I was saying that graphics cards don't address the memory throughput issue, and they don't, because what matters is not how fast you can access the on-board memory, but rather how fast you can stream data to that on-board memory (because even 2GB is only a fraction of overall system memory in these systems, and sever apps in particular tend not to be streaming but more random access), and that's fast but still slightly slower than CPU DRAM itself.
Specs don't tell you everything. You have to interpret them in the context of what you are doing.
Maybe they'll end up establishing new populations in different locations. It's certainly better to give them the chance rather than simply let them be exterminated.
I think it's important to note that while 70,000 eggs seems like a lot, it comes from only 800 nests in two states. Turtles lay a lot of eggs, because most of them are not going to survive in any case.
The point here is that they're hardly moving all the turtle hatchlings to the east coast, so if moving them is futile or even detrimental compared to leaving them in the Gulf, that'll be apparent in the outcome of the sea turtles that will hatch in the sands of the Gulf coast. On the other hand, if they have even a chance to survive in the Atlantic while the Gulf ends up being certain death for the turtles left there, then that's a huge win. The species occurs naturally on the Atlantic coast, so it's not like we're introducing a new species with potentially detrimental consequences.
There's very little downside here, and the potential for a huge upside. The experts may be right that it is futile, but it is absolutely worth trying and I commend these folks for it.
What I was arguing for was a reduced shuttle program as a backup regardless of what we end up deciding to develop for the next launcher. The shuttles are aging, but they currently work. Use Soyuz as a bridge for access to the ISS, but don't rely on it as a sole source.
The shuttle orbiters are aging, and currently on life support, surviving on scavenged working parts from other orbiters. The program has been on life support for years, and is way past due for retirement. Much of the program already has been retired. Consequently, we simply cannot continue the shuttle program until replacements are ready. The plan to add 1-2 more shuttle missions is simply taking advantage of what few spare parts we have left. After that, they will be gone, and the shuttle won't fly no matter who wants it to or how badly.
And so what? What's wrong with relying on Soyuz for a couple years? That program is still active, there are plenty of rockets and capsules, it's ridiculously cheaper to fly than our giant Jack-of-all-trades Flying Space Truck, and has an equivalent performance record with better crew survivability in the event of a disaster as well. It's better in every single way for the purpose of lifting crew or cargo to the ISS.
If you don't have the money to develop the next launcher without completely shutting down the existing program, that indicates to me you probably just don't have the money, period.
Not having the money to both fund the shuttle program indefinitely and build a new HLV does not imply we don't have the money for a new HLV. However, we don't have the money to do the HLV and all the vastly more important and useful stuff too. We need neither the shuttle nor the new HLV, and both should be canceled immediately in favor of actual progress.
Keeping the shuttle alive is the opposite of progress, and if they actually devoted the funds to do it, would succeed where the new HLV has failed to completely kill off all actual progress.
I'd say that pretty much all manned spaceflight from NASA is dead. I'd be very, very surprised if they get anything completed at all, considering their mandate seems to change every time you turn around.
When NASA astronauts fly to the ISS on a Falcon 9 rocket, I'm interested what you'll say about why that doesn't count.
And frankly I hope the 'mandate' for a shuttle contractor bailout gets turned around.
True enough, but using the Shuttle (or parts thereof) doesn't appear to be the way to go. Nothing about the Shuttle is cheap or easy.
For sure. By saying "We need to make it cheaper and easier", I was trying to implicitly agree with you that this means "not use the shuttle-derived HLV":)
IMHO, for the foreseeable future, I'd stay in LEO and work out the nuts and bolts engineering of keeping people alive in space for extended periods of time. When you take six months to plan each space walk, you're not quite ready to venture out of the Van Allen belts.
Also agreed. There's a ton of technology, capability, and process development we can do in LEO that will be highly useful for anything else we want to do. But we've neglected doing any of it in favor of maintaining the shuttle. Hopefully it will only be reduced by this compromise, not completely neglected.
Re:It's now clear where M$ is headed to!
on
Recomputing the Sky
·
· Score: 2, Insightful
Yes, trolling slashdot nerds to install silverlight to view the image out of uncontrollable curiousity.
Someone at M$ is now chuckling while you sell your soul clicking "install silverlight plugin": trolled hard.
It already worked on me when Microsoft put a series of Richard Feynman lectures online. Alas, Moonlight is too "advanced" to load the app. T_T
And I'm not getting trolled all the way into installing a MS operating system!
Obama says, "Screw the moon, I'm setting up a 20 year project to go to Mars."
But that's not what he said. He said "I'm creating projects to develop technology that could enable a mission to Mars in 20 years", and that's a huge difference. He's talking about developing general technologies and capabilities that would be useful for a wide variety of missions outside of Mars, and if nobody wants to pull the trigger on the Mars mission in 20 years, we still have all the technology and capabilities. Mars was only mentioned to make the people who think we must have a specific mission happy (and it's not a bad policy to at least have a practical application in mind).
