"That all said, humans are impulsive beings. It's entirely possible that human intervention could turn a bad situation into a much worse one. I believe that this is an argument that needs to be settled by a set of scientific data, rather than by emotional arguments about the "battle-hardened pilot." An analysis of incidents that have already occurred is also helpful, if we want to compare the relative number of crashes due to software as opposed to pilot errors.
Certain aspects of the software definitely should not allowed to be overriden (ie. the part that prevents pilots from putting enough stress on the plane to snap it in half)."
I disagree. In an extreme situation it might be very useful to exceed the limit on the airframe. That needs to be the pilot's call.
Look, who is in command of the flight? The computer, or the pilot? It's a REALLY bad idea to have two commanders calling the shots on an objective. The old saying about too many cooks in the kitchen would seem to apply.
Until such a time as the pilot is no longer required, the pilot is in command of the plane, and all of the plane's systems need to subordinate to pilot, on demand.
It is a disaster in the making when the pilot has to compete with automatic systems that can take actions that counter the pilot's authority to command the flight. A previous poster here made a valuable point: computers have their strengths in emergency situations, and the pilot has their strengths. These two dynamic solutions need to be coherent, and based on the anecdotal references here about various failures, it is clear they are not.
"And remember that the recent plane crash in NY was caused by human error: the autopilot responded to the ice buildup by diving to maintain speed, the pilot 'corrected' what he though was an error and the plane fell to the ground like a stone."
The ice buildup would have caused a loss of lift... the dive was the autopilot response to detecting the loss of lift and an impending stall.
I have a serious problem with blaming the pilot in this situation.
What kind of report did the autopilot give the pilot as it was taking this recovery action?
Autopilot: Dude! I'm detecting a loss of lift and an impending stall! I'm gonna drop the nose! Low temps on the leading edges indicate we might be icing up!
Pilot: Hmmm ok lets take a closer look....
This IMO is much better than AutoPilot silently commanding a nose dive while blinking the stall warning indicator.
The autopilot committing to a dangerous action during a critical, and dangerous maneuver (like takeoff or landing) without giving the pilot enough info to assess the reason the AutoPilot acted is begging for disaster.
But I am guessing the crew had their hands full when it became apparent that they were overspeed, and starting to come unglued.
Not a lot of time to send a message. If they even realized they were fuxed before the plane started coming apart.
A key point I think a lot of posters here missed is that they were flying through a storm.... already getting bounced around, and making difficult for the flight crew to tell that some of the rattling and shaking was due to being overspeed.
"But I guarantee you for every one of those there are a thousand situations where a well designed computer will outperform a human being. Any casino will tell you, play the odds."
Slot Machines don't rip a person's arms out of their sockets when they fail.
Oobviously, you have never been involved in a design process where a reasonable consequence of system failure was loss of human life, and/or massive loss of property.
From what I have been reading though, is that the Pitot tubes are only triple redundant and they are ALL of the same design.
If the pentuple-redundant flight computer array gets bad airspeed data from three sensors that are all subject to identical failure conditions (due to them being identical) and environmental conditions to achieve the failure condition are met, then the plane is fucked.
This could be avoided easily by providing heterogenous input systems that do not share common mode failures.
If the system has a 25 knot window of stability then it seems pretty obvious to me that above all else you cannot afford to lose reliable airspeed indication. And I don't care how good the pitot designer claims their product is, I am going to want matching equipment from another supplier, using a different design with compatible performance specs.
Homogenous, redundant input gives a false sense of security. full stop.
Sorry I don't buy that. Unfortunately you get modded up for being cheeky.
The solder will know full well that what they are doing. When they level up to taking on humans. I think that is going to be when they have serious problems.
OTOH: If they pull an Ender's Game on the soldiers.... then there is going to be a horrible backlash when the soldiers find out....
While there is no doubt that a 4 year degree can have a lot of valuable content, the only thing that 4 year degree really proves, at the end of the day, is that the student can follow through on a long term goal and jump through hoops on command.
Employers find that more comforting for some reason.
Any time you see a article headline that formatted as an interrogative you are reading A FICTION, not journalism.
Any time a journalist wants to Make shit Up(tm) for any reason. They can do so simply by forming their story as a speculation.
