> Dark matter [wikipedia.org] is required by looking at > galaxy rotation curves.
Which, in every case, suggests a number about 10 times visible mass. I point this out for reasons that will become clear later.
> Dark matter was first hypothesized by Zwicky in 1933
After examining galactic _clusters_. In his analisys, the number was 400 times.
And then you don't note the other argument: that the large-scale structure of the universe cannot be explained without adding lots of mass. The filamental structure of the superclusters are supposed to be gravitationally constructed, but there's no way they could have done so given the visible mass. Using this, we arrive at a figure of about 24x.
Now here's the problem. The _visible_ mass of the universe makes it very close to being energetically "flat", that is, the gravitational energy is very close to the mass energy. This is actually the most "obvious" way for a universe to be built.
So basically dark matter is a _bad_ thing. It ruins any simple model of the universe, and yet requires us to believe in an unseen mass which fails every theory we've thrown at it and remains unseen in experiment.
All this to remain tied to GR, which most QM theorists generally regard as being wrong (ie, a simplification) anyway. And let's not forget, the only evidence we have that GR "works" are local experiments, the more we look out into space, more more fudge factors we have to add.
> Dark energy [wikipedia.org] is required by looking at [snip] > In about 1998, the supernova observations were pinned down > well enough to show that the expansion rate is actually > increasing with time. Therefore, there must be some > "antigravity" force that causes the expansion of the > universe to accelerate. This is dubbed "dark energy."
Now perhaps I'm the only one, but this is where I start thinking the whole damb thing is just plain WRONG.
I mean doesn't anyone find it the least bit ironic that the imaginary mass we created ex-nihlo now has to be counterbalanced by _another_ even more wild fudge factor?
This is aether all over again. We're at the "let's pretend the aether gets dragged by mass" stage. Every time a new telescope gets first light you can hear the theories crashing, and the mad scramble to add in yet another fudge factor starts anew.
I mean let's face it, if this latest report is supportable, now we're adding _another_ curve-fit to the original curve-fit because the curve doesn't fit!
> The new claim in the current article is that the effect > of dark energy has changed over time. The fundamental > problem is that the new evidence relies on gamma-ray bursts, > which are not nearly as well established a standard candle > as the Type 1a supernovae, so it's much harder to say with > certainty what distance they are at.
Let's be honest, we don't even _really_ know what these things are. Until recently there were arguments they were in fact small events occuring in "local" space (ie, within light years, not mega-light-years). This report seems shaky to me.
Do you know why the Shuttle costs so much and keeps blowing up? Because it's too complex. It consists of four major "parts" that have to be mated in the VAB then carried out to the pad on an rediculous transport. The four "parts" are individually some of the most complex and part-intensive rockets even build. The SRB's, in turn, are fabricated in parts and shipped for assembly.
Failure is strongly related to the number of parts. More parts means more things that can break, and a highly failure rate as a result.
So maybe people out there will understand why I find the idea of strapping five of these things to a new and part-assembled core is an ABSOLUTELY REDICULOUS IDEA.
Sure, SSTO is difficult, but it IS doable. But even if you don't want to spend that much, we know for sure we can do low-risk DSTO like Shuttle II et all. All hail the reduction of development budgets!
This is simply the wrong way to go. Dump Fred into the ocean, where it belongs, and built a _real_ launch system. Enough already!
You can tell the iPod craze is reaching it's peak if people are willing to beat each other up to get to _buy_ an ancient machine from the same company.
The manned portion consists of an SRB. You know, the big candle thing that keeps blowing up either in-air or on the ground? And now that it goes higher I'm not sure they plan on gettng them back.
The unmanned portion consists of two SRB's (ditto x 2) along with the ET, which was designed for off-axis thrust. In fact it is the off-axis thrust that doomed the wire-round SRB the Air Force worked on, because it was too stiff and couldn't flex enough when the SME's lit up. Now to this we put 100 tonnes of load on top of the ET and 2 million pounds of thrust on the bottom, something it was never designed for (although should have been).
So basically they're going to redesign the ET for changes to the thrust line, which adds nothing of value, but they're NOT going to redesign the SRB which would improve safety and performance.
This has all been looked at in the past. Back then the conclusion was that the only safe way forward was Shuttle-C, which placed a mock-shuttle on the ET back. They concluded this was the only way to go, because an ET redesign would be too expensive and the launch sites would have to be rebuilt. 20 years later and none of this applies any more? Riiiight.
Is it just me, or are the "insiders" who can't even spell just a tad less than credible?
Why can't anyone take the announcement at face value? Clearly IBM (and Moto/Freescale) don't want to develop new top-end chips for a small market. Who can blame them?
