not an officer.
If you want to gamble and try to be rich in your 20s or 30s, go into the technology sector. If you want to be laughably well off in your 40s, go into the public sector/military, put in your 20 years and then do whatever you want. I did the former and, holy shit do I wish I'd done the latter.
An interesting stance. It kind of implies that my blame on "The War on Drugs" isn't the right origin. And, maybe it's not. But, I would say that a "War on Poverty" and "A War on the Poor" aren't the same. A "War on Poverty" is effectively socialism. However, I think the war on drugs is very much a war on the poor. And that war is insanely profitable. And, you are wrong that we didn't enslave those poor: The vast number of the casualties in our war end up in for-profit prisons. So, rather than import black people to work as slaves in cotton fields, we've rounded them up and let people profit by the simple headcount.
It's almost too horrible to think about it without getting sick to your stomach. It's literally just a less obvious form of slavery.
That's an excellent point and, if I hadn't started this chain of posts, I'd mod you up. Definitely something to think about.
The problem with "just say no" (and the resulting DARE programs) is that it's effectively the same as taking a policy of teaching "abstinence only" in schools. It would have been a lot more useful for me to have a drug education program in school that, yes, told me to not take drugs but, at the same time educated you on the real effects of drugs: "If you drink, don't drive or do anything else that requires precise motor coordination", "If you take LSD, do it with other people and, if possible, have a chaperon", "Hydration is vitally important if you are taking ecstasy", etc, etc.
People are going to take drugs. An "abstinence only" policy towards drugs is going to result in unnecessary deaths.
I kind of assumed I'd have some responses along these lines and, you make a valid point. But, I think there is a genuine difference between "The War on Drugs" or "The War on Terror" and "The War on Poverty". In the former cases, you are declaring war on something where there are definable enemy combatants (drug dealers and terrorists). In the latter case, it's more a symbolic gesture to acknowledge that the government understands that their is a problem and wants to address it with the same seriousness that they'd address a war.
It's maybe a pedantic difference, sure. But, the need to wage war on poverty didn't arise from the fact that the poor were enemies of the state. On the other hand, the wars on drugs and terror were very much inspired by the governments desire to incarcerate or kill people they deemed enemies of the state. In those two cases, we now have literal wars and, we are starting to see that maybe a war mindset wasn't the best solution to the problems.
You are correct. It was Nixon who coined the term. I apologize for the oversight. I just remember the term becoming very popular in the 80s and so assumed it was Reagan (considering how crazy Nancy was with her anti-drug campaigns).
"I mean you call something a war, and pretty soon everyone is going to be running around acting like warriors." -- Major Colvin
Regan declared The War On Drugs and, unsurprisingly, people started acting like warriors. We now have a militarized police force that, in many areas, is effectively an occupying military. Guess what happens when an occupying military starts killing civilians? Insurgents are created.
I have a feeling this situation is going to spiral out of control pretty quickly.
What's "legitimate"? Is it "legitimate" to want to be able to handle 50% more simultaneous TCP-connections to your server or from your desktop, for example? Look at my earlier post for differences in sizes of several vital data-structures.
If pointer size is your primary concern, why not run a 16-bit OS? I say that in jest but, come on... If you are running a 32-bit machine (without PAE), moving from 4GB (well, more likely 3.5-ish gigabytes) to 8GB on 64-bits is going to cost you about $30. I can't even fathom a mission critical system unable to absorb a $30 upgrade.
I was a holdout on 64-bit stuff as well. With PAE, you get 36-bit pointers in the kernel and that's pretty cool. But, it comes at a performance cost. And these days, the performance cost is a lot worse because it means you are missing out on the performance *benefits* you'd get by moving to 64-bits.
Now, is it "legitimate" for developers to "steal" the hardware gains for their own benefit?
I'm not even sure what this means. Who stole anything? And who benefited by this theft? As far as I can tell, distro maintainers have decided that the maintenance of an almost unused flavor of their distro is no longer beneficial. And, it makes sense: Almost all Intel/AMD hardware has been 64-bit for many years. But, again, as I said, even Ubuntu will support 32-bit hardware until 2021. More than a decade after 32-bit Intel/AMD hardware was completely fucking obsolete.
I agree with you to a certain point but, how many people are running a 32-bit OS for legitimate reasons? Some, without a doubt. But probably not enough to justify the effort of keeping it around. I understand the argument of 64-bit pointer size increasing memory usage but, when running in 64-bit mode, the CPU has access to WAY more registers, things like SSE are implied, etc. It's a performance to memory usage trade off. Memory has been abundant and cheap for years and even lowly Atom chips have been 64-bit for quite some time.
Now, if the kernel decided to drop support for 32-bit builds, that would be pretty crazy. But, having the major distros drop it isn't that big of a deal. If you have hardware/software that legitimately *needs* a 32-bit distro, there will be options available for quite some time. In fact, Ubuntu 16.04 is supported until 2021. That's a pretty large runway to get your poop in a group.
Exactly this. It was a silly show but, you couldn't help but watch it even if your didn't give a shit about cars. It was *fun*. And had the added benefit of seeing insanely expensive cars that you'll never drive. Being driven by a mute gimp in a helmet. Brilliant.
But, that's not how it works at all. It probably costs less than $100 to build the ASICs and their planned obsolescence is measured in months. So, they will sell you a mining rig that *might* pay for itself in a year. But, they are taking the money you paid and building more efficient mining rigs. The more powerful rigs, once in use, increase the complexity of computing the bitcoins and, pretty quickly, the cost of running your mining rig doesn't cover the electricity costs of running it.
It really is the ultimate scam and I found the entire experience to be fascinating. Costly but fascinating.