Sometimes if the content is provocative enough, the journalist gets to write another story about the subject because the misrepresented parties in the speculative story make a point of 'correcting' the fiction with some better material, and the journalist doesn't have to do the leg work...
The problem with 'last forever' incandescent bulbs is that they're horrible at efficiency. 'Energy Miser' bulbs tend to last a bit shorter than standard bulbs because their filament's are thinner.
This wasn't true in the early part of the light-bulb story. Efficiency was not a big consideration back then. Marketability WAS. Early bulbs cost considerably more than the electricity they consumed.
It took 30 years before tungsten was used in bulb filaments.
It wasn't until 1910 that tungsten filaments were commercially viable due to a process invented by W.D. Coolidge.
It took least another 50 years before we got to the "energy miser" bulb.
The RedVines Theory of the Confection of the Universe can model Superfruity Goodness as embodied in the Skittle particle, or the JuicyFruit Dynamic Flavor Field, and many others due to it's inherent flexibility.
Alas, so far, it cannot predict where we might look for as yet undiscovered Superfruity Goodness, or how many grams of simple carbs such a structure might have.
By examining my place in a REAL universe composed of structures and processes too vast for the typical layman to even comprehend... and combine that with the understanding that all of that vastness is composed of equally REAL and unfathomably small quanta... I find, attenuates the more typical human tendency to behave like a total prick.
"My dear Watson! If you remove from further consideration all the elements of the [observed event] that have adequate explanations. Anything that remains; no matter how outlandish, must contain the truth!"
It's an unfortunate fact that the naming convention included an superlative... now they have to keep upping the ante each time a new property of magnetoresistance is discovered.
Slashdot Mirror
"That all said, humans are impulsive beings. It's entirely possible that human intervention could turn a bad situation into a much worse one. I believe that this is an argument that needs to be settled by a set of scientific data, rather than by emotional arguments about the "battle-hardened pilot." An analysis of incidents that have already occurred is also helpful, if we want to compare the relative number of crashes due to software as opposed to pilot errors.
Certain aspects of the software definitely should not allowed to be overriden (ie. the part that prevents pilots from putting enough stress on the plane to snap it in half)."
I disagree. In an extreme situation it might be very useful to exceed the limit on the airframe. That needs to be the pilot's call.
Look, who is in command of the flight? The computer, or the pilot? It's a REALLY bad idea to have two commanders calling the shots on an objective. The old saying about too many cooks in the kitchen would seem to apply.
Until such a time as the pilot is no longer required, the pilot is in command of the plane, and all of the plane's systems need to subordinate to pilot, on demand.
It is a disaster in the making when the pilot has to compete with automatic systems that can take actions that counter the pilot's authority to command the flight. A previous poster here made a valuable point: computers have their strengths in emergency situations, and the pilot has their strengths. These two dynamic solutions need to be coherent, and based on the anecdotal references here about various failures, it is clear they are not.
"And remember that the recent plane crash in NY was caused by human error: the autopilot responded to the ice buildup by diving to maintain speed, the pilot 'corrected' what he though was an error and the plane fell to the ground like a stone."
The ice buildup would have caused a loss of lift... the dive was the autopilot response to detecting the loss of lift and an impending stall.
I have a serious problem with blaming the pilot in this situation.
What kind of report did the autopilot give the pilot as it was taking this recovery action?
Autopilot: Dude! I'm detecting a loss of lift and an impending stall! I'm gonna drop the nose! Low temps on the leading edges indicate we might be icing up!
Pilot: Hmmm ok lets take a closer look....
This IMO is much better than AutoPilot silently commanding a nose dive while blinking the stall warning indicator.
The autopilot committing to a dangerous action during a critical, and dangerous maneuver (like takeoff or landing) without giving the pilot enough info to assess the reason the AutoPilot acted is begging for disaster.
We will find out....
But I am guessing the crew had their hands full when it became apparent that they were overspeed, and starting to come unglued.
Not a lot of time to send a message. If they even realized they were fuxed before the plane started coming apart.
A key point I think a lot of posters here missed is that they were flying through a storm.... already getting bounced around, and making difficult for the flight crew to tell that some of the rattling and shaking was due to being overspeed.
"But I guarantee you for every one of those there are a thousand situations where a well designed computer will outperform a human being. Any casino will tell you, play the odds."