Whereas a definite "Mars in 20 years" would mean lots of development of tech designed for that mission and only that mission. 20 years to have enough technology in place that a Mars mission doesn't require that much specific development is a much more sensible, useful, and future-proof plan.
But hey, I guess having a giant expensive rocket that can't do anything rockets of 30 years ago couldn't do is nice too.:/
The phrase "baby with the bathwater" comes to mind here. NASA does some things that no other US entity currently does.
Completely agreed but none of the things I care about are tied to the shuttle or derived vehicles.
We're about to rely on a foreign country as our sole source supplier for manned access to the ISS for at least several years.
It gives us a backup that won't take years to be ready. Ultimately, a man rated Falcon 9 or some other private launcher would be a good solution. But, we don't have it yet.
Except it will take years to be ready. The new schedule has the new HLV's first launch in 2015. SpaceX has claimed they could have their first manned launch in 2013.
Frankly I don't expect either schedule to hold, but I still think it's likely that SpaceX will be delivering crew to the ISS before the shuttle-derived launcher can, and at a greatly reduced cost too.
There is no circumstance under which we aren't dependent on the Russians for some period of time, so what is this plan getting us exactly?
If you want NASA to push out of LEO, you need some better systems.
If we want to get out of LEO, then we need to make getting to LEO cheaper and easier, and develop technology that will let us go from there as a separate step. Lifting everything we need for a manned moon or (ha!) Mars mission from the surface of the earth one one giant rocket is foolish and will just mean the mission scope is cut down to the point of, well, pointlessness.
Keeping the Shuttle pieces parts going is mostly a make work project for a couple of Senators and their constituents. It has no scientific or engineering value.
Don't forget it also apparently keeps prices down on ICBM parts, because the DOD is so strapped for cash they need NASA to subsidize their equipment(?!)
Oh well. At least the pointless moon mission is dead. Hopefully this compromise doesn't cripple the actual useful and new projects that will expand our capabilities. And hey, maybe we'll actually find a good use for our HLV to LEO, and not just find arbitrary ways to justify its existence.
A bill that kills NASA entirely would be a better direction for space research and the United States.
Without NASA there would be virtually no space research in the United States, which is only "better" if you aren't in favor of space exploration to begin with. Nobody but NASA is going to launch missions like LISA, Cassini, Deep Impact, Mars Science Laboratory, etc etc. The only people on earth that are doing things like that are other governmental space agencies. Much like NSF, NASA serves a vital function of providing funding for projects that are infeasible for universities and unprofitable for private industry, with basic research that advances the state of knowledge and technology for the future.
The problem with NASA, the thing that makes it a political football, is the huge in-house rocket projects. The shuttle (and now derivatives) represent $billions/year all going to a single project and a small number of contractors. A giant target like that is tempting to get rid of, and nearly impossible for those profiting from it to let go of. Thus the political stalemate.
Yet all the interesting projects I mentioned, and all the technology programs that Obama wanted to have happen and which I pray to God won't be crippled by this compromise, are individually much cheaper. No single constituency has such a stake in them that they will fight tooth and nail to keep them, nor are they such tempting targets for cuts. They're more flexible, and also more broadly addressing the needs of future space exploration.
The shuttle-derived HLV, that does nothing but keep a contractor in business and let NASA have a rocket with its logo on the side, is the problem. Other than that, NASA is fine and does great work and saying it should be killed is the worst idea ever.
All you say makes sense, but I for one don't understand the market for this. Today, if you need a compute server that's good for stream (e.g. SIMD) workloads you get a dozen 1U/2U rackmounts and fill them up with as many GPU boards as they'll take.
Well I was mostly just trying to justify why the transition to the situation "today" is taking place, and why the multi-threading itself isn't that big a deal. "Yesterday" the biggest compute servers were still made from traditional CPUs. Only recently has the potential for GPUs as general purpose (if your "general" purpose is FP math) computational devices really captured significant mindshare, and APIs an methodologies are still being ironed out.
It seemed like the article was talking about having GPUs in the "server" room in general, not about a specific situation where there's only a few GPUs stuck in an otherwise normal rackmount server. That would be fairly pointless. Though on the other hand there is some research that suggests that having more than a token amount of regular CPUs close to the GPUs is useful.
Programmers of Server applications are already used to multithreading, and they've been able to make good use of systems with large numbers of processors on them even before the advent of virtualization.
But don't pay too much attention to the word "Server". Yes the machines that they're talking about are in the segment of the market referred to as "servers", as distinct from "desktops" or "mobile". But the target of GPU-based computing isn't "Servers" in the sense of the tasks you normally think of -- web servers, database servers, etc.
The real target is mentioned in the article, and it's HPC, aka scientific computing. Normal server apps are integer code, and depend more on high memory bandwidth and I/O, which GPGPU doesn't really address. HPC wants that stuff too, but they also want floating point performance. As much floating point math performance as you can possibly give them. And GPUs are way beyond what CPUs can provide in that regard. Plus a lot of HPC applications are easier to parallelize than even the traditional server codes, though not all fall in the "embarrassingly parallel" category.
