Slashdot Mirror
Nattering Nabobs Of NASA Negativity
code_rage writes "IEEE Spectrum Magazine has an article by James Oberg which enumerates some of the problems which have cropped up and will crop up during assembly of Space Station Alpha (or whatever it is called this week).
The article lists many software problems, including safety related issues. Also a problem which was news to me: the U.S.-supplied Solar Arrays operate at a high voltage, which would place astronauts at risk of a potentially deadly plasma discharge during EVA. The workarounds include some Catch-22's."
Sheesh, I wish for a return to the old days, when things would just blow up.
No boom today. Boom tomorrow. There's always a boom tomorrow. - Cmdr. Susan Ivanova
Don't practice your alliteration on me!
No boom today. Boom tomorrow. There's always a boom tomorrow. - Cmdr. Susan Ivanova
Oh... wait - we're talkin' 'bout that ion stuff...
Hi! This is the Sig, blatantly attached to the end of this comment.
let me guess, he wrote this article for his secret race of giant robots with which he plans to take over the world? (starting with the ISS).
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
...chuckle...
I see no reason to panic based on what was in that article.
Information wants to be anthropomorphized.
We are catching up to science fiction, and I for one am glad to see it. Now bring me my flying car.
*whup* "Get along, little electrons. Heeyah!"
Why should this be a surprise that there might be electrical problems? Haven't we learned from Star Trek that future space craft, when under any kind of stress, immediately give off massive sparks through the consoles?
________________
________________
Private Essayist
Older chips are more of a known factor. They've had more time to determine the necessary power and radiation shielding levels. Plus, if you write your software right, there's not a whole lot of things you really need a high speed processor for. Not like they're going to be doing some Q3 deathmatches up there.
"That's Tron. He fights for the Users."
Yup, it certainly is nice to have that reliable Mir technology to depend on in case the new stuff has problems.
I remember asking the same question when I worked at Rockwell Collins, why were they using 186's in the Boing 777. They don't need more. Plus that chip has been around long enough for them to have a damn good idea of what could go wrong with it. The more complex the chip, the higher the chances of something going wrong.
As x approaches total apathy I couldn't care less.
by the way, how does one pronounce "Zvezda"?
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
Alas, poor HAL-9000! I knew him, Arthur.
Well, at least it won't run Windows.
Want to learn about race cars? Read my Book
For chips to perform in space, they have to be able to perform under much more extreme conditions than those which are found in a comfortable, enormous multi-fan gargantuan heat-sinked p4 case. To this end, the silicon technology, I'm guessing the voltage swings of transistors and base voltages, etc, must be re-designed. This takes time, not only for design, but to get permission from the chip's manufacturer (Intel) to do this. For instance, a prof of mine is working at Sandia Labs on a pentium, a six year old chip, to perform in space.
Mike
Intel transfer the difficult from Hadware to software, for get more power, programmer need more technology. -- chinaitn
Yes, the solar arrays on the ISS are supposed to be around 160 volts, which is a lot higher than most satellites. They've designed around it, though. The ISS is connected to the solar array via a positive ground rather than a negative one, which should keep the station itself safe for astronauts. (They still should avoid the array if possible, though.) And the plasma contactor mentioned in the article is a pretty useful item that's worked on scientific satellites for years. With the PCU working, they shouldn't have many problems.
:) )
If the PCU goes out, though, plasma charging is a problem. You have the possibility of electrical arcs...which are equally dangerous to astronauts and to the electrical equipment on the station. The torques on the station change when the ground is disturbed, possibly changing its orbit or spin. Ion sputtering (erosion of the spacecraft hull) increases...although that's probably the least of your concerns. There may be periods in the orbit when the astronauts, if they work quickly, can get out and fix things safely. That'd be tough, though, as they hit the aurorae belts every orbit and the South Atlantic Anomaly at least once every seven. You don't want to be EVA over south america next to an ungrounded high voltage space station.
But the folks who build the ISS know what they're doing, and I think they'll have the plasma environment under control. Some of the other problems mentioned in that article I did not know about and do look like a worry, but I'm sure things aren't as dire as the article writer is predicting.