But Intel is going to build their next generation anyway. Apple's small marketshare is meaningless in this context, they're in a race with AMD for a huge market no matter what else happens.
Let's remember that Intel has been courting Apple for well over a decade now. They're also clearly unhappy with the crappy boxes being offered by their existing vendors. Having Apple onboard making cool products with their systems must be a dream come true -- "See, THIS is what an Intel machine can do".
But no, not enough of a conspiracy in that I suppose.
You know what I hate? Watching one company copy another's program without looking at any other examples for good ideas. This seems to be happening MORE these days, notably in the free software world.
So what WOULD make a good spreadsheet? Here's some ideas...
1) start with Lotus Improv - the key idea here is the separation of sheets, temporary work, and formulas
2) add 3D sheets from Stories, they would fit into Improv's "sheetlette" idea perfectly
3) there's got to be an idea or two from Spreadsheet 2000 worth using
4) Now make every *&%&^% part of it AppleScriptable
Try this experiment: replace every mention of "PowerPC" with "68x" and every mention of "Intel" with "PowerPC". You end up with a factually correct article that predicts the end of Apple because they pre-announced their new order-of-mag-faster machines in advance and are thus killing sales.
Of course the world generally credits the 68k->PPC transition as one of the best handled ever, but that could _never_ work twice, right? I mean, after all, it _didn't_ work once, so that must mean it can never work.
As others have pointed out, this white elephant has been going on for years now. It's what, seven years late and 4 billion over budget so far? There were letters being written about it being only 350 million over and a year late in 1999, and asking serious questions about whether or not it _could_ work. That was five years ago.
And to everyone who thinks it's a great idea that we build this thing for fusion power, consider this: rhe electical industry has already pretty much told the fusion researchers that they can't imagine ever building a fusion plant.
No, really.
While the researchers continue to support their budgets with "cheap unlimited clean power" mantras, the reality is that fusion plants are rediculously expensive, and show no signs that they won't always be so. Additionally they require a fuel infrastructure that does not currently exist, and as the Canadians learned with CANDU, going this route for long-term savings is fraught with peril.
The industry also has the problem that these things only work if they're big. That's a problem because when you bring the plant online, supply-n-damand makes the price of electrity drop. That's one of the reasons no one builds nukes any more, you loose money when you turn them on.
If you're worried about the enviornment, buy a hybrid. Fusion is not your saviour.
My point was limited to the time for the switching itself. Perhaps I should have been more clear on this.
The "at best" is assuming that the forgoing issues don't cause things like cache faults. Passing parameters in registers won't. Thus the performance really can be MUCH worse than twice even in the case of minimal calls.
But even in that case the real-world performance of Mach is, in fact, much worse that twice as slow. I believe it was Chan that did all the big measurement runs, and - going on memory here - BSD on a 50MHz 486DX took 21 usec per syscall, and Mach 3 on the same machine took 114. I may have the number for BSD wrong, that might be the L4 number, BSD might be 9 usec.
The other case is, in fact, even worse. This is the case where the system is constructed of a number of cleanly separated servers running as Mach tasks. In this case a system call may spawn off a series of calls, consider a page fault in the VM for instance. In this case you might end up with a chain of 5 IPC's causing traps. And each one of them is 5-15 times as slow. This is real-world noticable, even for I/O.
One of the papers I have on L4 shows theoretical peak and sustained network throughput for IP on Mach, L4 and BSD. Mach had about 1/10th the performance IIRC. So that "best case" is pretty rare.
> Windows started off microkernel-y but has ended > up rather monolithic (at least partly for performance reasons).
I've heard this claim before, but I'm not sure what to make of it. I always suspected this claim comes from MS's marketting efforts, trying to make it cooler than it is. Remember, back then ukernels were still "cool".
But the only actual example I can think of is the display driver, which used to run outside the kernel. AFAIK most everything else is done internally. And even the display driver used some sort of custom data-passing system, as opposed to a universal IPC, did it not? All of this is fuzzy... forgetting past...
So does Mach, and it's slow. I've never seen real-world measures to suggest that QnX is fast. All we know is that the performance of the OS itself is good, and that's a VERY DIFFERENT measure.
The slow performance is due to a number of problems:
1) not all MMU's are really suited to this task. Many are slower to set up than just copying the memory around. Sun found this to be at around 5k, below that, it was faster to just copy memory physically.
2) MMUs/VM are based on pages, 2 or 4k typically. Thus passing in a single 32-bit int parameter causes big page hits. You can tune this out, but it's still annoying.
3) Each copy takes TWO context switches - one to switch into the kernel to copy the memory across ports, another back out to the called program. This means that even the simplest "system calls" are twice as slow as under a monokernel, AT BEST.
4) Additionally the data has to be examined to see if it contains ports being passed around, and if so, they have to be translated because the port codes are private to a program (and thus different in the other one).