There will be a few growing pains, but once APIs get straightened out and programmers get used to it (which shouldn't take too long for the ones writing HPC code), this is going to be a huge win for scientific computing.
True, but it sure would be a bummer to successfully decelerate from mach25 and 100 miles to mach "something" and 100Kft and then discover the parachute won't unfurl properly or whatever.
Yeah, but what a fucking awesome story to tell in Heaven/the afterlife's waiting room (depending on if you are Christian or Beetlejuiceian).
Though on the other hand, colloquially "suicidal behavior" often refers to actions taken that aren't necessarily intended to result in death but for which it is a nearly inevitable outcome.
You may not say the guy who drives around at night with the headlights off while wasted is "committing suicide", but you might say he's engaged in "suicidal behavior". Same with the rat infected with toxoplasma who is attracted to the smell of cat urine; the fungus is making the rat behave suicidally even if the rat isn't consciously trying to die. Same with the shrimp.
I agree that the headline seems to imply an emotional state of "suicidal" which is obviously nonsense. But I think you could say that it makes them engage in suicidal behavior.
It doesn't matter how often people say this, it won't make it true. Case history even in the early '80s already confirmed that it was fairly difficult to prove that derived firmware work had been created (for reasons I explained in the '99 case). IBM, unlike you, had smart enough lawyers to know that they weren't going to be able to protect their whole PC over a few kilobytes of ROM BIOS.
It doesn't matter how often you say IBM didn't want to stop Compaq, it only demonstrates your ignorance. You're nearly right about the rest, though, in that IBM's lawyers understood copyright law well enough that once the facts of the case were in the open, they knew that they didn't have a chance and endless appeals would be pointless.
You seem to think every company is like SCO, and will pursue a case indefinitely if they want to win regardless of facts, so if they accept a defeat, it means they didn't want to win. Which is cute and funny and stupid.
That's ridiculous and you know it. If you look at what X's work does and how it works, and use that information to copy what it does, then you are using X's creative input in order to make a creative work of your own. Thus a derivative work has been created.
And here you demonstrate your ignorance regarding what constitutes a derivative work, and then in your analogies ignorance of the specific situation. Maybe read up on the legal code and acquire a clue.
Today, a "clean-room reinterpretation" is used fairly honestly: you can only read what you're not restricted by or from reading. IOW, you implement from published licence-unencumbered specs. 30 years ago it meant, "Pretend that a derived work hadn't been created by handwaving an imaginary wall. This wall ignores the nature of creative expression in software and takes advantage of a technically illiterate court."
The wall was real, no handwaving, otherwise Compaq would have lost the case. And the same technique is used today which you'd know if you had any relevant experience and weren't using your imagination in place of knowledge -- the court exceeding you both in technical experience and knowledge of the law despite your hilarious claims otherwise. The main difference today is in that word "license-unencumbered" since back then software patents were much more rare. Had IBM held patents on pieces of the BIOS, not even a clean-room implementation would have been legally safe. But I'm sure that's not the point you were trying to make, since it's far too intelligent.
You're angry because you know that so much of the geek dream lore surrounding the development of PC clones is sophomoric bullshit.
I'm laughing my ass off. You're mad because your attempt to re-imagine reality, the public record, and established case history is failing so spectacularly.
BIOS. Clones were born, IBM moved on, and the victory for straight "piracy" was rechronicled lest the nascent commercial PC software market be scared away. Perhaps had people such as yourself not been so dishonest at the time, we'd see a very different software landscape where copying is considered far more acceptable.
Bwa ha ha! Oh yeah! If IBM and others hadn't realized there was a legally valid way to implement "secret" functionality, they totally would have been happy to have people outright copy later! Please. The connection to "straight piracy" is in your imagination only. The rest of the engineering profession knows that reverse-engineering is important to the point where they were even able to get exceptions in the DMCA for it. The "pro-piracy" crowd you are so hilariously trying to paint the accurate historical view as coming from has not such power. IEEE and ACM? That's a different story.
IBM hadn't moved on even in the 90s, btw, when I was there. They were still bitter over losing control of the PC market, though their impression of the case was the same as everyone else, at least those who know what happened. I.e. not you. But keep it up! You're cracking me up here with your "truth".
IBM's mock lawsuit against Compaq (I say "mock" because it expended pretty much no resources by IBM standards, and probably existed only to placate certain investors)
This "IBM didn't actually want to win the lawsuit" theory belies a hilarious ignorance of IBM in the 80s. They absolutely wanted to stop any unlicensed BIOS. They absolutely wanted to control the home PC market. And they absolutely knew what the consequences would be -- loss of market control. Which is exactly what happened.
It may surprise you, but some companies actually don't pour endless resources down lawsuits that are lost causes. Some lawyers actually inform their clients of the legal realities, not the fantasies the client wants to hear. Actually... most, in both cases.
Compaq did a good job of creating a BIOS which looked nothing like IBM's BIOS
And ergo was not a derivative work. That's the long and short of it. If you implement a piece of code based simply on a description of how it's supposed to work (which is what the AC is trying to describe to you, him playing the role of the desciber) and the result doesn't look substantially like the original work that was described, then that's not a derivative work in any sense.
