WindBourne said, "As such we NEED multiple architectures... "
I agree with you. Yet it takes a look at history to see how we got here.
"The above will prevent Congress from doing what it is doing AND will prevent an accident in a rocket from shutting down the entire space program."
Your point here escapes me. Congress controls spending; that's in the Constitution. How could we prevent Congress from doing anything? Recently, Bush's space policy (go back to the Moon, etc) was a good one. However, Congress did not fund it. While I see you faulting "neo-cons", the truth is that the bean counters at OMB (Office of Management & Budget) have control. When the Republicans were in office, no funding. Now the Democrats are in office. No funding.
Obama has had a year to express interest in the space program, or even allocate keep-the-workforce-alive funding.
You wrote, "Nixon killed Skylab because he did not fund NASA properly for building the Shuttle after shutting down Apollo in 1970."
I came across a detailed and well written history of the Space Shuttle. It's a detailed history of the winding down of Apollo, the effort to fund a Mars mission, a space station, and many different incarnations of the Shuttle. The OMB constantly shows up in these decisions. I highly recommend this history.
As an example, here is a quote from that history concerning Nixon's decision to approve the Shuttle:
"Why, finally, did Nixon decide to build the Shuttle? One must not underestimate the tendency of the federal government to look after its own; few major Washington programs reach an end, to vanish into the night. Nixon had no wish to shut down piloted space flight; he wanted to keep it alive. He also was concerned over aerospace employment. Yet he could have addressed such issues with nothing more than Big Gemini riding atop a Titan III-M, to fly occasionally and show the flag.
The key to the Shuttle was its well-founded prospect of low cost and routine operation. This promise did not rest on the cost-benefit studies of Mathematica, which the Flax Committee largely refuted and the OMB rejected out of hand. Rather, it rested on technical developments: automated onboard checkout, reusable thermal protection, rocket engines with long life. No OMB internal memo or White House report ever denied this promise; only experience would do that, years later. The Shuttle thus could find its way to approval, within a nation and government that remained willing to embrace the new."
... The discussion is wobbling around Microsoft's coders and tools and why machine code isn't necessary anymore...
And here we were just discussing yet-another-chip-bug in an Intel chip which affects, as they say, 'three guesses and two don't count', yes, a Microsoft product. (Remember? Something about timer interrupts which can get stepped on by other timers, hanging the system. It's a known bug. In fact, there's 69 pages of known bugs in the i7 chip "errata" sheet.)
I would (gently) ask this: Which programmer is more likely to make the mistake of relying on this timer:
Programmer A, who learned C++ in college and is wizard at high level abstraction, or,
Programmer B, who has gone down to the bare metal and programmed chips?
Who is more likely to actually read the errata sheet that says, 'Don't use this frickin' timer'?
My answer is that people who have read chip datasheets (and errata sheets!) before are far more likely to read them again before trying to use that (broken) timer on the i7 chip.
Something else which really troubles me is how hard this bug may have been to find. Windows crashed? And the processor hung, not "leaving any tracks" in memory? I will just mention that this has happened to me a time or two (*cough, cough*). Just realizing it was a seperate and new bug was probably awful.
Now imagine you're the poor slob at Redmond who has to fix this! You're told: "Windows Server crashes. Fix it." What are you gonna do? If you don't have an in-circuit emulator that can "watch" the crash, you are in for some real debugging hell. I've been there.
Someone had to fix this problem and make a "hot patch". What language do you think they used, APL?
Someone still has to read the datasheets that chip manufacturers put out, in particular the North/Southbridge chipsets. The Linux people are getting really good at it.
People overclock RAM. They have to understand enough about SDRAM to do that. That's machine level clock-cycle stuff. They also overclock CPU's and use amazing cooling systems (anyone for liquid nitrogen?).
Back When Dinosaurs Ruled The Earth and I was in college, I was told that assembly was on its way out and a High Level Language, PASCAL, was going to be the standard. This was 1976-1980.
It's been awhile since I've even seen a PASCAL compiler, but the Visual C debugger shows me RAM, disassembly of opcodes, and all the Intel registers step-by-step... and man, does it ever help to know what the hell each assembly instruction does.
There's a time and place for high level abstraction. And there's a time and place for talking to the chips. I suspect that people's opinions about this are something like the C "brace wars" (the "proper" placement of "{}".)
Tom, you might want to look over the "Intel (r) Core (tm) i7-900 Desktop Processor Extreme Edition Series and Intel (r) Core (tm) i7-900 Desktop Processor Series".
It has around 69 pages of... well, I'll let Microsoft say it... the processors "may contain design defects or errors known as errata..."
Rather than "hate" or "like" NASA, there's a very well written history of how Apollo was compromised (Apollo 18-20 were cancelled), and how the Shuttle ended up with solid rocket boosters, at: http://history.nasa.gov/SP-4221/contents.htm .
It boils down to: This was what NASA could do with the money given them by the OMB (Office of Management and Budget).
I found this process to be quite interesting.
From a historical point of view, history seems to be repeating itself. Another "outside" committee has looked into NASA's plans and other options. The committee has priced the options for lowest possible, plod along, and even (remarkably) do something worthwhile.
"However, in these modern times with object-oriented multithreaded programming, BASIC is kind of a throwback, don't you think?"...
I would say that different computer languages tend to have different strengths. People are different and they naturally gravitate to different languages, over the dead bodies of their teachers, but it remains true. Knowing several computer languages never hurt anyone. And I'll happily mention that BASIC really does have some strengths. No, I don't think Linux should be written in BASIC. But --
I've found over time, that for me, BASIC can be really quite useful for whomping up little things, utilities that I need "Right Now, Just This Second". The syntax is friendly and easygoing, and it doesn't throw "avalanche fits". And if I mess up a line and something is wrong, figuring out what's wrong is not an exercise in puzzle solving, (not that I'm referring to "Macho C Coding" or anything [dark muttering] ).
Basic's fast enough for a lot of things. And I should mention that XBASIC, over at Sourceforge, has an astonishing, superb compiler that makes very, very fast 80x86 code indeed. If you'd like a pleasant surprise, give it a try.
I would say there's a place for BASIC. Don't underestimate the people how many people really like it. Possibly we should be dragging them, kicking, screaming, their fingernails digging into the wood, into C++ classes.
