MS makes utterly stupid decision, different from what anybody competent has done before for this task, insecurity ensues. Nobody with a clue is surprised.
I see that all the time in the IT industry. Process is more important than solid technology. And it is getting worse. At one customer, I basically have to sneak the solid tech past their processes. Of course the paper-pushers do not want that and make it harder. Until things break and cannot be easily repaired anymore.
Personally, I think what is missing is a "Chief Engineer" that is an actual engineer, very senior, very experienced and that has both final say and final responsibility on all tech decisions. What they have today is a CTO that is an administrative position. That just does not cut it.
I think this is the most important story on Slashdot in a long time.
Hopefully, it will make a lot of engineers remember their training and remember that redundancy is the one thing you always need and want, and that no MBA nil-whit has any business to tell you otherwise.
Part of the problem is Boeing didn't want pilots to have to retrain and certify under a different type of aircraft.
So they've jiggled things around to make it look like it's just like any other 737, but it now has different flight characteristics.
Indeed. "Jiggle" too much and things break. They jiggled way more than was sane and they ended up killing a lot of people. This is basically a textbook-example of how to mess it up. It will be used in teaching as it was so extremely clear from the onset that this would not work. It is also an example of large complex organizations completely losing sight of what is important.
It is not as though the testing authority *wanted* people to die.
No. But they got sloppy and trusted too much in theoretical calculations. And they put no redundancy in place for the case that their predictions turned out to be wrong.
That is why in safety-critical systems you use reliable, established designs and you put in redundancy. Boeing failed to do both. The accident you quote was a failure to do both. Fuckushima was a failure to do both. Etc. These are well-established principles of engineering for safety critical systems. Ignore them and you _will_ kill people, sometimes a lot of them.
This is subject to diminishing returns. If you do it right, the 3 sensors will be of different make. Then not all 3 will fail at the same time. When one fails, you do an not-so urgent emergency landing with the remaining two working. If two fail, you do an urgent emergency landing with one still working. If all three fail, something larger is wrong (plane on fire?) and that is what likely will kill you. 7 offer no advantage here.
That said, I can't speak for the airline industry but one thing has been clear, more often than not doing reliability calculations is borderline a waste of time. Saying a unit fails ever x number of units is irrelevant as the vast majority of sensor problems across multiple industries are systematic. Incorrectly designed or not suitable for service. Assembly faults, poor maintenance, sheer dumb luck, all of these don't play out in manufacturers reliability figures.
As can be seen regularly. For example, look at reactor safety predictions. We should not have any of the 4 large catastrophes that happened so far. The numbers are no good. Solid engineering puts in redundancy exactly to be prepared for things you do not know in advance, and they do happen because no model is perfect.
And every good engineer in the safety/security space knows that. But bring in some MBAs and they will find statistics that say this can be done cheaper. And cheaper. And then a lot of people die.
This is pretty much what I would have written when I had heard of the single sensor before these crashes. It is bloody obvious.
"Hard work" is pretty much a failed direction. What you need to do is works smart. But most people cannot do that and hence fall for the "hard work" meme. Also, "hard work" does tend to not produce a lot of value in these times.
Indeed. Being qualified and _interested_ in what you do always puts you ahead of the pack. It gives you options. Sure, those that cannot hack it would really like to get it all and not have to compete against people with an actual clue. And that is pretty much the direction this story comes from.
Ok we get it. She thinks tax/redistribution is the answer. That just feels like half the solution. The other half is what will the people DO with their UBI or similar style income?
That is indeed the question. Of course, that average backwards-looking moron is still fighting hard against the very idea of an UBI, despite it being extremely obvious that there will not be another choice to keep society functioning. But an UBI is only part of the solution and not enough. Most people need work to have meaning in their lives. Sure, there are those that will find this very easy to handle, but it will probably be restricted to the 10-15% of independent thinkers. The rest will find it really hard to deal with the situation. They will get sick, turn to drugs, get aggressive, become extremists. That could destroy society just as easily.
These morons think they define reality and they can make it whatever they want. In actual reality, that is of course not how things work at all.
I do agree that it is not a good idea to fear having your job automated away. It will happen, but fearing it will just make things even worse. And there is absolutely nothing that can be done about it, except some temporary stop-gaps that will make things even worse though.
That is even worse than decision by committee. This is decision by uninformed masses that have no clue how things work and what is and is not important.
While I am all for moving forward with extremely care in this area, letting the public decide about it is the worst idea possible. They will either be panicked irrationally or overoptimistic just as irrationally. Not good at all.
Slashdot Mirror
MS makes utterly stupid decision, different from what anybody competent has done before for this task, insecurity ensues. Nobody with a clue is surprised.
I see that all the time in the IT industry. Process is more important than solid technology. And it is getting worse. At one customer, I basically have to sneak the solid tech past their processes. Of course the paper-pushers do not want that and make it harder. Until things break and cannot be easily repaired anymore.
Personally, I think what is missing is a "Chief Engineer" that is an actual engineer, very senior, very experienced and that has both final say and final responsibility on all tech decisions. What they have today is a CTO that is an administrative position. That just does not cut it.