The lie to the technically inexperienced court is in claiming that you didn't essentially copy the IBM BIOS in all but machine code byte order.
If you were technically experienced you'd understand what a clean-room reimplementation was and this conversation never would have begun.
No it wasn't. That guy read the disassembled code, someone else wrote a new BIOS based on the first guy's description of how it was supposed to work. They didn't directly copy any code, they didn't directly copy any algorithms. They did a clean-room reverse engineered implementation. Maybe google that for en-clue-ification. It's a perfectly legitimate technique.
Because that's not what we want to do. ISS resupply is for commercial endeavours. Moon, Mars, BEO: that's what we're supposed to be shooting for.
But we shouldn't be shooting for the moon in a giant rocket. That's a recipe for re-creating Apollo, the most useless endeavor I can imagine for the space program. Instead we should be first developing the orbital technology where we can get mass to orbit cheaply, then refuel in orbit for the next stage. Bring crew up separately. Assemble a lunar vehicle in orbit. And so on. So that we don't need a single rocket that lifts everything needed to go to the moon and return in one shot. So that once we arrive at the moon, we'll have supplies and even habitats waiting for our astronauts there. That's what we're supposed to be shooting for, and neither Constellation nor DIRECT get us there. They get us DIRECTly to a mid-life-crisis re-enactment of our glory days. But guess what? Re-creating our glory days does not recreate the glory. It does nothing to advance us.
The only way we're going to do Mars is if we do it in a piecemeal fashion. The impetus and more importantly funds for a one-shot mission are never going to actually appear -- even before Bush left office his "Mars, Bitches!" plan had been reduced to "Apollo again, maybe!". Constellation isn't going to be able to do it. And even if we could, all we'd do is put boot-prints on the red planet. Woopty-fucking-do.
What?!? Apart from one catastrophic O-ring problem the SRBs have performed flawlessly.
Yeah, aside from that. Oh and except for cost and reuse, which once upon a time were actually part of the supposed features. They do actually haul the SRBs out of the ocean, but it's not even worth it. The main cost is in the fuel. As is all the unreliability and limitations. A flawless solid rocket still can't be shut off or throttled or refueled except on the ground at great expense. But let's not pretend SRBs are flawless anyway.
There's a place for solid rockets, but there's no reason they need to be SSRBs specifically as newer designs exist, and even less reason they need to be used on a manned craft. Even less reason they need to be the only choice of launch vehicle. Strap solid fuel boosters on a Delta, Falcon, or Ariane if you must, but don't require them especially if we're operating under the "only enough budget for one vehicle" regime you suggest.
Can't really argue with that. NASA should work like the army. The army doesn't design tanks. They say "we need a tank that does this". The various companies submit bids and the army picks one. NASA should just say "we need a rocket that can lift x tons into such and such an orbit."
Which is what they're going to be doing if the plan takes effect. They'll be paying contracts to perform X task, not contracts to develop a vehicles with specified features where the government ends up paying for the inevitable cost overruns.
Everything about the new plan is better than the old. The giant rocket is going to castrate NASAs ability to do new and important things. Things that will largely obviate the giant rocket. If we could do both, that'd be great, but that's not the way it's turning out, is it? So we can keep NASA with basically the same capabilities it used to have -- a space-SUV that's crazy expensive and operates infrequently -- or we can do something new and better. It's not a tough call for me, nostalgia, "prestige", or any of that notwithstanding one bit.
If you've got lots of money to throw at the problem, then by all means, develop two completely new rockets and perform those tasks.
You don't need to develop two completely different rockets if you focus on developing modular, reusable engines. SpaceX developed the Falcon I and Falcon IX for less than the cost of a single shuttle launch, and they'll develop the Falcon IX Heavy for an even smaller incremental cost.
And if you don't have lots of money, why exactly are you wasting it launching a heavy lift vehicle when all you want is to get some crew to the ISS?
However, if you've got limited funds and you've already got a nice, man-rated, heavy lift system that has 30+ years of near flawless operation, it makes more sense to take a more DIRECT approach.
Actually if you have limited funds it makes more sense to get private industry to provide truly reusable rockets using liquid fuel, since the main thing we've learned in those 30+ years is that solid fuel boosters are problematic and not very suitable to reuse.
If we simply must keep using shuttle technology for the sake of the DoD (who can't afford solid rockets for ICBMs but NASA can in what universe?), then DIRECT is better than Constellation. But neither make any sense for the future of space travel.
Slashdot Mirror
Arthur C Clarke did a much better job of explaining the concept with his "Sunjammer" story that appeared in Boy's Life ca 1964.
What?! Clarke was writing for Boy's Life in 1964? When and why did that lame-ass publication become so lame-ass when I was reading it in the 80s?!
You were saying?