But possibly BASIC is all they have time to learn, and all they really -want- to learn, about computers. They may be heathens, but they are also programmers.
These are just my thoughts. Feel free to disagree with me; I welcome thoughts. I clearly learn the most from people who disagree with me!
I've had technical books published by various publishing houses.
What troubles me these days is seeing other author's books being scanned in and put up for download as PDF's up on the Net.
I have no problem with open-source programming if the programmer wants to.
But I think most programmers would be rather annoyed to find their source code published to the Net without their permission, and without any attempt to get their permission.
Authors feel the same way. Writing a book is generally a hell of a lot of work.
To me this seems a pretty easy answer once you look at the raw numbers.
http://en.wikipedia.org/wiki/Uranium-235 shows that when one atom of U-235, once fissioned, releases 202.5 MeV of heat. That's 202,500,000 electron volts, a.k.a., one huge amount of energy.
An atom of carbon when burned (C + 2 Ox->COx2) releases a few electron volts of energy and gives us carbon dioxide, which is said to be a "greenhouse gas". (I'm not debating that point).
Let's just do it with money, okay?
Hold an atom of carbon in one hand. Hold an atom of uranium in the other hand. The carbon's worth a few dollars. The uranium's worth Two Hundred Million Dollars. Which one do you pick? If you pick uranium, you just hit the Lotto Jackpot!
Bear in mind that you have to get enough of either to meet the energy needs of the country, and it's very hard to get enough coal, and much easier, by a factor of two hundred million, to get uranium.
Jimmy Carter made the unfortunate decision (funny how those words appear next to his name) not to include used fuel rods in reprocessing. There's a lot of energy there awaiting.
I think what we ought to do as a country is swallow some pride, go to France, which gets about 80% of its energy from nuclear, and say, "Obviously you have a well debugged design. Help! Show us how to do it!" The French do it right. You know how useful debugged code is.
This is the end of Manned Space Flight for the United States.
If you, or your children, want to see the launch of a manned American bird, you'd better do it immediately, in the remainder of 2009 or 2010. You have nine Shuttle launches left.
NASA has been promised funding sufficient for nine more Shuttle missions to finish the ISS, and that is all. (2009 & 2010). This is essentially funding for the coffin and headstone.
One year from now, we will be back to 1957. The Russians will still have their heavy-lift rockets and Soyuz capsules, and we won't have anything. We'll have to beg/buy rides from Russia to get to the ISS. Should the Russians decide they're not in the mood, they can easily say: "And what are you going to do about it?".
The Russians do not have a good "Partners in Space" history. Remember when the ISS was very nearly lost to [Russian] computer problems, the Russians were swift to blame the Americans for "bad power". When the Soyuz's explosive bolts failed, the Russians blamed the ISS for mysterious "plasma discharges" that, somehow (don't ask me how) damaged explosive bolts.
Now NASA is trying to build its "Ares-1" rocket. Two stages. Its lower stage is a Shuttle SRB (Solid Rocket Booster) with an extra solid segment, which means higher pressures, more joints, and this time the vibration is directly transmitted to the backs of the crew.
(I wish someone would tell me why we're not just using the Saturn V main booster technology for the first stage. It has the advantage of being repeatedly tested and it works. Is it just too easy or something?)
It's revealing to see what former NASA officials say when they are not muzzled anymore:
Former NASA Administrator Mike Griffin said in a recent speech the projected funding shortfalls threaten America's leadership in manned space flight.
"In a democracy, the proper purpose of the OMB (Office of Management and Budget) is not to find a way to create a Potemkin Village at NASA. It is not to create the appearance of having a real space program without having to pay for it. It is not to specify to NASA how much money shall be allocated for human lunar return by 2020. The proper purpose of the OMB is to work with NASA, as a partner in good government, to craft carefully vetted estimates of what is required to achieve national policy goals. The judgment as to whether the stated goals are too costly, or not, is one to be made by the nation's elected leadership, not career civil service staff.
No one can wrest leadership in space from the United States. We're that good. But we can certainly cede it, and that is the path we are on."
Something like the Ares-1 design has happened before; see the history of the Space Shuttle:
You will find that what Congress/OMB and Obama are doing ia just about precisely what Congress/BoB and Nixon did.
Funding was cut right through the fat, muscle, and into the bone, and the only way NASA could figure out to launch anything was using solid rocket boosters (SRB's), which add significant vibration and which cannot be throttled, or switched off and reignited. Many engineers left NASA when the decision to use SRB's came down.
A stunning quote from one Douglas Cookie, 'Associate Administrator for the Exploration Systems Mission Directorate at NASA HQ', is:
[Cooke] said the biggest single reason he has favored Ares 1 over competing designs is _crew safety_.
"That's in part due to the Ares 1 first stage, an extended five-segment shuttle booster. Two four-segment boosters are used for every shuttle flight and in 125 missions to date - 250 booster flights - only one booster has ever failed, the one that doomed Challenger.
Sure, I "get" the rational answers about clearing your mind and letting your subconscious work on problems and stuff. Then there those other problems. When the network is down for no discernible reason. When it's a "bad day".
Here's a very weird bit of magic to know about. I know how weird this is, but it works. Oh, ye Gods, does it work! I have come to absolutely hate doing it, but, it just [dammit] works.
You won't believe it until you try it, and then you, too, will (a) hate it and (b) see it work.
Get Neil Young's "Live Rust" album. It's live, so studio engineers haven't fixed up all the mistakes.
Instead of hitting notes, Neil... shotguns notes, all around them. For some awful reason he writes songs on the very wobbling top edge of his vocal range. He's just one of those people who should be banned from singing the songs he writes.
The lyrics are really depressing. Cortez is massacring the Mayans. We got Mother Nature on the run. It's better to burn out than to fade away. And his favorite song topic, band people dying of overdoses: Every junkie's like a setting sun. Heard that he died, out on a mainline.
(Neil has not made a 2+2 = ? connection yet. If I were in his band and I had to listen to him every night, both (1) heroin and (2) death would look pretty good.)
I discovered this by accident while coding keyboard and mouse drivers. They were extremely hard. I was desperate. I saw one of those skiier T-shirts that says, "No Pain, No Gain", and I logically inverted it to "If Pain, Gain". I looked through my record albums to find True Pain and there was Neil. Put it on side 3 Repeat and started coding.