Slot Machines don't rip a person's arms out of their sockets when they fail.
Oobviously, you have never been involved in a design process where a reasonable consequence of system failure was loss of human life, and/or massive loss of property.
From what I have been reading though, is that the Pitot tubes are only triple redundant and they are ALL of the same design.
If the pentuple-redundant flight computer array gets bad airspeed data from three sensors that are all subject to identical failure conditions (due to them being identical) and environmental conditions to achieve the failure condition are met, then the plane is fucked.
This could be avoided easily by providing heterogenous input systems that do not share common mode failures.
If the system has a 25 knot window of stability then it seems pretty obvious to me that above all else you cannot afford to lose reliable airspeed indication. And I don't care how good the pitot designer claims their product is, I am going to want matching equipment from another supplier, using a different design with compatible performance specs.
Homogenous, redundant input gives a false sense of security. full stop.
Sorry I don't buy that. Unfortunately you get modded up for being cheeky.
The solder will know full well that what they are doing. When they level up to taking on humans. I think that is going to be when they have serious problems.
OTOH: If they pull an Ender's Game on the soldiers.... then there is going to be a horrible backlash when the soldiers find out....
This is going to create all new forms of Post Traumatic Stress Disorders. :(
Remote control sociopaths maybe?
I think this is more sizzle than steak.
While there is no doubt that a 4 year degree can have a lot of valuable content, the only thing that 4 year degree really proves, at the end of the day, is that the student can follow through on a long term goal and jump through hoops on command.
Employers find that more comforting for some reason.
Yeah $250 - $500 bucks is enough to get bad parents to game this.
Any time you see a article headline that formatted as an interrogative you are reading A FICTION, not journalism.
Any time a journalist wants to Make shit Up(tm) for any reason. They can do so simply by forming their story as a speculation.
Sometimes if the content is provocative enough, the journalist gets to write another story about the subject because the misrepresented parties in the speculative story make a point of 'correcting' the fiction with some better material, and the journalist doesn't have to do the leg work...
mod parent up!
Ya know you might want to hire a therapist RE: your issues with Homophobia. You might actually learn something about yourself.
The problem with 'last forever' incandescent bulbs is that they're horrible at efficiency. 'Energy Miser' bulbs tend to last a bit shorter than standard bulbs because their filament's are thinner.
This wasn't true in the early part of the light-bulb story.
Efficiency was not a big consideration back then. Marketability WAS. Early bulbs cost considerably more than the electricity they consumed.
It took 30 years before tungsten was used in bulb filaments.
It wasn't until 1910 that tungsten filaments were commercially viable due to a process invented by W.D. Coolidge.
It took least another 50 years before we got to the "energy miser" bulb.
If the economic market had fluidity.... there would not be a credit crunch.
The RedVines Theory of the Confection of the Universe can model Superfruity Goodness as embodied in the Skittle particle, or the JuicyFruit Dynamic Flavor Field, and many others due to it's inherent flexibility.
Alas, so far, it cannot predict where we might look for as yet undiscovered Superfruity Goodness, or how many grams of simple carbs such a structure might have.
42 for large values of 42.
Wow a 5D /.er promoting a 7D. o.0 How rare is that!
Citation please... and maybe a direct quote?
Or carve it into the handrail of a bridge with a pen-kinfe like Hamilton did.
http://en.wikipedia.org/wiki/Quaternion#History
By examining my place in a REAL universe composed of structures and processes too vast for the typical layman to even comprehend... and combine that with the understanding that all of that vastness is composed of equally REAL and unfathomably small quanta... I find, attenuates the more typical human tendency to behave like a total prick.
and as long as you stay out of the universe of the very, very big, and the very very small, Newton is accurate out to about 1e-6.
Good enough for g'varmit work.
"My dear Watson! If you remove from further consideration all the elements of the [observed event] that have adequate explanations. Anything that remains; no matter how outlandish, must contain the truth!"
My apologies to A.C. Doyle.
Not taking sides. For me, string theory is all about tying my shoes.
Who says the GUT has to be elegant?
Reality is messy. Deal with it.
looks like 4Chan wins again....
Fuxing stupid.
It's an unfortunate fact that the naming convention included an superlative... now they have to keep upping the ante each time a new property of magnetoresistance is discovered.