5) Using mapped memory ignores all the hardware specific solutions to these problems, many of which are built into modern processors.
It's exactly the sort of one-size-fits-all solution that you'd expect from a research project. One that doesn't work in the real world. One that should have been replaced, and was in L4, Spring, etc.
For instance, Spring included three different IPC systems, each tuned to certain types of data, each used in different ways on different CPUs. The "fast-path" used a half-switch into the kernel by mapping off registers, allowing IPC to degenerate into register passing largely identical to a procedure call. As long as the message fit within the limitations -- 8 registers, no port identifiers, etc. -- it was faster than a traditional Unix trap. These limitations seem serious, but were in fact used for 80% of calls and 60% of returns (you often say "getDiskSector(integer value)" which could fit into the fast-path, and get back 2k of data which wouldn't).
Although interesting, Mach was developed at a university and shows a huge number of problems as a result. Notably performance is terrible, due largely to the IPC performance. When people actually tried the "collection of servers" operating system in Mach 3, it was clear it was simply not a workable solution. Workplace OS, Star Trek and any number of other OS's died as a result.
What's sad about this is that the failure of Mach tainted ALL ukernels. By the mid-1990s the idea was basically dead. But what an idea! Don't have your machine on a network? Simply don't run the network program. Using a diskless system? Don't run the disk server. Don't want _VM_... no problem. You can use the exact same OS image to build anything from a minimal OS for a handheld to a full-blown multi-machine cluster, without even compiling. No pluggable kernels, no shared libraries, no stackable file systems, nothing but top and ls.
But it just didn't work. IIRC performance of a Unix app on a truly collection-of-servers Mach was 56% slower than BSD. Unusable. Of course you can compile the entire thing into a single app, the "co-located servers" idea, but then all the advantages of Mach go away, every single one.
Now, given this, the question has to be asked: why anyone would still use it? Don't get me wrong, there are real advantages to Mach, notably for Apple who ship a number of multiprocessor machines. But the same support can be added to monokernals. Likewise Apple's version has support for soft realtime, which has also been added to monokernels. So in the end the Mac runs slower than it could, and I am hard pressed to find an upside.
Of course it didn't have to be this way. The problems in Mach led from the development process, not the concepts within. As L4 shows, it is possible to make a cross-platform IPC system that is not a serious drag on performance. And Sun's Spring went further than anyone, really re-writing the entire OS into something I find really interesting, and still providing fast Unix at the same time. I'd love to see someone build Mac OS X on Spring...
Once again we have another example of real science being killed off so NASA can continue it's slavish and visionless mission from before we even landed on the Moon.
So we watch while they desperately try to scape up every amount, no matter how tiny, from worthy missions such as this in order to feed the Space Station and Shuttle programs.
When the post-Apollo era was first being studied NASA came back with a 1-2-3 punch, a space station to study deep-space and long-duration missions, a space shuttle to support cheap, timely and safe crew exchange, both in order to get ready for a mission to Mars. Nixon balked (rightfully) and told them to pick one. They picked the Shuttle, justifying it by saying the cheap access to space would let them go back to the station in the late-70s/early-80s. That turned out well.
What's sad about all of this is that the missions only support each other, neither, on it's own, would have ever made it to bent metal. They built the shuttle to make the station cheap, but when the shuttle turned out to be the most expensive launch system in history, they STILL kept to the original plan -- and now we have the most expensive launch system supplying the most expensive space station. And since the budgets go down (inflation adjusted) every year, NASA has to turn off every other project and feed every dollar into these useless projects.
Someone needs to stop the madness. No one will. What's the sound of freedom? "Oink!"
Once again I find an example of the confusion between gravity and acceleration, or energy and force.
The Moon's gravity is 1/81st that of Earth, because it masses 1/81st as much. It's that simple. Gravity is the result of energy, energy in this case is basically entirely contained in the rest mass, so presto, 1/81st the mass means 1/81st the gravity.
So what about the whole "lunastationary elevator" idea? If there's less gravity, isn't that going to make it much easier? Not really. Think about a line extending from a feature on the moon out into space until it reaches the stationary point -- that is, the point that an orbit remains over that feature on the moon. Where is that point? The Earth. Think about it for a second...
There are other places it could be anchored, the L1 point is often talked about. Yet that is still 58,000km, even longer than Earth's geostationary points at 36,000km. So as am experiment all it does is let you use some other material that's not impossible to make -- but that hardly seems like a good reason to try.
So you need some excuse to make one, and the answer is "there is no reason whatsoever". All of the materials available on the Moon are available for a lot less money here on Earth. Sure, if you have a massive space infrastructure perhaps there is some point at which there is an economy of scale in its favour, but that already presupposes lifting all that stuff off the Earth already.