Six plus hours and 1,500 lines of 68000 assembly later, I was done coding. I finally ran the AS68 assembler on it, and it said back to me: 0 Errors.
Floating in a sea of shocked surprise, I linked it and ran it, and my God, the keyboard and mouse code worked the first time.
This does not happen. As Luke Skywalker put it, "That's Impossible!" I've been terrified to touch that code.
The Neil effect has continued to work.
Last time I used Neil was on a laptop that WinXP would not install onto. Tried 3 times, the install hung each time. The guy's son Anthony said, "Dave, why aren't you playing Neil Young?" Good point. I put it on. I explained this to the guy. He looked at me like I was insane. I said, "You don't believe it right now. That's okay. You'll see it. Then you'll believe." With Neil wailing away about his first cigarette, and changing absolutely nothing, the XP install went without a hitch.
All I can say is I tripped over it by accident, I don't know how it works, but damn!, it does work. And you'll start to dislike that album as much as I do. Maybe the cost is listening to something you really don't care for. (I should try rap!)
I tested hundreds of Mac software applications 'Back In The Day'. A bit more than half of all of them banged onto location 0. About 100% of Microsoft Mac apps crashed. This happened to not (immediately) crash the Mac because the Mac just happened to have RAM mapped there, but, since multiple applications were resident in memory, taking turns shooting at 0.Long, "Difficult to debug problems, often not repeatable, soon arose". [I do love that phrase; it covers so much agony so concisely].
The Mac had a memory heap structure with "handles", which aimed at "pointers", which aimed at "memory blocks". From time to time, just to amuse developers, the Mac would compress memory to get rid of stale stuff. The handles would be updated but, a zero would go into the pointer you were trusting...
I was busy teaching the Atari ST 68000 computer to be a Mac at the time. The Atari ST had ROM mapped to locations 0-7. It bus-errored every time someone used a Nil pointer. This looked unsolvable unless I wanted to patch two zillion Mac programs.
I finally looked at the problem in a very, very simple manner, and discovered the IW (instruction word) that had caused the crash was on the stack, and the PC (program counter) was a bit past the crash point due to prefetch. A dim flickering flight appeared over my head. What if I "back up" the PC to the Evil Instruction, disassemble it to get its length [I hear you sigh at the thought of writing a disassembler, but I did], advance the PC to the next instruction (so we're just skipping over "the store into 0.Long" here), put all this on the stack, and treat it like a "normal" exception interrupt... "Return from Exception."
And it worked! The 68000 came back from a crash!
That one thing made all the difference.
To this day I have wondered, though... if the Atari ST had ROM there... what happened to all the data written to location 0?
Well, depending on the reaction (D+D, D+T, T+T), you get different levels of energy, but in general, nearly all of a fusion reaction's energy goes into making very, very energetic neutrons. At above 1 MeV these neutrons can split even "depleted" uranium or U-238. That's why half the energy from a "fusion" bomb is from fissioning the uranium wrapper, and the fission amount has been as large as 7.5 MT (Castle Bravo).
I haven't seen data on what very energetic neutrons do to isotopes like U-233 or other junk left after U-235 fission; that's one detail the article doesn't go into. I'd sure like to know.
Certainly the heat generated could spin a turbine. Which leads me to ask:
"Spent" fuel casks are typically stored underwater for shielding. If they're as radioactive as stated, they really warm the water up. Why isn't this used to generate electricity?
The US was a rich country once. We aren't anymore. Making decisions like "no more reactors" got us here.
The basic deal here is that each U-235 atom that splits yield 170 MeV of energy; 1.7 x 10^8. The average chemical reaction yields a few eV. (Source: Los Alamos Primer, Robert Serber)
As a country we can't afford to throw away fuel which is that much more powerful. It's not only stupid, it's cruel.
For example, we have people working in coal mines because we won't build nuclear reactors. Coal mines are not very safe at all.
About 20% of our electricity now comes from 1st generation nuclear reactors. They're being run past their design life. We need to build more just to replace the old ones, and they need to be 1000 Megawatt size. We really should ask the French, because they have had a lot of success in this area.
People are talking about "conservation". They've been talking about that since the first Earth Day, which I remember. When you add a lot more people, you still need more electricity. Without electricity, we go back a hundred years as a civilization.
It's time to make the tough choices instead of letting them go to the next generation by taking no action.
Well of course the first thing that John Coster-Mullen would have to do is determine what microprocessor family was used on Little Boy. Is it an 8051, Z80, 68000, x386, PIC ? Generally even on OEM parts there's a basic part number. On DIP ROMS there has been a real effort to stay consistent.
John should bear in mind that the working timing of a nuclear bomb is 10 nanoseconds. This makes his microprocessor choices more limited; he'll need some speed here.
Then he needs to get to the code and examine it. If it's in a separate ROM chip this is much easier. For example, I was looking at a flashable EEPROM on a PCI card just yesterday evening and found it had a 29F001-TPC in a nice 32 pin DIP package. 128K x 8 bytes and nice little extras like auto-protection. It's a 120 ns part (from address setup to output enable* and the data pins settled.)
"Little Boy" was not a very efficient bomb at all. It was heavily overengineered. I can remember Ted Taylor calling it a "committee bomb" and a "stupid bomb" in McPhee's book. The committee wanted to make sure it went off and didn't just dig a hole, and hand the Japanese more than a hundred pounds of U-235. If I recall right, and this is just off the top of my head, it yielded only 12.5 kilotons or so.
Robert Serber's "Los Alamos Primer" is really useful here. It takes 10 nanoseconds (one "shake") for a fission to happen ("neutron multiplication time"). "The direct energy release per atom is 170 MeV".
"The energy release of TNT is ~~ 4 x 10^10 erg / gram, or ~~ 3.6 x 10^16 erg / ton.... Hence, 1 kg of U-235 completely fissioned yields about 20,000 tons of TNT equivalent, or ~~ 20 kilotons." [Serber is referring to tons as 2,000 lb tons.]
From this you can see that with a 12.5 kiloton yield, about two thirds of 1 kg of U-235 fissioned, so more or less, plus or minus, about a pound fissioned. Wikipedia says the bomb had 64 kg of uranium and only 600 mg. actually fissioned. I don't understand that number.
To which I can only say, see what happens when you use embedded Windows instead of Linux to drive your timing signals? I can just see it now. The bomb releases, starts falling, and in its microprocessor, it draws The Blue Screen Of Death and halts.