(Full disclosure: I work (subcontract) for NASA on a satellite program unrelated to the ISS. Whether that makes me knowledgeable or just biased is your decision.
Genocide Man -- Life is funny. Death is funnier. Mass murder can be hilarious.
is that noise is a major consideration of space station design. I saw something like that on Discovery Channel once and they put a mic in on of the noisier modules in Mir and it sounded like a small NOC with too much equipment in it.
:-)
They said that the tolerable noise level in any module (according to NASA spec) was somewhere around 30 dB but a lot of the modules being made by other countries (notably Russia) exceeded those specs by as much a 40dB. Compund that with the fact that the sound has nowhere to go but in.
In space no one can hear your server crash!
"Me Ted"
BOSTON SUCKS!
Actually, it was stolen from Monty Python. Specifically, the Church Missile sketch (not sure of the exact name of the sketch, but if you've heard it, you know what I mean).
No boom today. Boom tomorrow. There's always a boom tomorrow. - Cmdr. Susan Ivanova
The payload I've been working on--and from the best I can tell, most of the other payloads on UF-1, the first of the many Utilization Flights--was bumped from its flight. Technically, we weren't on schedule, but the schedule is unrealistic to begin with.
The manifest is full of lies, damned lies, and statistics, but that's no different than any other NASA program. It's the typical NASA FUD: make the schedules unreasonable, and when the contractors fail to meet specs, blame the contractors, slip the schedule, and ask Congress for more money.
It makes one wish for the days of carte blanche, when the schedules were unreasonable, but you could at least throw enough money and brainpower at a situation to get the thing solved. People worked long hours, slept at their desks, had recreation at work, and took simple pleasure at their jobs being finally completed--then moved to another job.
You see, the geek culture today has a lot of roots in the geek culture of the '60s--but instead of Apollo and Saturn, we work on Linux and Gnome. Rather than the Evil Empire of the Soviet Union, which hid all their secrets behind an impenetrable Iron Curtain, we now fight the Evil Empire of Redmond, which hides all their secrets behind the impentrable Closed-Source Curtain.
All of which begs to ask: where's the deals with Life, and when does Tom Wolfe write a book on the open-source movement?
--
-- Geof F. Morris
Also, at one time, at least, Intel provided MIL-spec and rad-hardened 386SX chips.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
Coining the phrase "Nattering Nabobs of Negativity" is one of the two memorable things that Spiro Agnew accomplished while he was Vice President. The other was managing to get forced to resign his office during the middle of the Watergate scandal for something totally unrelated to Watergate.
The next Cmdr Taco duplicate will be ready soon, but subscribers can beat the rush and see it early!
The big, fat gates in a 386SX are also nice and sturdy from an electrical perspective.
You also have to factor in deaths from the Russian space program (Vladimir Komarov, Georgi Dobrovolsky, Vladislav Volkov, and Viktor Patsayev) and the countless animal deaths in the early days. Even when they were not directly involved in NASA's programs, it's not as though NASA didn't take heed of the mistakes learned by their deaths as well.
And modern Space exploration has only been going on for fifty years now. Presidents have been getting assasinated for the last couple hundred years. Since the space program started, only one president (JFK) has been assasinated, though their have been other attempts (Reagan, etc.) By your logic, space exploration remains ten times as fatal as the office of president.
-- Anne Marie
In one error discovered earlier this year, the corruption of two adjacent flags (bits in a status word) would command an air valve to open while locking out the "valve close" command; only a power cycle could reset the system and prevent all the air from leaking out.
What is the point of making things like this computerized with no manual workaround? That sounds poorly thought out. Surely a valve could be made so you could also close it by hand?? This reminds me of the models of BMW where you can't unlock the doors by hand, so if the power locks fail, you're locked in your own car (this actually happened to someone I know). Madness...