I was saying that graphics cards don't address the memory throughput issue, and they don't, because what matters is not how fast you can access the on-board memory, but rather how fast you can stream data to that on-board memory (because even 2GB is only a fraction of overall system memory in these systems, and sever apps in particular tend not to be streaming but more random access), and that's fast but still slightly slower than CPU DRAM itself.
Specs don't tell you everything. You have to interpret them in the context of what you are doing.
This is slashdot. Brace yourself.
Bracing yourself is futile, and only idiots would try it.
Maybe they'll end up establishing new populations in different locations.
It's certainly better to give them the chance rather than simply let them be exterminated.
I think it's important to note that while 70,000 eggs seems like a lot, it comes from only 800 nests in two states. Turtles lay a lot of eggs, because most of them are not going to survive in any case.
The point here is that they're hardly moving all the turtle hatchlings to the east coast, so if moving them is futile or even detrimental compared to leaving them in the Gulf, that'll be apparent in the outcome of the sea turtles that will hatch in the sands of the Gulf coast. On the other hand, if they have even a chance to survive in the Atlantic while the Gulf ends up being certain death for the turtles left there, then that's a huge win. The species occurs naturally on the Atlantic coast, so it's not like we're introducing a new species with potentially detrimental consequences.
There's very little downside here, and the potential for a huge upside. The experts may be right that it is futile, but it is absolutely worth trying and I commend these folks for it.
What I was arguing for was a reduced shuttle program as a backup regardless of what we end up deciding to develop for the next launcher. The shuttles are aging, but they currently work. Use Soyuz as a bridge for access to the ISS, but don't rely on it as a sole source.
The shuttle orbiters are aging, and currently on life support, surviving on scavenged working parts from other orbiters. The program has been on life support for years, and is way past due for retirement. Much of the program already has been retired. Consequently, we simply cannot continue the shuttle program until replacements are ready. The plan to add 1-2 more shuttle missions is simply taking advantage of what few spare parts we have left. After that, they will be gone, and the shuttle won't fly no matter who wants it to or how badly.
And so what? What's wrong with relying on Soyuz for a couple years? That program is still active, there are plenty of rockets and capsules, it's ridiculously cheaper to fly than our giant Jack-of-all-trades Flying Space Truck, and has an equivalent performance record with better crew survivability in the event of a disaster as well. It's better in every single way for the purpose of lifting crew or cargo to the ISS.
If you don't have the money to develop the next launcher without completely shutting down the existing program, that indicates to me you probably just don't have the money, period.
Not having the money to both fund the shuttle program indefinitely and build a new HLV does not imply we don't have the money for a new HLV. However, we don't have the money to do the HLV and all the vastly more important and useful stuff too. We need neither the shuttle nor the new HLV, and both should be canceled immediately in favor of actual progress.
Keeping the shuttle alive is the opposite of progress, and if they actually devoted the funds to do it, would succeed where the new HLV has failed to completely kill off all actual progress.
I'd say that pretty much all manned spaceflight from NASA is dead. I'd be very, very surprised if they get anything completed at all, considering their mandate seems to change every time you turn around.
When NASA astronauts fly to the ISS on a Falcon 9 rocket, I'm interested what you'll say about why that doesn't count.
And frankly I hope the 'mandate' for a shuttle contractor bailout gets turned around.
True enough, but using the Shuttle (or parts thereof) doesn't appear to be the way to go. Nothing about the Shuttle is cheap or easy.
For sure. By saying "We need to make it cheaper and easier", I was trying to implicitly agree with you that this means "not use the shuttle-derived HLV" :)
IMHO, for the foreseeable future, I'd stay in LEO and work out the nuts and bolts engineering of keeping people alive in space for extended periods of time. When you take six months to plan each space walk, you're not quite ready to venture out of the Van Allen belts.
Also agreed. There's a ton of technology, capability, and process development we can do in LEO that will be highly useful for anything else we want to do. But we've neglected doing any of it in favor of maintaining the shuttle. Hopefully it will only be reduced by this compromise, not completely neglected.
Yes, trolling slashdot nerds to install silverlight to view the image out of uncontrollable curiousity.
Someone at M$ is now chuckling while you sell your soul clicking "install silverlight plugin": trolled hard.
It already worked on me when Microsoft put a series of Richard Feynman lectures online. Alas, Moonlight is too "advanced" to load the app. T_T
And I'm not getting trolled all the way into installing a MS operating system!
Obama says, "Screw the moon, I'm setting up a 20 year project to go to Mars."
But that's not what he said. He said "I'm creating projects to develop technology that could enable a mission to Mars in 20 years", and that's a huge difference. He's talking about developing general technologies and capabilities that would be useful for a wide variety of missions outside of Mars, and if nobody wants to pull the trigger on the Mars mission in 20 years, we still have all the technology and capabilities. Mars was only mentioned to make the people who think we must have a specific mission happy (and it's not a bad policy to at least have a practical application in mind).
Whereas a definite "Mars in 20 years" would mean lots of development of tech designed for that mission and only that mission. 20 years to have enough technology in place that a Mars mission doesn't require that much specific development is a much more sensible, useful, and future-proof plan.