Thanks,
Dave Small
p.s. I really don't know why I wrote this. Maybe it was the coffee this morning.
The high altitude nuclear shots the U.S. did to study the effects (and try for an effective ABM) did indeed kill some early commercial satellites.
This was most embarrassing, even if it was an "effect" to study.
There was an informal understanding between the Soviet and U.S. space program that no one would fire off a megaton at high altitude when there was a spacecraft and crew up there.
I wasn't around for most of the 1950's, but I believe we also managed to EMP Hawaii dark, or at least kill a whole bunch of fuses. But in the 1950's we didn't use much transistor technology; we were busy learning about it and developing it. (Please, I know that some transistorized machines came out in this era; I'm saying we didn't use them to control main power switches.)
"Oh, please don't turn this into a zombie apocalypse survivalist fantasy! "
No! No! DO turn it into a Zombie Apocalypse Survivalist Fantasy!!! It'll be a heck of a lot more fun to read!
If I may put a vote in... go find the movie "Fido". Easily the funniest Zombie movie I've ever seen. Stars Carrie-Ann Moss. Yep, that girl from the Matrix... and she can actually smile! Set in 1950's America after a Zombie Apocalypse... with a Zombie and a boy named Timmy...
-- David
This may look like a signature, but it's only yet another buffer overrun.
I've had very good luck with a couple 6 volt "lantern" batteries, a standard 4-pin disk drive power connector, and a few alligator clips. Connect the 6 V batteries + to -. On the "+" battery, connect its minus to the center two black "ground" leads. Connect the "+" of the "+" battery to the +5 line (red) going to the drive. Finally, connect the "+" of the other battery to the +12V (yellow) wire. So we're supplying ground, +6, and +12. Heck, it has a voltage regulator on +5... and it doesn't have long to live anyway.
Plug the drive in. Spin it up. IMPORTANT: Put it so the platters are at 90 degrees to you! In other words, put it so that if the platters explode, they won't hit you. You don't want it sitting on the ground with the platters parallel to the ground; you want them spinning at 90 degrees to the ground and to you.
Hike back a bit and shoot with a heavy slug. A slug from a 12 gauge works fine. All the rotational (angular) momentum is transferred to the frame as the platters stop almost instantly. The drive goes whirl-whirl-whirl, sometimes up in the air!
I found that the.223 round from an M16A1 was strangely ineffective, but going to full auto made me feel better, and that's the point of all this. [ Yes it's legal to own an M16A1 here if you hop through the paperwork hoops]. Alas, the M16 was a toy I got rid of later...
9 mm works very well at disassembling the drives, as does.45 cal.
By the way, there are **absolutely terrific** magnets inside modern drives. Open 'em up and use a little acetone to unglue the magnets. They're the "supermagnets". And if you disassemble the drive with shooting, it is commonly opened FOR you!
Shooting a disk drive that crashed and lost you a bunch of work is terrific stuff. In Las Vegas there are a couple places that let you bring in your computer or whatever and shoot it with automatic weapons. Hmmmm, maybe I should try an AK-47 someday, that's a.30 cal round...
Take a million car batteries. Yup, lead acid. Why a million car batteries? Because they're so cheap -- we made a hundred million of them last year, and they're one of the best recycling stories out there). Don't be clever. New tech means R&D, and that means unexpected surprises.
Store them in buildings which are above 50 degrees F so that they last a very long time (and certainly not conditions under a hood). Some place near a large interconnect.
Use an inverter, or, just hook the batteries up with IGBT switches to "thermometer" up the voltage, and then back down, making AC.
I'm aware that this is limited storage; the batteries don't like to run at load capacity for long. But also note that Fairbanks has, I believe, 25 megawatts of battery backup.
I'll throw this into the "fresh meat" bowl here...
Thank you very much for pointing me at good programs on the magnetic field reversal. I appreciate it.
I do know that in my lifetime, the amount in degrees that has to be (added/subtracted) from "Magnetic North" to get "True North" has drifted several degrees, for here in my home state of Colorado. I can remember my surprise on reading that!
It's always disturbing to me when something that I tended to regard as taking "geologic time" is now running at, well, "human time" -- this speed.
SSTO (Single Stage To Orbit) is absolutely the way to go. Gary Hudson has been pushing that idea for some time. Go up, come down, refuel, and go back up.
The Rotary Rocket idea didn't work out for unforeseen reasons; this is why we have "research".
It's a shocker, but implementing a fleet of SSTO's would be cheap and would work . . .
which means our government isn't interested! *grin*
To "slashdot". Verb. 1) To immediately shift the discussion 90 degrees from any known frame of reference in space-time. e.g., "Hey, you just slashdotted my proposal!" 2) To devolve into arguing over something 90 degrees from any known frame of reference in space-time. e.g., "I know we were talking about Microsoft patenting the Page Up key, but dammit!, you can't express that Sanskrit character properly without patching the X Server with grep,sed,unk,dwork,glink, and you must be running a kernel later than 2.26.0!" 3) To perform steps (1) and (2) so rapidly that missing one reply will leave the reader 90 degrees from any known frame of reference. 4) The hell you say! What really pisses me off about the Mac are the Microsoft commercials are so incredibly lame, they're no real competition for the Mac commercials. In fact, it must be a conspiracy, started by SCO, the RIAA, the MPAA, under the DMCA, and many, many other four letter acronyms.
Slashdot Mirror
WindBourne said, ... "
"As such we NEED multiple architectures
I agree with you. Yet it takes a look at history to see how we got here.
"The above will prevent Congress from doing what it is doing AND will prevent an accident in a rocket from shutting down the entire space program."
Your point here escapes me. Congress controls spending; that's in the Constitution. How could we prevent Congress from doing anything? Recently, Bush's space policy (go back to the Moon, etc) was a good one. However, Congress did not fund it. While I see you faulting "neo-cons", the truth is that the bean counters at OMB (Office of Management & Budget) have control. When the Republicans were in office, no funding. Now the Democrats are in office. No funding.
Obama has had a year to express interest in the space program, or even allocate keep-the-workforce-alive funding.
You wrote, "Nixon killed Skylab because he did not fund NASA properly for building the Shuttle after shutting down Apollo in 1970."