Well, since the CEO probably needs to go to prison for a few hundred years for this, fat chance.
I think this is the most important story on Slashdot in a long time.
Hopefully, it will make a lot of engineers remember their training and remember that redundancy is the one thing you always need and want, and that no MBA nil-whit has any business to tell you otherwise.
Part of the problem is Boeing didn't want pilots to have to retrain and certify under a different type of aircraft.
So they've jiggled things around to make it look like it's just like any other 737, but it now has different flight characteristics.
Indeed. "Jiggle" too much and things break. They jiggled way more than was sane and they ended up killing a lot of people. This is basically a textbook-example of how to mess it up. It will be used in teaching as it was so extremely clear from the onset that this would not work. It is also an example of large complex organizations completely losing sight of what is important.
It is not as though the testing authority *wanted* people to die.
No. But they got sloppy and trusted too much in theoretical calculations. And they put no redundancy in place for the case that their predictions turned out to be wrong.
That is why in safety-critical systems you use reliable, established designs and you put in redundancy. Boeing failed to do both. The accident you quote was a failure to do both. Fuckushima was a failure to do both. Etc. These are well-established principles of engineering for safety critical systems. Ignore them and you _will_ kill people, sometimes a lot of them.
This is subject to diminishing returns. If you do it right, the 3 sensors will be of different make. Then not all 3 will fail at the same time. When one fails, you do an not-so urgent emergency landing with the remaining two working. If two fail, you do an urgent emergency landing with one still working. If all three fail, something larger is wrong (plane on fire?) and that is what likely will kill you. 7 offer no advantage here.
That said, I can't speak for the airline industry but one thing has been clear, more often than not doing reliability calculations is borderline a waste of time. Saying a unit fails ever x number of units is irrelevant as the vast majority of sensor problems across multiple industries are systematic. Incorrectly designed or not suitable for service. Assembly faults, poor maintenance, sheer dumb luck, all of these don't play out in manufacturers reliability figures.
As can be seen regularly. For example, look at reactor safety predictions. We should not have any of the 4 large catastrophes that happened so far. The numbers are no good. Solid engineering puts in redundancy exactly to be prepared for things you do not know in advance, and they do happen because no model is perfect.
If the numbers are correct. Two failures in a few months say something different...
And every good engineer in the safety/security space knows that. But bring in some MBAs and they will find statistics that say this can be done cheaper. And cheaper. And then a lot of people die.
This is pretty much what I would have written when I had heard of the single sensor before these crashes. It is bloody obvious.
I can see why chronic fatigue sufferers might not want to wait half a century for an effective treatment.
You mean they prefer "never" to half a century? Because that is the effect they are having.
I do that on every Debian installation: I just put in sysVinit. Debian still supports that and you get rid of a ton of problems if you do it.
You can still install Debian without the systemd atrocity. Works pretty well.
Se my self-answer above.
Funny. Or tragic.
"Hard work" is pretty much a failed direction. What you need to do is works smart. But most people cannot do that and hence fall for the "hard work" meme. Also, "hard work" does tend to not produce a lot of value in these times.
Indeed. All the "luck" in the world will not help you if an opportunity comes along, but you cannot use it because you are not prepared.
First, stop equating money with success.
Indeed. If you had no life, or did a _lot_ of damage (like Gates did), but you have money, you wasted your life.
But equating money with success is the American Thing. Recognizing actual reality is to difficult for the people that do it.
Indeed. Being qualified and _interested_ in what you do always puts you ahead of the pack. It gives you options. Sure, those that cannot hack it would really like to get it all and not have to compete against people with an actual clue. And that is pretty much the direction this story comes from.
He is an incompetent hack. He is an extraordinary salesman though. You know the scum that sells you trash at hugely inflated prices.
Ok we get it. She thinks tax/redistribution is the answer. That just feels like half the solution. The other half is what will the people DO with their UBI or similar style income?
That is indeed the question. Of course, that average backwards-looking moron is still fighting hard against the very idea of an UBI, despite it being extremely obvious that there will not be another choice to keep society functioning. But an UBI is only part of the solution and not enough. Most people need work to have meaning in their lives. Sure, there are those that will find this very easy to handle, but it will probably be restricted to the 10-15% of independent thinkers. The rest will find it really hard to deal with the situation. They will get sick, turn to drugs, get aggressive, become extremists. That could destroy society just as easily.
Please ignore comment. In good old /. tradition I only read the headline.
Well, she does seem to get what is happening. I am less sure she understands how extremely difficult it will be to deal with it.
These morons think they define reality and they can make it whatever they want. In actual reality, that is of course not how things work at all.
I do agree that it is not a good idea to fear having your job automated away. It will happen, but fearing it will just make things even worse. And there is absolutely nothing that can be done about it, except some temporary stop-gaps that will make things even worse though.
That is even worse than decision by committee. This is decision by uninformed masses that have no clue how things work and what is and is not important.
While I am all for moving forward with extremely care in this area, letting the public decide about it is the worst idea possible. They will either be panicked irrationally or overoptimistic just as irrationally. Not good at all.