But hey, I guess having a giant expensive rocket that can't do anything rockets of 30 years ago couldn't do is nice too. :/
The phrase "baby with the bathwater" comes to mind here. NASA does some things that no other US entity currently does.
Completely agreed but none of the things I care about are tied to the shuttle or derived vehicles.
We're about to rely on a foreign country as our sole source supplier for manned access to the ISS for at least several years.
It gives us a backup that won't take years to be ready. Ultimately, a man rated Falcon 9 or some other private launcher would be a good solution. But, we don't have it yet.
Except it will take years to be ready. The new schedule has the new HLV's first launch in 2015. SpaceX has claimed they could have their first manned launch in 2013.
Frankly I don't expect either schedule to hold, but I still think it's likely that SpaceX will be delivering crew to the ISS before the shuttle-derived launcher can, and at a greatly reduced cost too.
There is no circumstance under which we aren't dependent on the Russians for some period of time, so what is this plan getting us exactly?
If you want NASA to push out of LEO, you need some better systems.
If we want to get out of LEO, then we need to make getting to LEO cheaper and easier, and develop technology that will let us go from there as a separate step. Lifting everything we need for a manned moon or (ha!) Mars mission from the surface of the earth one one giant rocket is foolish and will just mean the mission scope is cut down to the point of, well, pointlessness.
Keeping the Shuttle pieces parts going is mostly a make work project for a couple of Senators and their constituents. It has no scientific or engineering value.
Don't forget it also apparently keeps prices down on ICBM parts, because the DOD is so strapped for cash they need NASA to subsidize their equipment(?!)
Oh well. At least the pointless moon mission is dead. Hopefully this compromise doesn't cripple the actual useful and new projects that will expand our capabilities. And hey, maybe we'll actually find a good use for our HLV to LEO, and not just find arbitrary ways to justify its existence.
A bill that kills NASA entirely would be a better direction for space research and the United States.
Without NASA there would be virtually no space research in the United States, which is only "better" if you aren't in favor of space exploration to begin with. Nobody but NASA is going to launch missions like LISA, Cassini, Deep Impact, Mars Science Laboratory, etc etc. The only people on earth that are doing things like that are other governmental space agencies. Much like NSF, NASA serves a vital function of providing funding for projects that are infeasible for universities and unprofitable for private industry, with basic research that advances the state of knowledge and technology for the future.
The problem with NASA, the thing that makes it a political football, is the huge in-house rocket projects. The shuttle (and now derivatives) represent $billions/year all going to a single project and a small number of contractors. A giant target like that is tempting to get rid of, and nearly impossible for those profiting from it to let go of. Thus the political stalemate.
Yet all the interesting projects I mentioned, and all the technology programs that Obama wanted to have happen and which I pray to God won't be crippled by this compromise, are individually much cheaper. No single constituency has such a stake in them that they will fight tooth and nail to keep them, nor are they such tempting targets for cuts. They're more flexible, and also more broadly addressing the needs of future space exploration.
The shuttle-derived HLV, that does nothing but keep a contractor in business and let NASA have a rocket with its logo on the side, is the problem. Other than that, NASA is fine and does great work and saying it should be killed is the worst idea ever.
All you say makes sense, but I for one don't understand the market for this. Today, if you need a compute server that's good for stream (e.g. SIMD) workloads you get a dozen 1U/2U rackmounts and fill them up with as many GPU boards as they'll take.
Well I was mostly just trying to justify why the transition to the situation "today" is taking place, and why the multi-threading itself isn't that big a deal. "Yesterday" the biggest compute servers were still made from traditional CPUs. Only recently has the potential for GPUs as general purpose (if your "general" purpose is FP math) computational devices really captured significant mindshare, and APIs an methodologies are still being ironed out.
It seemed like the article was talking about having GPUs in the "server" room in general, not about a specific situation where there's only a few GPUs stuck in an otherwise normal rackmount server. That would be fairly pointless. Though on the other hand there is some research that suggests that having more than a token amount of regular CPUs close to the GPUs is useful.
Programmers of Server applications are already used to multithreading, and they've been able to make good use of systems with large numbers of processors on them even before the advent of virtualization.
But don't pay too much attention to the word "Server". Yes the machines that they're talking about are in the segment of the market referred to as "servers", as distinct from "desktops" or "mobile". But the target of GPU-based computing isn't "Servers" in the sense of the tasks you normally think of -- web servers, database servers, etc.
The real target is mentioned in the article, and it's HPC, aka scientific computing. Normal server apps are integer code, and depend more on high memory bandwidth and I/O, which GPGPU doesn't really address. HPC wants that stuff too, but they also want floating point performance. As much floating point math performance as you can possibly give them. And GPUs are way beyond what CPUs can provide in that regard. Plus a lot of HPC applications are easier to parallelize than even the traditional server codes, though not all fall in the "embarrassingly parallel" category.
There will be a few growing pains, but once APIs get straightened out and programmers get used to it (which shouldn't take too long for the ones writing HPC code), this is going to be a huge win for scientific computing.