I found it particularily funny that the author complains about several "problems" in the wiki that he displays right in his own article.
For instance, he notes that he checked the history of the article he chose to examine. Then at the end claims "What he certainly does not know is who has used the facilities before him."
But that's just facts, even more disconcerting is the trust of his article, which he also contradicts.
After noting that the basic idea is that article will approach perfection after many edits, he questions why anyone would believe that. Well I believe it because I've seen it happen in the real world - people edit my articles and most of the time they do get better.
As his example he notes a single article in which 150 edits has resulted in what he considers a poor article (rightfully IMHO). But then he undermines his entire argument by NOT FIXING IT. So basically his complaint appears to be that since he's lazy, and that everone else must be too, that the wiki can't work.
What, do you think coding is different than any other job? How about all those 18th century factory workers at Bolton's button-polishing plants who worked 12 hours a day (or more) and had families of 8? Get over yourself.
Here's how you do it: you go home at five. Every day, period. Wave goodbye to the boss, and say "well I'm off to see the kid". When they say "crunch time", say "see you". When they say "death march" , say "see you".
I told the boss I wasn't coming in till noon twice a week so I could have the kid mornings. Moan, whine, bitch... ok, see you at noon.
You will not lose your job. You will not lose your bonus. You might get a raise, and maybe even a promotion. If you're so insecure at your job that going home at 5 loses it for you, you lost it already.
Face it, you work long hours because you want to. Don't tell me different, I was there too. With a kid you just won't want to any more, so you won't. That's all there is to it.
Devices like this, known generally as "reformers", have beeen in use for a decade at least. They universally share the problem of leaking contaminents into the hydrogen output, where these stray molecules stick to the catalyst inside the fuel cell and slowly degrade it.
If this team has invented a new type of reformer, great, but as it stands the article is a joke.
the awesome Manifold trilogy ?
on
Coalescent
·
· Score: 1
I read the trilogy while on a trip, and frankly I can understand any of the hype. I found the work dull and lifeless, with equally lifeless characters. Moreover the books just didn't make any sense, while it was claimed they were "hard science" the only thing "hard" was trying to understand the gigantic leaps of logic in his apparent attempt to make a linear plotline. Replace any technobabble in the story with any other technobabble and the story remains the same -- that's Star Trek, not hard sci fi.
> Pretty simple, I'd guess. Look up any > information on MIG-25 development. > Shortages of titanium led them to basically > rivet the thing together out of steel plates;
Sigh...
Titantium is one of two materials useable in the Mach 2.5-3.0 regime, _stainless_ steel being the other. The MiG-25, Bristol 188, several Concorde runners-up and the US B-70 all used stainless because it was cheaper, easier to work with, and a known quantity. Titanium was, in the 1960s, a still largely unknown quantity, the development risks were much higher. The fSovs didn't have infinite money to play with, and in that light they made the right choice.
As to them being "plates", one could only use that term if you also call saran wrap "plates" because the material in question was so thin as to be less strong than paper (except in tension). It was only the careful machining and construction that turns into into a structural material. Indeed, the MiG-25 has a fairly good material weight/volume of it's era.
> the air-to-air radar was powered by a bunch > of massive vacuum tubes.
The vast majority of air-search radars today still use at least one tube, as do most radars in general. Note that the vast majority of radars in use today are from the 1970s, the MiG is from the 1960s, so none of this is surprising.
There is a common misconception that transistors replaced tubes everywhere. This is silly. You likely have at least a half-dozen tubes in your house and didn't even realize it. Every TV, computer monitor and microwave oven has one.
> Remember the story of how the US spent $5 > million to develop a space pen
The urban rumor you mean? This is evidence of some point?
> Different design philosophies. I've been > inside a reconstruction of MIR--the thing's > pretty massive, and you definitely get the > feeling that some of the engineers had a > blacksmithing background...
Mir was designed in the 1970s. Spacelab was no different.
The fSovs rarely had different design philosophies. What they had was 1/10th to 1/100th's the money. They didn't have the luxury of being able to try three things and find out which one worked, they had to pick one and make it work the first time. This led to a low-risk approach and longer development times in many situations, but the starting-point technology was largely identical to the west until the 1980s.
"Infrasound is also produced by storms, seasonal winds and weather patterns and some types of earthquakes. Animals such as elephants also use infrasound to communicate over long distances or as weapons to repel foes."
So now we just have to explain how the elephants got into the haunted houses. Or how it is we don't see ghosts every time there's a thundershower.
Seriously, trying to come up with a physical explaination of ghost stories that doesn't include the mind of the person is dumb. The range of reported phenomina is so wide as to be clearly "made up".