I came across a detailed and well written history of the Space Shuttle. It's a detailed history of the winding down of Apollo, the effort to fund a Mars mission, a space station, and many different incarnations of the Shuttle. The OMB constantly shows up in these decisions. I highly recommend this history.
You can find it at:
http://history.nasa.gov/SP-4221/contents.htm
As an example, here is a quote from that history concerning Nixon's decision to approve the Shuttle:
"Why, finally, did Nixon decide to build the Shuttle? One must not underestimate the tendency of the federal government to look after its own; few major Washington programs reach an end, to vanish into the night. Nixon had no wish to shut down piloted space flight; he wanted to keep it alive. He also was concerned over aerospace employment. Yet he could have addressed such issues with nothing more than Big Gemini riding atop a Titan III-M, to fly occasionally and show the flag.
The key to the Shuttle was its well-founded prospect of low cost and routine operation. This promise did not rest on the cost-benefit studies of Mathematica, which the Flax Committee largely refuted and the OMB rejected out of hand. Rather, it rested on technical developments: automated onboard checkout, reusable thermal protection, rocket engines with long life. No OMB internal memo or White House report ever denied this promise; only experience would do that, years later. The Shuttle thus could find its way to approval, within a nation and government that remained willing to embrace the new."
Thanks,
David Small
You can buy Gadgets online?
Errrr... Ummm...
Why didn't anybody tell... me?
Thanks,
Dave Small
(formerly of Gadgets by Small)
... The discussion is wobbling around Microsoft's coders and tools and why machine code isn't necessary anymore ...
And here we were just discussing yet-another-chip-bug in an Intel chip which affects, as they say, 'three guesses and two don't count', yes, a Microsoft product. (Remember? Something about timer interrupts which can get stepped on by other timers, hanging the system. It's a known bug. In fact, there's 69 pages of known bugs in the i7 chip "errata" sheet.)
I would (gently) ask this: Which programmer is more likely to make the mistake of relying on this timer:
Programmer A, who learned C++ in college and is wizard at high level abstraction, or,
Programmer B, who has gone down to the bare metal and programmed chips?
Who is more likely to actually read the errata sheet that says, 'Don't use this frickin' timer'?
My answer is that people who have read chip datasheets (and errata sheets!) before are far more likely to read them again before trying to use that (broken) timer on the i7 chip.
Something else which really troubles me is how hard this bug may have been to find. Windows crashed? And the processor hung, not "leaving any tracks" in memory? I will just mention that this has happened to me a time or two (*cough, cough*). Just realizing it was a seperate and new bug was probably awful.
Now imagine you're the poor slob at Redmond who has to fix this! You're told: "Windows Server crashes. Fix it." What are you gonna do? If you don't have an in-circuit emulator that can "watch" the crash, you are in for some real debugging hell. I've been there.
Someone had to fix this problem and make a "hot patch". What language do you think they used, APL?
Someone still has to read the datasheets that chip manufacturers put out, in particular the North/Southbridge chipsets. The Linux people are getting really good at it.
People overclock RAM. They have to understand enough about SDRAM to do that. That's machine level clock-cycle stuff. They also overclock CPU's and use amazing cooling systems (anyone for liquid nitrogen?).
Back When Dinosaurs Ruled The Earth and I was in college, I was told that assembly was on its way out and a High Level Language, PASCAL, was going to be the standard. This was 1976-1980.
It's been awhile since I've even seen a PASCAL compiler, but the Visual C debugger shows me RAM, disassembly of opcodes, and all the Intel registers step-by-step ... and man, does it ever help to know what the hell each assembly instruction does.
There's a time and place for high level abstraction. And there's a time and place for talking to the chips. I suspect that people's opinions about this are something like the C "brace wars" (the "proper" placement of "{}".)
Enjoy --
Dave Small
Tom, you might want to look over the
"Intel (r) Core (tm) i7-900 Desktop Processor Extreme Edition Series and Intel (r) Core (tm) i7-900 Desktop Processor Series".
It has around 69 pages of ... well, I'll let Microsoft say it ... the processors "may contain design defects or errors known as errata..."
It's dated November 2009.
Here's a URL:
http://download.intel.com/design/processor/specupdt/320836.pdf
Thanks, Dave
Rather than "hate" or "like" NASA, there's a very well written history of how Apollo was compromised (Apollo 18-20 were cancelled), and how the Shuttle ended up with solid rocket boosters, at: http://history.nasa.gov/SP-4221/contents.htm .
It boils down to: This was what NASA could do with the money given them by the OMB (Office of Management and Budget).
I found this process to be quite interesting.
From a historical point of view, history seems to be repeating itself. Another "outside" committee has looked into NASA's plans and other options. The committee has priced the options for lowest possible, plod along, and even (remarkably) do something worthwhile.
I assume "lowest possible" will be picked.
I recommend the above hyperlink to anyone.
Thanks,
Dave
...quote from parent
"However, in these modern times with object-oriented multithreaded programming, BASIC is kind of a throwback, don't you think?"...
I would say that different computer languages tend to have different strengths. People are different and they naturally gravitate to different languages, over the dead bodies of their teachers, but it remains true. Knowing several computer languages never hurt anyone. And I'll happily mention that BASIC really does have some strengths. No, I don't think Linux should be written in BASIC. But --
I've found over time, that for me, BASIC can be really quite useful for whomping up little things, utilities that I need "Right Now, Just This Second". The syntax is friendly and easygoing, and it doesn't throw "avalanche fits". And if I mess up a line and something is wrong, figuring out what's wrong is not an exercise in puzzle solving, (not that I'm referring to "Macho C Coding" or anything [dark muttering] ).
Basic's fast enough for a lot of things. And I should mention that XBASIC, over at Sourceforge, has an astonishing, superb compiler that makes very, very fast 80x86 code indeed. If you'd like a pleasant surprise, give it a try.
I would say there's a place for BASIC. Don't underestimate the people how many people really like it. Possibly we should be dragging them, kicking, screaming, their fingernails digging into the wood, into C++ classes.
But possibly BASIC is all they have time to learn, and all they really -want- to learn, about computers. They may be heathens, but they are also programmers.
These are just my thoughts. Feel free to disagree with me; I welcome thoughts. I clearly learn the most from people who disagree with me!
Thanks,
David Small
davetracer at-sign aol DOT com
I've had technical books published by various publishing houses.