There is a 186 or more properly a 80186. it was never used in a PC. It was never intended to be used in a PC. It wasn't a prediessor to the 286. They were actually sybling similar architectures, one targeting the desktop and the other targeting embeded devices
As x approaches total apathy I couldn't care less.
Cool, you mean I've been using NASA jargon all this time?
Wah!
The reason they have 386SX chips is because of an array of things that come together.
A) They need to design new cooling systems since radiant heat dissipation does not exist in zero gravity (heat cannot rise in zero g).
B) They need to harden some of the shielding on the systems so that stray radiation does not have a negative effect on bits.
C) They need to have special chips made for stray radiation.
D) They need to be flight tested and sturdy.
E) They don't necessarily NEED blazingly fast computers to do the tasks required.
F) Development cycles for the software are much slower (There is not Fix/Release type thing, its Fix/Test/Test/Test/Test/Test/Fix/Test.../Release)
G) This is what is proven to work through time.
Is that clear enough?
Jay
"What's this script do? unzip ; touch ; finger ; mount ; gasp ; yes ; umount ; sleep Hint for the answer: not everyth
Eric
--
Be who you are...and be it in style!
... NASA should use several poorly-trained astronaut teams, instead of one well-trained team. This way, they could afford to lose one or more astronaut teams, but the mission could still continue because of the hot spares available.
Would this be a RAID system (Redundent Array of Inexperienced Dudes?)
DrLunch.com The site that tells you what's for lunch!
Comment removed based on user account deletion
This is a "person of great wealth or prominence" from arabic derived through Urdu. This is how a provincial governer in the Mogul Empire in India was called.
BTW, there is a site www.m-w.com, Merriam-Webster Online.
Tigers respect lions, elephants and hippos. Maggots respect no one. (C) S. Dovlatov
Others are less upbeat. For example, even optimists admit the development of space station software has been difficult.
And it will continue to be so, across the entire software industry, until software developers force the chip manufacturers to provide hard MMU encapsulation for fine-grain objects like they do for processes now.
Programs are unreliable to a very significant extent because their internal objects all live in the same address space and can merrily tramp all over anything they like under fault conditions. And fault conditions always arise in any non-trivial program, yet recovery is impossible in the general case because there is no internal protection against fault propagation.
Today's software developers are still using a 30-year old hardware model. Is it any surprise that software is still as flakey as ever?
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
The 80186 is basically an 8086 with a bunch of chipset and I/O features integrated on a single chip. It was popular for controller applications, like smart disk and serial I/O adapters. Tandy used it in a PC, but it had software compatibility problems with PC software due to the fact that the original IBM PC BIOS used interrupt vectors that Intel had reserved for their own use.
Mea navis aericumbens anguillis abundat
If you have ever written custom code for anything you recognize that truth. The code for the space station is essentially 'alpha' code. How could they get it to the 'beta' stage? Where would they get any 'users' to test it? Can any of you write millions of lines of alpha code with no errors?
The reason that nobody can write alpha code error free is the same reason that nobody can go out and shoot '18' for a round of golf; the job is too difficult for anyone to accomplish. That is why software requires several versions to get it right.
The computing section of the space station is far more extensive than any previous space flight. It was done that way because of the advantages that computer control brings. Because of Yin and Yang there is always a down side to anything which has an upside. The down side is that computer controlling everything necessarily increases the complexity of the computer code. With that complexity comes increased error problems. Sorry, that is the way that reality works.
There is one more truth - NASA has never managed a software project this complex for space use. As a result the management process has problems also.
Here is a management truth: nobody ever has enough time to spend doing the job right in the first place, but somehow they always find enough time to do the job over when their work breaks. In other words there always is enough time to do the job right . Doing it wrong and trying to fix the screw ups with kluges later always takes longer. That ought to be software management 101 - but it is something which most managers never understand.
The only way to solve the complexity issues that computer control brings is to do away with the computer controllers. That costs a lot more money and weighs a lot more. Either live with the problems that computers bring or live with the problems that not having computers bring.
I have to agree with the NASA veteran on the preparedness issue: it costs far less to be prepared at the start than to find out later that you weren't prepared.