Slashdot Mirror
> Dark matter [wikipedia.org] is required by looking at
> galaxy rotation curves.
Which, in every case, suggests a number about 10 times visible mass. I point this out for reasons that will become clear later.
> Dark matter was first hypothesized by Zwicky in 1933
After examining galactic _clusters_. In his analisys, the number was 400 times.
And then you don't note the other argument: that the large-scale structure of the universe cannot be explained without adding lots of mass. The filamental structure of the superclusters are supposed to be gravitationally constructed, but there's no way they could have done so given the visible mass. Using this, we arrive at a figure of about 24x.
Now here's the problem. The _visible_ mass of the universe makes it very close to being energetically "flat", that is, the gravitational energy is very close to the mass energy. This is actually the most "obvious" way for a universe to be built.
So basically dark matter is a _bad_ thing. It ruins any simple model of the universe, and yet requires us to believe in an unseen mass which fails every theory we've thrown at it and remains unseen in experiment.
All this to remain tied to GR, which most QM theorists generally regard as being wrong (ie, a simplification) anyway. And let's not forget, the only evidence we have that GR "works" are local experiments, the more we look out into space, more more fudge factors we have to add.
> Dark energy [wikipedia.org] is required by looking at [snip]
> In about 1998, the supernova observations were pinned down
> well enough to show that the expansion rate is actually
> increasing with time. Therefore, there must be some
> "antigravity" force that causes the expansion of the
> universe to accelerate. This is dubbed "dark energy."
Now perhaps I'm the only one, but this is where I start thinking the whole damb thing is just plain WRONG.
I mean doesn't anyone find it the least bit ironic that the imaginary mass we created ex-nihlo now has to be counterbalanced by _another_ even more wild fudge factor?
This is aether all over again. We're at the "let's pretend the aether gets dragged by mass" stage. Every time a new telescope gets first light you can hear the theories crashing, and the mad scramble to add in yet another fudge factor starts anew.
I mean let's face it, if this latest report is supportable, now we're adding _another_ curve-fit to the original curve-fit because the curve doesn't fit!
> The new claim in the current article is that the effect
> of dark energy has changed over time. The fundamental
> problem is that the new evidence relies on gamma-ray bursts,
> which are not nearly as well established a standard candle
> as the Type 1a supernovae, so it's much harder to say with
> certainty what distance they are at.
Let's be honest, we don't even _really_ know what these things are. Until recently there were arguments they were in fact small events occuring in "local" space (ie, within light years, not mega-light-years). This report seems shaky to me.
Maury
A fine example of the lowest common denominator.
When tasked with the highly difficult job of blowing up a balloon for science, one of the "crew" had a problem...
"Keri told Capcom she couldn't see any instructions on animals.
"Have you taken the instructions out of the envelope?" asked Capcom.
She had not."
*sigh*
Do you know why the Shuttle costs so much and keeps blowing up? Because it's too complex. It consists of four major "parts" that have to be mated in the VAB then carried out to the pad on an rediculous transport. The four "parts" are individually some of the most complex and part-intensive rockets even build. The SRB's, in turn, are fabricated in parts and shipped for assembly.
Failure is strongly related to the number of parts. More parts means more things that can break, and a highly failure rate as a result.
So maybe people out there will understand why I find the idea of strapping five of these things to a new and part-assembled core is an ABSOLUTELY REDICULOUS IDEA.
Sure, SSTO is difficult, but it IS doable. But even if you don't want to spend that much, we know for sure we can do low-risk DSTO like Shuttle II et all. All hail the reduction of development budgets!
This is simply the wrong way to go. Dump Fred into the ocean, where it belongs, and built a _real_ launch system. Enough already!
Maury
You can tell the iPod craze is reaching it's peak if people are willing to beat each other up to get to _buy_ an ancient machine from the same company.
It is broke.
The manned portion consists of an SRB. You know, the big candle thing that keeps blowing up either in-air or on the ground? And now that it goes higher I'm not sure they plan on gettng them back.
The unmanned portion consists of two SRB's (ditto x 2) along with the ET, which was designed for off-axis thrust. In fact it is the off-axis thrust that doomed the wire-round SRB the Air Force worked on, because it was too stiff and couldn't flex enough when the SME's lit up. Now to this we put 100 tonnes of load on top of the ET and 2 million pounds of thrust on the bottom, something it was never designed for (although should have been).
So basically they're going to redesign the ET for changes to the thrust line, which adds nothing of value, but they're NOT going to redesign the SRB which would improve safety and performance.
This has all been looked at in the past. Back then the conclusion was that the only safe way forward was Shuttle-C, which placed a mock-shuttle on the ET back. They concluded this was the only way to go, because an ET redesign would be too expensive and the launch sites would have to be rebuilt. 20 years later and none of this applies any more? Riiiight.