What troubles me these days is seeing other author's books being scanned in and put up for download as PDF's up on the Net.
I have no problem with open-source programming if the programmer wants to.
But I think most programmers would be rather annoyed to find their source code published to the Net without their permission, and without any attempt to get their permission.
Authors feel the same way. Writing a book is generally a hell of a lot of work.
-- thanks,
Dave Small
A prayer wheel application ...
Just the ticket for the on-the-go iPhoner.
(flip) (flip) (flip) (flip)
-- D
To me this seems a pretty easy answer once you look at the raw numbers.
http://en.wikipedia.org/wiki/Uranium-235 shows that when one atom of U-235, once fissioned, releases 202.5 MeV of heat. That's 202,500,000 electron volts, a.k.a., one huge amount of energy.
An atom of carbon when burned (C + 2 Ox->COx2) releases a few electron volts of energy and gives us carbon dioxide, which is said to be a "greenhouse gas". (I'm not debating that point).
Let's just do it with money, okay?
Hold an atom of carbon in one hand. Hold an atom of uranium in the other hand. The carbon's worth a few dollars. The uranium's worth Two Hundred Million Dollars. Which one do you pick? If you pick uranium, you just hit the Lotto Jackpot!
Bear in mind that you have to get enough of either to meet the energy needs of the country, and it's very hard to get enough coal, and much easier, by a factor of two hundred million, to get uranium.
Jimmy Carter made the unfortunate decision (funny how those words appear next to his name) not to include used fuel rods in reprocessing. There's a lot of energy there awaiting.
I think what we ought to do as a country is swallow some pride, go to France, which gets about 80% of its energy from nuclear, and say, "Obviously you have a well debugged design. Help! Show us how to do it!" The French do it right. You know how useful debugged code is.
Thanks,
Dave Small
This is terribly sad.
This is the end of Manned Space Flight for the United States.
If you, or your children, want to see the launch of a manned American bird, you'd better do it immediately, in the remainder of 2009 or 2010. You have nine Shuttle launches left.
NASA has been promised funding sufficient for nine more Shuttle missions to finish the ISS, and that is all. (2009 & 2010). This is essentially funding for the coffin and headstone.
One year from now, we will be back to 1957. The Russians will still have their heavy-lift rockets and Soyuz capsules, and we won't have anything. We'll have to beg/buy rides from Russia to get to the ISS. Should the Russians decide they're not in the mood, they can easily say: "And what are you going to do about it?".
The Russians do not have a good "Partners in Space" history. Remember when the ISS was very nearly lost to [Russian] computer problems, the Russians were swift to blame the Americans for "bad power". When the Soyuz's explosive bolts failed, the Russians blamed the ISS for mysterious "plasma discharges" that, somehow (don't ask me how) damaged explosive bolts.
Now NASA is trying to build its "Ares-1" rocket. Two stages. Its lower stage is a Shuttle SRB (Solid Rocket Booster) with an extra solid segment, which means higher pressures, more joints, and this time the vibration is directly transmitted to the backs of the crew.
(I wish someone would tell me why we're not just using the Saturn V main booster technology for the first stage. It has the advantage of being repeatedly tested and it works. Is it just too easy or something?)
It's revealing to see what former NASA officials say when they are not muzzled anymore:
Former NASA Administrator Mike Griffin said in a recent speech the projected funding shortfalls threaten America's leadership in manned space flight.
"In a democracy, the proper purpose of the OMB (Office of Management and Budget) is not to find a way to create a Potemkin Village at NASA. It is not to create the appearance of having a real space program without having to pay for it. It is not to specify to NASA how much money shall be allocated for human lunar return by 2020. The proper purpose of the OMB is to work with NASA, as a partner in good government, to craft carefully vetted estimates of what is required to achieve national policy goals. The judgment as to whether the stated goals are too costly, or not, is one to be made by the nation's elected leadership, not career civil service staff.
No one can wrest leadership in space from the United States. We're that good. But we can certainly cede it, and that is the path we are on."
Something like the Ares-1 design has happened before; see the history of the Space Shuttle:
http://history.nasa.gov/SP-4221/contents.htm
You will find that what Congress/OMB and Obama are doing ia just about precisely what Congress/BoB and Nixon did.
Funding was cut right through the fat, muscle, and into the bone, and the only way NASA could figure out to launch anything was using solid rocket boosters (SRB's), which add significant vibration and which cannot be throttled, or switched off and reignited. Many engineers left NASA when the decision to use SRB's came down.
A stunning quote from one Douglas Cookie, 'Associate Administrator for the Exploration Systems Mission Directorate at NASA HQ', is:
[Cooke] said the biggest single reason he has favored Ares 1 over competing designs is _crew safety_.
"That's in part due to the Ares 1 first stage, an extended five-segment shuttle booster. Two four-segment boosters are used for every shuttle flight and in 125 missions to date - 250 booster flights - only one booster has ever failed, the one that doomed Challenger.
"I personally believe the risk is l
Sure, I "get" the rational answers about clearing your mind and letting your subconscious work on problems and stuff. Then there those other problems. When the network is down for no discernible reason. When it's a "bad day".
Here's a very weird bit of magic to know about. I know how weird this is, but it works. Oh, ye Gods, does it work! I have come to absolutely hate doing it, but, it just [dammit] works.
You won't believe it until you try it, and then you, too, will (a) hate it and (b) see it work.
Get Neil Young's "Live Rust" album. It's live, so studio engineers haven't fixed up all the mistakes.
Instead of hitting notes, Neil... shotguns notes, all around them. For some awful reason he writes songs on the very wobbling top edge of his vocal range. He's just one of those people who should be banned from singing the songs he writes.
The lyrics are really depressing. Cortez is massacring the Mayans. We got Mother Nature on the run. It's better to burn out than to fade away. And his favorite song topic, band people dying of overdoses: Every junkie's like a setting sun. Heard that he died, out on a mainline.
(Neil has not made a 2+2 = ? connection yet. If I were in his band and I had to listen to him every night, both (1) heroin and (2) death would look pretty good.)
I discovered this by accident while coding keyboard and mouse drivers. They were extremely hard. I was desperate. I saw one of those skiier T-shirts that says, "No Pain, No Gain", and I logically inverted it to "If Pain, Gain". I looked through my record albums to find True Pain and there was Neil. Put it on side 3 Repeat and started coding.