Yeah, hey man, I hate to call you out, but those weren't binaural beats. That was just regular bad music.
Skin flutes are a kind of flute...
True, but it sure would be a bummer to successfully decelerate from mach25 and 100 miles to mach "something" and 100Kft and then discover the parachute won't unfurl properly or whatever.
Yeah, but what a fucking awesome story to tell in Heaven/the afterlife's waiting room (depending on if you are Christian or Beetlejuiceian).
Though on the other hand, colloquially "suicidal behavior" often refers to actions taken that aren't necessarily intended to result in death but for which it is a nearly inevitable outcome.
You may not say the guy who drives around at night with the headlights off while wasted is "committing suicide", but you might say he's engaged in "suicidal behavior". Same with the rat infected with toxoplasma who is attracted to the smell of cat urine; the fungus is making the rat behave suicidally even if the rat isn't consciously trying to die. Same with the shrimp.
I agree that the headline seems to imply an emotional state of "suicidal" which is obviously nonsense. But I think you could say that it makes them engage in suicidal behavior.
Since you can't really put a shrimp on a shrink-couch and ask it about its feelings,
Of course you can!
it is very hard to say whether the shrimp are "suicidal" or whether their fear responses are being blunted.
Ah well, that's true, since they aren't so big in the "answering" department.
And this is where some fucking idiot chimes in about the 3 laws of robotics.
I love these little self-fulfilling prophecies.
It doesn't matter how often people say this, it won't make it true. Case history even in the early '80s already confirmed that it was fairly difficult to prove that derived firmware work had been created (for reasons I explained in the '99 case). IBM, unlike you, had smart enough lawyers to know that they weren't going to be able to protect their whole PC over a few kilobytes of ROM BIOS.
It doesn't matter how often you say IBM didn't want to stop Compaq, it only demonstrates your ignorance. You're nearly right about the rest, though, in that IBM's lawyers understood copyright law well enough that once the facts of the case were in the open, they knew that they didn't have a chance and endless appeals would be pointless.
You seem to think every company is like SCO, and will pursue a case indefinitely if they want to win regardless of facts, so if they accept a defeat, it means they didn't want to win. Which is cute and funny and stupid.
That's ridiculous and you know it. If you look at what X's work does and how it works, and use that information to copy what it does, then you are using X's creative input in order to make a creative work of your own. Thus a derivative work has been created.
And here you demonstrate your ignorance regarding what constitutes a derivative work, and then in your analogies ignorance of the specific situation. Maybe read up on the legal code and acquire a clue.
Today, a "clean-room reinterpretation" is used fairly honestly: you can only read what you're not restricted by or from reading. IOW, you implement from published licence-unencumbered specs. 30 years ago it meant, "Pretend that a derived work hadn't been created by handwaving an imaginary wall. This wall ignores the nature of creative expression in software and takes advantage of a technically illiterate court."
The wall was real, no handwaving, otherwise Compaq would have lost the case. And the same technique is used today which you'd know if you had any relevant experience and weren't using your imagination in place of knowledge -- the court exceeding you both in technical experience and knowledge of the law despite your hilarious claims otherwise. The main difference today is in that word "license-unencumbered" since back then software patents were much more rare. Had IBM held patents on pieces of the BIOS, not even a clean-room implementation would have been legally safe. But I'm sure that's not the point you were trying to make, since it's far too intelligent.
You're angry because you know that so much of the geek dream lore surrounding the development of PC clones is sophomoric bullshit.
I'm laughing my ass off. You're mad because your attempt to re-imagine reality, the public record, and established case history is failing so spectacularly.
BIOS. Clones were born, IBM moved on, and the victory for straight "piracy" was rechronicled lest the nascent commercial PC software market be scared away. Perhaps had people such as yourself not been so dishonest at the time, we'd see a very different software landscape where copying is considered far more acceptable.
Bwa ha ha! Oh yeah! If IBM and others hadn't realized there was a legally valid way to implement "secret" functionality, they totally would have been happy to have people outright copy later! Please. The connection to "straight piracy" is in your imagination only. The rest of the engineering profession knows that reverse-engineering is important to the point where they were even able to get exceptions in the DMCA for it. The "pro-piracy" crowd you are so hilariously trying to paint the accurate historical view as coming from has not such power. IEEE and ACM? That's a different story.
IBM hadn't moved on even in the 90s, btw, when I was there. They were still bitter over losing control of the PC market, though their impression of the case was the same as everyone else, at least those who know what happened. I.e. not you. But keep it up! You're cracking me up here with your "truth".
IBM's mock lawsuit against Compaq (I say "mock" because it expended pretty much no resources by IBM standards, and probably existed only to placate certain investors)
This "IBM didn't actually want to win the lawsuit" theory belies a hilarious ignorance of IBM in the 80s. They absolutely wanted to stop any unlicensed BIOS. They absolutely wanted to control the home PC market. And they absolutely knew what the consequences would be -- loss of market control. Which is exactly what happened.