Your tax dollars at work.
I "clearly" didn't read the article? I am "enraged"? The WWDC roadmap and actual hardware is a marketing "anouncment"?
.25, dude, buy a clue.
Here's
Is it just me, or are the "insiders" who can't even spell just a tad less than credible?
Why can't anyone take the announcement at face value? Clearly IBM (and Moto/Freescale) don't want to develop new top-end chips for a small market. Who can blame them?
But Intel is going to build their next generation anyway. Apple's small marketshare is meaningless in this context, they're in a race with AMD for a huge market no matter what else happens.
Let's remember that Intel has been courting Apple for well over a decade now. They're also clearly unhappy with the crappy boxes being offered by their existing vendors. Having Apple onboard making cool products with their systems must be a dream come true -- "See, THIS is what an Intel machine can do".
But no, not enough of a conspiracy in that I suppose.
You know what I hate? Watching one company copy another's program without looking at any other examples for good ideas. This seems to be happening MORE these days, notably in the free software world.
So what WOULD make a good spreadsheet? Here's some ideas...
1) start with Lotus Improv - the key idea here is the separation of sheets, temporary work, and formulas
2) add 3D sheets from Stories, they would fit into Improv's "sheetlette" idea perfectly
3) there's got to be an idea or two from Spreadsheet 2000 worth using
4) Now make every *&%&^% part of it AppleScriptable
THAT is the spreadsheet you want.
Try this experiment: replace every mention of "PowerPC" with "68x" and every mention of "Intel" with "PowerPC". You end up with a factually correct article that predicts the end of Apple because they pre-announced their new order-of-mag-faster machines in advance and are thus killing sales.
Of course the world generally credits the 68k->PPC transition as one of the best handled ever, but that could _never_ work twice, right? I mean, after all, it _didn't_ work once, so that must mean it can never work.
I would really like to see a comparison using better compilers though, Intel on x86 and IBM on Power.
As others have pointed out, this white elephant has been going on for years now. It's what, seven years late and 4 billion over budget so far? There were letters being written about it being only 350 million over and a year late in 1999, and asking serious questions about whether or not it _could_ work. That was five years ago.
And to everyone who thinks it's a great idea that we build this thing for fusion power, consider this: rhe electical industry has already pretty much told the fusion researchers that they can't imagine ever building a fusion plant.
No, really.
While the researchers continue to support their budgets with "cheap unlimited clean power" mantras, the reality is that fusion plants are rediculously expensive, and show no signs that they won't always be so. Additionally they require a fuel infrastructure that does not currently exist, and as the Canadians learned with CANDU, going this route for long-term savings is fraught with peril.
The industry also has the problem that these things only work if they're big. That's a problem because when you bring the plant online, supply-n-damand makes the price of electrity drop. That's one of the reasons no one builds nukes any more, you loose money when you turn them on.
If you're worried about the enviornment, buy a hybrid. Fusion is not your saviour.
My point was limited to the time for the switching itself. Perhaps I should have been more clear on this.
The "at best" is assuming that the forgoing issues don't cause things like cache faults. Passing parameters in registers won't. Thus the performance really can be MUCH worse than twice even in the case of minimal calls.
But even in that case the real-world performance of Mach is, in fact, much worse that twice as slow. I believe it was Chan that did all the big measurement runs, and - going on memory here - BSD on a 50MHz 486DX took 21 usec per syscall, and Mach 3 on the same machine took 114. I may have the number for BSD wrong, that might be the L4 number, BSD might be 9 usec.
The other case is, in fact, even worse. This is the case where the system is constructed of a number of cleanly separated servers running as Mach tasks. In this case a system call may spawn off a series of calls, consider a page fault in the VM for instance. In this case you might end up with a chain of 5 IPC's causing traps. And each one of them is 5-15 times as slow. This is real-world noticable, even for I/O.
One of the papers I have on L4 shows theoretical peak and sustained network throughput for IP on Mach, L4 and BSD. Mach had about 1/10th the performance IIRC. So that "best case" is pretty rare.
Maury
> Windows started off microkernel-y but has ended
> up rather monolithic (at least partly for performance reasons).
I've heard this claim before, but I'm not sure what to make of it. I always suspected this claim comes from MS's marketting efforts, trying to make it cooler than it is. Remember, back then ukernels were still "cool".
But the only actual example I can think of is the display driver, which used to run outside the kernel. AFAIK most everything else is done internally. And even the display driver used some sort of custom data-passing system, as opposed to a universal IPC, did it not? All of this is fuzzy... forgetting past...
Maury
> QNX uses shared memory to pass messages
So does Mach, and it's slow. I've never seen real-world measures to suggest that QnX is fast. All we know is that the performance of the OS itself is good, and that's a VERY DIFFERENT measure.