Six plus hours and 1,500 lines of 68000 assembly later, I was done coding. I finally ran the AS68 assembler on it, and it said back to me: 0 Errors.
YouHaveGotToBeKiddingMeTheresAlwaysErrors. TyposIfNothingElse.
Floating in a sea of shocked surprise, I linked it and ran it, and my God, the keyboard and mouse code worked the first time.
This does not happen. As Luke Skywalker put it, "That's Impossible!" I've been terrified to touch that code.
The Neil effect has continued to work.
Last time I used Neil was on a laptop that WinXP would not install onto. Tried 3 times, the install hung each time. The guy's son Anthony said, "Dave, why aren't you playing Neil Young?" Good point. I put it on. I explained this to the guy. He looked at me like I was insane. I said, "You don't believe it right now. That's okay. You'll see it. Then you'll believe." With Neil wailing away about his first cigarette, and changing absolutely nothing, the XP install went without a hitch.
All I can say is I tripped over it by accident, I don't know how it works, but damn!, it does work. And you'll start to dislike that album as much as I do. Maybe the cost is listening to something you really don't care for. (I should try rap!)
I'm not making this up.
-- Dave Small
The article quotes Mark Studness from Verizon.
Mark Studness ?? !!
"Is someone playing silly buggers here?"
-- Archchancellor Ridcully
I tested hundreds of Mac software applications 'Back In The Day'. A bit more than half of all of them banged onto location 0. About 100% of Microsoft Mac apps crashed. This happened to not (immediately) crash the Mac because the Mac just happened to have RAM mapped there, but, since multiple applications were resident in memory, taking turns shooting at 0.Long, "Difficult to debug problems, often not repeatable, soon arose".
[I do love that phrase; it covers so much agony so concisely].
The Mac had a memory heap structure with "handles", which aimed at "pointers", which aimed at "memory blocks". From time to time, just to amuse developers, the Mac would compress memory to get rid of stale stuff. The handles would be updated but, a zero would go into the pointer you were trusting ...
I was busy teaching the Atari ST 68000 computer to be a Mac at the time. The Atari ST had ROM mapped to locations 0-7. It bus-errored every time someone used a Nil pointer. This looked unsolvable unless I wanted to patch two zillion Mac programs.
I finally looked at the problem in a very, very simple manner, and discovered the IW (instruction word) that had caused the crash was on the stack, and the PC (program counter) was a bit past the crash point due to prefetch. A dim flickering flight appeared over my head. What if I "back up" the PC to the Evil Instruction, disassemble it to get its length [I hear you sigh at the thought of writing a disassembler, but I did], advance the PC to the next instruction (so we're just skipping over "the store into 0.Long" here), put all this on the stack, and treat it like a "normal" exception interrupt ... "Return from Exception."
And it worked! The 68000 came back from a crash!
That one thing made all the difference.
To this day I have wondered, though ... if the Atari ST had ROM there ... what happened to all the data written to location 0?
*grin*
David Small
Well, depending on the reaction (D+D, D+T, T+T), you get different levels of energy, but in general, nearly all of a fusion reaction's energy goes into making very, very energetic neutrons. At above 1 MeV these neutrons can split even "depleted" uranium or U-238. That's why half the energy from a "fusion" bomb is from fissioning the uranium wrapper, and the fission amount has been as large as 7.5 MT (Castle Bravo).
I haven't seen data on what very energetic neutrons do to isotopes like U-233 or other junk left after U-235 fission; that's one detail the article doesn't go into. I'd sure like to know.
Certainly the heat generated could spin a turbine. Which leads me to ask:
"Spent" fuel casks are typically stored underwater for shielding. If they're as radioactive as stated, they really warm the water up. Why isn't this used to generate electricity?
-- Thanks, Dave
The US was a rich country once. We aren't anymore. Making decisions like "no more reactors" got us here.
The basic deal here is that each U-235 atom that splits yield 170 MeV of energy; 1.7 x 10^8. The average chemical reaction yields a few eV. (Source: Los Alamos Primer, Robert Serber)
As a country we can't afford to throw away fuel which is that much more powerful. It's not only stupid, it's cruel.
For example, we have people working in coal mines because we won't build nuclear reactors. Coal mines are not very safe at all.
About 20% of our electricity now comes from 1st generation nuclear reactors. They're being run past their design life. We need to build more just to replace the old ones, and they need to be 1000 Megawatt size. We really should ask the French, because they have had a lot of success in this area.
People are talking about "conservation". They've been talking about that since the first Earth Day, which I remember. When you add a lot more people, you still need more electricity. Without electricity, we go back a hundred years as a civilization.
It's time to make the tough choices instead of letting them go to the next generation by taking no action.
-- Dave
Well of course the first thing that John Coster-Mullen would have to do is determine what microprocessor family was used on Little Boy. Is it an 8051, Z80, 68000, x386, PIC ? Generally even on OEM parts there's a basic part number. On DIP ROMS there has been a real effort to stay consistent.
John should bear in mind that the working timing of a nuclear bomb is 10 nanoseconds. This makes his microprocessor choices more limited; he'll need some speed here.
Then he needs to get to the code and examine it. If it's in a separate ROM chip this is much easier. For example, I was looking at a flashable EEPROM on a PCI card just yesterday evening and found it had a 29F001-TPC in a nice 32 pin DIP package. 128K x 8 bytes and nice little extras like auto-protection. It's a 120 ns part (from address setup to output enable* and the data pins settled.)
"Little Boy" was not a very efficient bomb at all. It was heavily overengineered. I can remember Ted Taylor calling it a "committee bomb" and a "stupid bomb" in McPhee's book. The committee wanted to make sure it went off and didn't just dig a hole, and hand the Japanese more than a hundred pounds of U-235. If I recall right, and this is just off the top of my head, it yielded only 12.5 kilotons or so.
Robert Serber's "Los Alamos Primer" is really useful here. It takes 10 nanoseconds (one "shake") for a fission to happen ("neutron multiplication time"). "The direct energy release per atom is 170 MeV".
"The energy release of TNT is ~~ 4 x 10^10 erg / gram, or ~~ 3.6 x 10^16 erg / ton.... Hence, 1 kg of U-235 completely fissioned yields about 20,000 tons of TNT equivalent, or ~~ 20 kilotons." [Serber is referring to tons as 2,000 lb tons.]