If people don't pump money into mining, what will happen to the currency? I think it's pretty obvious that the "currency" will crash because it's largely propped up by Chinese miners and miner hype. It really is genuinely as scam. Sure, a handful of people have gotten rich from bitcoins but, unless you were in on the ground floor, trying to mine bitcoins is going to be a money loss. Trying to triage them on a meta-market to make some money is pure insanity. Unless you own a silicon fabrication plant, in China, your ability to make money with bitcoin is just a flip of a coin.
For a pure user of bitcoin, the Chinese control of the mining is not really an issue, sure. But, when you pull back the curtains, it's very, very obvious that the currency exists solely for the benefit of the companies that produce the ASICs that allow the computational complexity of the algorithms to grow. I understand that the difficulty of computing a bitcoin grows with the CPU power that is utilized to compute it. That's *exactly* what the Chinese manufacturers are banking on you *not* understanding.
When they sell you powerful ASICs that can compute bitcoins, they are banking on the fact that, in 6 months, when they've produce a chip that is 10x as fast, they'll be able to sell it to miners *again* at an absurd premium. And, for a short time, those miners might make a small amount of money. But, in the end, the profits will mostly lie with the company that sells you the ASIC. Because they control the computational price of the currency.
Sure one can for example smuggle some drugs and weapons to the nearest star system and make good profit
Hello. Out of curiosity, can you point me towards those star systems? I've got a few bars of gold plated latinum for you, if you would be so kind...
No, what happened is that the cheap bitcoins went away. A cursory examination of the dynamics of the thing would have revealed that mining bit coins increases in computational/electricity cost over time. So as long as value of bit coins exceeds the electricity cost of mining them, then there will be massive computational resources thrown at them.
This is wrong. You have to take into account the cost of the mining rig as well. If a $1000 mining rig is generating enough bitcoin to pay the electricity bill (which is non-trivial) then you have to look at how long it will take to pay back that $1000. Bitcoins are highly volatile so, today it might seem like it will repay itself in a few weeks. Tomorrow, it might literally look like centuries.
I stand by my statement that the only people that make money from bitcoin mining are the people that make the mining equipment.
No, by "scam" I mean exactly what I said: Why would someone sell you a magic money printing machine instead of running it themselves? The answer is: Because they can make more money selling it to you than they would make by running it themselves. Which kind of implies that buying a mining rig is a losing proposition. A scam.
Agreed. But, it's not something you realize until you've bought your shovels. I did it as an experiment so, I didn't mind losing some money to learn about the industry but, I'd like to educate potential miners: The magic money printing machine you bought was sold to you because the people who made it thought it was more profitable to sell it to you than to run it themselves.
Bitcoin as an idea is very interesting but, in actual function, it's a scam. I wanted to learn more about it a year ago and so bought some mining ASICs. It doesn't take long before you realize that your magic money printing device was sold to you at a cost that means you will invariably lose money. Which makes sense. If I have a magic money printing device, why the hell would I sell it to you instead of running it myself?
At the time when I bought my ASICs, the Chinese companies that were making them started changing directions and were moving towards a model of letting you rent ASICs that they would host. It's kind of a brilliant plan because it means they can reduce the costs of ASICs by creating WAY more than people will ever buy and then just renting people as many as they want. The rent more than pays for the electricity and manufacturing costs so, no matter what happens, the ASIC manufacturers always win. Unfortunately, they are the *only* people that win.
It's not surprising at all that China controls the bitcoin market. They *are* the market. All the ASICs come from China and if they don't sell the ASICs, they make a profit by running them. If they do sell the ASICs, they make more profit by selling them at an inflated price that speculators will pay. It's a hell of a racket.
Well, I'm sad to say that it was actually my fault. I have a habit when I'm debugging particularly tricky MPI code (we do massively parallel scientific simulation) to use "sleep(pid)" where "pid" is the "processor ID" or "rank" of the current MPI process.
See, your problem was that you should have used "sleep(pid % 10)". That way when you release your debug code, performance is just "not quite right" instead of "broken in an unfathomable way". Then, when you realize you've released your debug code, you can quietly go in, take it out, bless a new build and humbly accept your accolades for your engineering prowess.
Slack was the first distro I ever loaded back in the 90s.
I imagine that a large portion of slashdot readers cut their teeth on slack. Slack was probably the most popular user distro when slashdot was founded. And I imagine most people installed it by getting a CDROM out of the back of their first linux book. Downloading slack over a 2400bps connection was practically a sisyphean task.
They might be opposed to systemd on a moral level but, they may not be able to avoid it on a technical level. As systemd absorbs more and more things, and more things start to depend on systemd, it's a simple problem of manpower: Unless your distro is run by a multi-billion dollar company that has the resources to undo the damage caused by systemd, you may have no choice but to adopt it.
It's interesting to think of systemd in terms of The Cathedral and Bazaar. RedHat built a really nice stall in the middle of the bazaar. Initially everyone loved it and it was very healthy for the bazaar. Then one day they looked around and realized that they owned considerable interests in all the surrounding stalls in the bazaar and decided, "You know what? We should just build a cathedral right here in the middle of the bazaar and bring all these stalls inside".
Definitely. Whenever I find deadlocks and race conditions in my code, I just sprinkle the code with sleep(1). PROBLEM SOLVED!
Eudev is more or less a fork of udev before it was absorbed by systemd. In my dealings with it, it works exactly like how udev worked before the systemd developers hijacked the project so, I would imagine it was a trivial but necessary change for Slackware.
It will be interesting to see how long Slackware can resist systemd. Even venerable projects like LFS (Linux From Scratch) seem to be leaning towards its adoption. They still provide the non-systemd book as the default but, looking at the mailing lists, I'm not sure how long it will remain the default.