The slow performance is due to a number of problems:
1) not all MMU's are really suited to this task. Many are slower to set up than just copying the memory around. Sun found this to be at around 5k, below that, it was faster to just copy memory physically.
2) MMUs/VM are based on pages, 2 or 4k typically. Thus passing in a single 32-bit int parameter causes big page hits. You can tune this out, but it's still annoying.
3) Each copy takes TWO context switches - one to switch into the kernel to copy the memory across ports, another back out to the called program. This means that even the simplest "system calls" are twice as slow as under a monokernel, AT BEST.
4) Additionally the data has to be examined to see if it contains ports being passed around, and if so, they have to be translated because the port codes are private to a program (and thus different in the other one).
5) Using mapped memory ignores all the hardware specific solutions to these problems, many of which are built into modern processors.
It's exactly the sort of one-size-fits-all solution that you'd expect from a research project. One that doesn't work in the real world. One that should have been replaced, and was in L4, Spring, etc.
For instance, Spring included three different IPC systems, each tuned to certain types of data, each used in different ways on different CPUs. The "fast-path" used a half-switch into the kernel by mapping off registers, allowing IPC to degenerate into register passing largely identical to a procedure call. As long as the message fit within the limitations -- 8 registers, no port identifiers, etc. -- it was faster than a traditional Unix trap. These limitations seem serious, but were in fact used for 80% of calls and 60% of returns (you often say "getDiskSector(integer value)" which could fit into the fast-path, and get back 2k of data which wouldn't).
Maury
Although interesting, Mach was developed at a university and shows a huge number of problems as a result. Notably performance is terrible, due largely to the IPC performance. When people actually tried the "collection of servers" operating system in Mach 3, it was clear it was simply not a workable solution. Workplace OS, Star Trek and any number of other OS's died as a result.
What's sad about this is that the failure of Mach tainted ALL ukernels. By the mid-1990s the idea was basically dead. But what an idea! Don't have your machine on a network? Simply don't run the network program. Using a diskless system? Don't run the disk server. Don't want _VM_... no problem. You can use the exact same OS image to build anything from a minimal OS for a handheld to a full-blown multi-machine cluster, without even compiling. No pluggable kernels, no shared libraries, no stackable file systems, nothing but top and ls.
But it just didn't work. IIRC performance of a Unix app on a truly collection-of-servers Mach was 56% slower than BSD. Unusable. Of course you can compile the entire thing into a single app, the "co-located servers" idea, but then all the advantages of Mach go away, every single one.
Now, given this, the question has to be asked: why anyone would still use it? Don't get me wrong, there are real advantages to Mach, notably for Apple who ship a number of multiprocessor machines. But the same support can be added to monokernals. Likewise Apple's version has support for soft realtime, which has also been added to monokernels. So in the end the Mac runs slower than it could, and I am hard pressed to find an upside.
Of course it didn't have to be this way. The problems in Mach led from the development process, not the concepts within. As L4 shows, it is possible to make a cross-platform IPC system that is not a serious drag on performance. And Sun's Spring went further than anyone, really re-writing the entire OS into something I find really interesting, and still providing fast Unix at the same time. I'd love to see someone build Mac OS X on Spring...
> Why should MSFT be held to some high standard
> for a tool that they include in their software?
You're kidding, right?
Maury
Once again we have another example of real science being killed off so NASA can continue it's slavish and visionless mission from before we even landed on the Moon.
So we watch while they desperately try to scape up every amount, no matter how tiny, from worthy missions such as this in order to feed the Space Station and Shuttle programs.
When the post-Apollo era was first being studied NASA came back with a 1-2-3 punch, a space station to study deep-space and long-duration missions, a space shuttle to support cheap, timely and safe crew exchange, both in order to get ready for a mission to Mars. Nixon balked (rightfully) and told them to pick one. They picked the Shuttle, justifying it by saying the cheap access to space would let them go back to the station in the late-70s/early-80s. That turned out well.
What's sad about all of this is that the missions only support each other, neither, on it's own, would have ever made it to bent metal. They built the shuttle to make the station cheap, but when the shuttle turned out to be the most expensive launch system in history, they STILL kept to the original plan -- and now we have the most expensive launch system supplying the most expensive space station. And since the budgets go down (inflation adjusted) every year, NASA has to turn off every other project and feed every dollar into these useless projects.
Someone needs to stop the madness. No one will. What's the sound of freedom? "Oink!"
Once again I find an example of the confusion between gravity and acceleration, or energy and force.
The Moon's gravity is 1/81st that of Earth, because it masses 1/81st as much. It's that simple. Gravity is the result of energy, energy in this case is basically entirely contained in the rest mass, so presto, 1/81st the mass means 1/81st the gravity.