It may surprise you, but some companies actually don't pour endless resources down lawsuits that are lost causes. Some lawyers actually inform their clients of the legal realities, not the fantasies the client wants to hear. Actually... most, in both cases.
Compaq did a good job of creating a BIOS which looked nothing like IBM's BIOS
And ergo was not a derivative work. That's the long and short of it. If you implement a piece of code based simply on a description of how it's supposed to work (which is what the AC is trying to describe to you, him playing the role of the desciber) and the result doesn't look substantially like the original work that was described, then that's not a derivative work in any sense.
The lie to the technically inexperienced court is in claiming that you didn't essentially copy the IBM BIOS in all but machine code byte order.
If you were technically experienced you'd understand what a clean-room reimplementation was and this conversation never would have begun.
IBM's work was stolen.
Plain and simple.
No it wasn't. That guy read the disassembled code, someone else wrote a new BIOS based on the first guy's description of how it was supposed to work. They didn't directly copy any code, they didn't directly copy any algorithms. They did a clean-room reverse engineered implementation. Maybe google that for en-clue-ification. It's a perfectly legitimate technique.
Because that's not what we want to do. ISS resupply is for commercial endeavours. Moon, Mars, BEO: that's what we're supposed to be shooting for.
But we shouldn't be shooting for the moon in a giant rocket. That's a recipe for re-creating Apollo, the most useless endeavor I can imagine for the space program. Instead we should be first developing the orbital technology where we can get mass to orbit cheaply, then refuel in orbit for the next stage. Bring crew up separately. Assemble a lunar vehicle in orbit. And so on. So that we don't need a single rocket that lifts everything needed to go to the moon and return in one shot. So that once we arrive at the moon, we'll have supplies and even habitats waiting for our astronauts there. That's what we're supposed to be shooting for, and neither Constellation nor DIRECT get us there. They get us DIRECTly to a mid-life-crisis re-enactment of our glory days. But guess what? Re-creating our glory days does not recreate the glory. It does nothing to advance us.
The only way we're going to do Mars is if we do it in a piecemeal fashion. The impetus and more importantly funds for a one-shot mission are never going to actually appear -- even before Bush left office his "Mars, Bitches!" plan had been reduced to "Apollo again, maybe!". Constellation isn't going to be able to do it. And even if we could, all we'd do is put boot-prints on the red planet. Woopty-fucking-do.
What?!? Apart from one catastrophic O-ring problem the SRBs have performed flawlessly.
Yeah, aside from that. Oh and except for cost and reuse, which once upon a time were actually part of the supposed features. They do actually haul the SRBs out of the ocean, but it's not even worth it. The main cost is in the fuel. As is all the unreliability and limitations. A flawless solid rocket still can't be shut off or throttled or refueled except on the ground at great expense. But let's not pretend SRBs are flawless anyway.
There's a place for solid rockets, but there's no reason they need to be SSRBs specifically as newer designs exist, and even less reason they need to be used on a manned craft. Even less reason they need to be the only choice of launch vehicle. Strap solid fuel boosters on a Delta, Falcon, or Ariane if you must, but don't require them especially if we're operating under the "only enough budget for one vehicle" regime you suggest.
Can't really argue with that. NASA should work like the army. The army doesn't design tanks. They say "we need a tank that does this". The various companies submit bids and the army picks one. NASA should just say "we need a rocket that can lift x tons into such and such an orbit."
Which is what they're going to be doing if the plan takes effect. They'll be paying contracts to perform X task, not contracts to develop a vehicles with specified features where the government ends up paying for the inevitable cost overruns.
Everything about the new plan is better than the old. The giant rocket is going to castrate NASAs ability to do new and important things. Things that will largely obviate the giant rocket. If we could do both, that'd be great, but that's not the way it's turning out, is it? So we can keep NASA with basically the same capabilities it used to have -- a space-SUV that's crazy expensive and operates infrequently -- or we can do something new and better. It's not a tough call for me, nostalgia, "prestige", or any of that notwithstanding one bit.
If you've got lots of money to throw at the problem, then by all means, develop two completely new rockets and perform those tasks.
You don't need to develop two completely different rockets if you focus on developing modular, reusable engines. SpaceX developed the Falcon I and Falcon IX for less than the cost of a single shuttle launch, and they'll develop the Falcon IX Heavy for an even smaller incremental cost.
And if you don't have lots of money, why exactly are you wasting it launching a heavy lift vehicle when all you want is to get some crew to the ISS?
However, if you've got limited funds and you've already got a nice, man-rated, heavy lift system that has 30+ years of near flawless operation, it makes more sense to take a more DIRECT approach.
Actually if you have limited funds it makes more sense to get private industry to provide truly reusable rockets using liquid fuel, since the main thing we've learned in those 30+ years is that solid fuel boosters are problematic and not very suitable to reuse.
If we simply must keep using shuttle technology for the sake of the DoD (who can't afford solid rockets for ICBMs but NASA can in what universe?), then DIRECT is better than Constellation. But neither make any sense for the future of space travel.