Well, I think the most likely candidate is not anything *simple* at all (although human error may certainly be involved). Most catastrophic spacecraft failures (and other engineering failures, for that matter) typically result from a couple of things going wrong in a complex, yet interconnected, way - often referred to as "the failure chain".
In the case of a simple flight software error, not only would the boneheaded engineer that wrote the code have screwed up, but also the organization that is supposed to validate and test the software (and they are usually fairly independent, if things are set up correctly).
Another thing I've encountered with working on NASA spacecraft is that the systems engineering is just a bitch - the thing is so complex nobody really understands how it works from one end to the other. The guidance guys may have this slick new way to do an inertion burn - but they don't realize the duty cycle on the engine will create a heat load that the thermal system can't dissipate - and the thermal guys don't know the first thing about how the engine and guidance works so they don't catch it until it's too late. Misunderstanding of complex systems is often a problem that leads to failures.
It's certainly a significantly different kind of problem. You don't have to worry about parachutes and atmospheric entry but the orbital mechanics are extremely tricky due to the weak gravity field.
Fortunately, they do have some experience to draw upon. NASA's NEAR mission managed to land on a 21x8x8 mile asteroid named Eros and operated afterwards, despite the fact it was not actually designed to land. Performing that end of mission "stunt" contributed greatly to the overall knowledge of operating around small bodies, although the comet will be even less massive and more challenging.
Besides testing for the impact, we ought to try to hang on to the resources we have to track these storms. I work in satellite ops, and one of the best resources I've got is The Space Environment Center, which was referenced in the story and provides real time solar monitoring and space weather predictions.
What wasn't mentioned in the story is that there is a move afoot in Congress to either drastically reduce it's funding or eliminate it altogether: U.S. Space Weather Service in Deep Trouble . All for a service that costs a measily $8 million bucks. Think about how much a "surprise" solar event costs. Since space weather can wreak havoc with high technology systems as well as a hurricane can, it's an appropriate thing for NOAA to be providing to the public, IMHO
Re:Is Ed Lu running lunix?
on
Space Blog
·
· Score: 5, Informative
Only the laptops and the network that are used for housekeeping stuff like email, downlinking digital images, the crew's private phone (yes, they have one), procedures, and timelines are windows based.
The laptops that plug directly into the core flight computers to allow the crew to do command and control of the station itself (critical things like maneuvering, opening valves, etc.) are running Solaris 2.5.
These laptops are currently IBM 760 XDs but they are going to be upgraded next year, at which point the OS will be transitioned to Linux.
Some of the payloads in the racks use various flavors of Unix depending on what the investigators put together but I'm not sure how many (if any) of those are Linux.
I'm an aerospace engineer in the space/defense sector, which is notoriously cyclical. Sometimes I questioned my sanity in staying in it, but working on spacecraft was just too fun to leave.
Ironically, it's a pretty good place to be these days. Defense spending is way up for obvious reasons, and it's poured into high-tech programs that demand a lot of engineering. Many, many people bailed out for the tech boom in the late '90s, which thinned out the employee base. I won't say there is a massive shortage of people or a huge number of unfilled openings, but there is a healthy market with employers like Lockheed, Raytheon, CIA - calls from recruiters and job hopping are pretty common where I work.
And, national security jobs can't be outsourced to India - 99% of them have US citizenship as a basic requirement.
Slashdot Mirror
In the case of a simple flight software error, not only would the boneheaded engineer that wrote the code have screwed up, but also the organization that is supposed to validate and test the software (and they are usually fairly independent, if things are set up correctly).
Another thing I've encountered with working on NASA spacecraft is that the systems engineering is just a bitch - the thing is so complex nobody really understands how it works from one end to the other. The guidance guys may have this slick new way to do an inertion burn - but they don't realize the duty cycle on the engine will create a heat load that the thermal system can't dissipate - and the thermal guys don't know the first thing about how the engine and guidance works so they don't catch it until it's too late. Misunderstanding of complex systems is often a problem that leads to failures.
Fortunately, they do have some experience to draw upon. NASA's NEAR mission managed to land on a 21x8x8 mile asteroid named Eros and operated afterwards, despite the fact it was not actually designed to land. Performing that end of mission "stunt" contributed greatly to the overall knowledge of operating around small bodies, although the comet will be even less massive and more challenging.
Besides testing for the impact, we ought to try to hang on to the resources we have to track these storms. I work in satellite ops, and one of the best resources I've got is The Space Environment Center, which was referenced in the story and provides real time solar monitoring and space weather predictions.
What wasn't mentioned in the story is that there is a move afoot in Congress to either drastically reduce it's funding or eliminate it altogether: U.S. Space Weather Service in Deep Trouble . All for a service that costs a measily $8 million bucks. Think about how much a "surprise" solar event costs. Since space weather can wreak havoc with high technology systems as well as a hurricane can, it's an appropriate thing for NOAA to be providing to the public, IMHO
The laptops that plug directly into the core flight computers to allow the crew to do command and control of the station itself (critical things like maneuvering, opening valves, etc.) are running Solaris 2.5.
These laptops are currently IBM 760 XDs but they are going to be upgraded next year, at which point the OS will be transitioned to Linux. Some of the payloads in the racks use various flavors of Unix depending on what the investigators put together but I'm not sure how many (if any) of those are Linux.
Ironically, it's a pretty good place to be these days. Defense spending is way up for obvious reasons, and it's poured into high-tech programs that demand a lot of engineering. Many, many people bailed out for the tech boom in the late '90s, which thinned out the employee base. I won't say there is a massive shortage of people or a huge number of unfilled openings, but there is a healthy market with employers like Lockheed, Raytheon, CIA - calls from recruiters and job hopping are pretty common where I work.
And, national security jobs can't be outsourced to India - 99% of them have US citizenship as a basic requirement.