From this you can see that with a 12.5 kiloton yield, about two thirds of 1 kg of U-235 fissioned, so more or less, plus or minus, about a pound fissioned. Wikipedia says the bomb had 64 kg of uranium and only 600 mg. actually fissioned. I don't understand that number.
To which I can only say, see what happens when you use embedded Windows instead of Linux to drive your timing signals? I can just see it now. The bomb releases, starts falling, and in its microprocessor, it draws The Blue Screen Of Death and halts.
Thanks,
Dave Small
p.s. I really don't know why I wrote this. Maybe it was the coffee this morning.
The high altitude nuclear shots the U.S. did to study the effects (and try for an effective ABM) did indeed kill some early commercial satellites.
This was most embarrassing, even if it was an "effect" to study.
There was an informal understanding between the Soviet and U.S. space program that no one would fire off a megaton at high altitude when there was a spacecraft and crew up there.
I wasn't around for most of the 1950's, but I believe we also managed to EMP Hawaii dark, or at least kill a whole bunch of fuses. But in the 1950's we didn't use much transistor technology; we were busy learning about it and developing it. (Please, I know that some transistorized machines came out in this era; I'm saying we didn't use them to control main power switches.)
Thanks,
Dave
"Oh, please don't turn this into a zombie apocalypse survivalist fantasy! "
No! No! DO turn it into a Zombie Apocalypse Survivalist Fantasy!!! It'll be a heck of a lot more fun to read!
If I may put a vote in ... go find the movie "Fido". Easily the funniest Zombie movie I've ever seen. Stars Carrie-Ann Moss. Yep, that girl from the Matrix ... and she can actually smile! Set in 1950's America after a Zombie Apocalypse ... with a Zombie and a boy named Timmy ...
-- David
This may look like a signature, but it's only yet another buffer overrun.
I've had very good luck with a couple 6 volt "lantern" batteries, a standard 4-pin disk drive power connector, and a few alligator clips. Connect the 6 V batteries + to -. On the "+" battery, connect its minus to the center two black "ground" leads. Connect the "+" of the "+" battery to the +5 line (red) going to the drive. Finally, connect the "+" of the other battery to the +12V (yellow) wire. So we're supplying ground, +6, and +12. Heck, it has a voltage regulator on +5 ... and it doesn't have long to live anyway.
Plug the drive in. Spin it up. IMPORTANT: Put it so the platters are at 90 degrees to you! In other words, put it so that if the platters explode, they won't hit you. You don't want it sitting on the ground with the platters parallel to the ground; you want them spinning at 90 degrees to the ground and to you.
Hike back a bit and shoot with a heavy slug. A slug from a 12 gauge works fine. All the rotational (angular) momentum is transferred to the frame as the platters stop almost instantly. The drive goes whirl-whirl-whirl, sometimes up in the air!
I found that the .223 round from an M16A1 was strangely ineffective, but going to full auto made me feel better, and that's the point of all this. [ Yes it's legal to own an M16A1 here if you hop through the paperwork hoops]. Alas, the M16 was a toy I got rid of later ...
9 mm works very well at disassembling the drives, as does .45 cal.
By the way, there are **absolutely terrific** magnets inside modern drives. Open 'em up and use a little acetone to unglue the magnets. They're the "supermagnets". And if you disassemble the drive with shooting, it is commonly opened FOR you!
Shooting a disk drive that crashed and lost you a bunch of work is terrific stuff. In Las Vegas there are a couple places that let you bring in your computer or whatever and shoot it with automatic weapons. Hmmmm, maybe I should try an AK-47 someday, that's a .30 cal round...
*grin*
David Small
Think big.
Take a million car batteries. Yup, lead acid. Why a million car batteries? Because they're so cheap -- we made a hundred million of them last year, and they're one of the best recycling stories out there). Don't be clever. New tech means R&D, and that means unexpected surprises.
Store them in buildings which are above 50 degrees F so that they last a very long time (and certainly not conditions under a hood). Some place near a large interconnect.
Use an inverter, or, just hook the batteries up with IGBT switches to "thermometer" up the voltage, and then back down, making AC.
I'm aware that this is limited storage; the batteries don't like to run at load capacity for long. But also note that Fairbanks has, I believe, 25 megawatts of battery backup.
I'll throw this into the "fresh meat" bowl here ...
-- Dave Small
Someone tries to grab the smiley face?
I say, it's time a missile test up in Minot.
We need to target this guy, and their patent office.
I would think that 350 kt should be sufficient to stop this nonsense.
-- grin,
David
Damn, here I thought that with Star Trek only, what, 5 months away that this would be Yet Another Article on it.
So Annoying.
*grin*
David
Thank you very much for pointing me at good programs on the magnetic field reversal. I appreciate it.
I do know that in my lifetime, the amount in degrees that has to be (added/subtracted) from "Magnetic North" to get "True North" has drifted several degrees, for here in my home state of Colorado. I can remember my surprise on reading that!
It's always disturbing to me when something that I tended to regard as taking "geologic time" is now running at, well, "human time" -- this speed.
Again, thanks for your links.
David Small
SSTO (Single Stage To Orbit) is absolutely the way to go. Gary Hudson has been pushing that idea for some time. Go up, come down, refuel, and go back up.
The Rotary Rocket idea didn't work out for unforeseen reasons; this is why we have "research".
It's a shocker, but implementing a fleet of SSTO's would be cheap and would work . . .
which means our government isn't interested! *grin*
-- Dave
To "slashdot".
Verb.
1) To immediately shift the discussion 90 degrees from any known frame of reference in space-time.
e.g., "Hey, you just slashdotted my proposal!"
2) To devolve into arguing over something 90 degrees from any known frame of reference in space-time. e.g., "I know we were talking about Microsoft patenting the Page Up key, but dammit!, you can't express that Sanskrit character properly without patching the X Server with grep,sed,unk,dwork,glink, and you must be running a kernel later than 2.26.0!"
3) To perform steps (1) and (2) so rapidly that missing one reply will leave the reader 90 degrees from any known frame of reference.
4) The hell you say! What really pisses me off about the Mac are the Microsoft commercials are so incredibly lame, they're no real competition for the Mac commercials. In fact, it must be a conspiracy, started by SCO, the RIAA, the MPAA, under the DMCA, and many, many other four letter acronyms.