So what about the whole "lunastationary elevator" idea? If there's less gravity, isn't that going to make it much easier? Not really. Think about a line extending from a feature on the moon out into space until it reaches the stationary point -- that is, the point that an orbit remains over that feature on the moon. Where is that point? The Earth. Think about it for a second...
There are other places it could be anchored, the L1 point is often talked about. Yet that is still 58,000km, even longer than Earth's geostationary points at 36,000km. So as am experiment all it does is let you use some other material that's not impossible to make -- but that hardly seems like a good reason to try.
So you need some excuse to make one, and the answer is "there is no reason whatsoever". All of the materials available on the Moon are available for a lot less money here on Earth. Sure, if you have a massive space infrastructure perhaps there is some point at which there is an economy of scale in its favour, but that already presupposes lifting all that stuff off the Earth already.
I found it particularily funny that the author complains about several "problems" in the wiki that he displays right in his own article.
For instance, he notes that he checked the history of the article he chose to examine. Then at the end claims "What he certainly does not know is who has used the facilities before him."
But that's just facts, even more disconcerting is the trust of his article, which he also contradicts.
After noting that the basic idea is that article will approach perfection after many edits, he questions why anyone would believe that. Well I believe it because I've seen it happen in the real world - people edit my articles and most of the time they do get better.
As his example he notes a single article in which 150 edits has resulted in what he considers a poor article (rightfully IMHO). But then he undermines his entire argument by NOT FIXING IT. So basically his complaint appears to be that since he's lazy, and that everone else must be too, that the wiki can't work.
Perhaps he should look up "irony".
What, do you think coding is different than any other job? How about all those 18th century factory workers at Bolton's button-polishing plants who worked 12 hours a day (or more) and had families of 8? Get over yourself.
Here's how you do it: you go home at five. Every day, period. Wave goodbye to the boss, and say "well I'm off to see the kid". When they say "crunch time", say "see you". When they say "death march" , say "see you".
I told the boss I wasn't coming in till noon twice a week so I could have the kid mornings. Moan, whine, bitch... ok, see you at noon.
You will not lose your job. You will not lose your bonus. You might get a raise, and maybe even a promotion. If you're so insecure at your job that going home at 5 loses it for you, you lost it already.
Face it, you work long hours because you want to. Don't tell me different, I was there too. With a kid you just won't want to any more, so you won't. That's all there is to it.
Devices like this, known generally as "reformers", have beeen in use for a decade at least. They universally share the problem of leaking contaminents into the hydrogen output, where these stray molecules stick to the catalyst inside the fuel cell and slowly degrade it.
If this team has invented a new type of reformer, great, but as it stands the article is a joke.
I read the trilogy while on a trip, and frankly I can understand any of the hype. I found the work dull and lifeless, with equally lifeless characters. Moreover the books just didn't make any sense, while it was claimed they were "hard science" the only thing "hard" was trying to understand the gigantic leaps of logic in his apparent attempt to make a linear plotline. Replace any technobabble in the story with any other technobabble and the story remains the same -- that's Star Trek, not hard sci fi.
> Pretty simple, I'd guess. Look up any
> information on MIG-25 development.
> Shortages of titanium led them to basically
> rivet the thing together out of steel plates;
Sigh...
Titantium is one of two materials useable in the Mach 2.5-3.0 regime, _stainless_ steel being the other. The MiG-25, Bristol 188, several Concorde runners-up and the US B-70 all used stainless because it was cheaper, easier to work with, and a known quantity. Titanium was, in the 1960s, a still largely unknown quantity, the development risks were much higher. The fSovs didn't have infinite money to play with, and in that light they made the right choice.
As to them being "plates", one could only use that term if you also call saran wrap "plates" because the material in question was so thin as to be less strong than paper (except in tension). It was only the careful machining and construction that turns into into a structural material. Indeed, the MiG-25 has a fairly good material weight/volume of it's era.
> the air-to-air radar was powered by a bunch
> of massive vacuum tubes.
The vast majority of air-search radars today still use at least one tube, as do most radars in general. Note that the vast majority of radars in use today are from the 1970s, the MiG is from the 1960s, so none of this is surprising.
There is a common misconception that transistors replaced tubes everywhere. This is silly. You likely have at least a half-dozen tubes in your house and didn't even realize it. Every TV, computer monitor and microwave oven has one.
> Remember the story of how the US spent $5
> million to develop a space pen
The urban rumor you mean? This is evidence of some point?
> Different design philosophies. I've been
> inside a reconstruction of MIR--the thing's
> pretty massive, and you definitely get the
> feeling that some of the engineers had a
> blacksmithing background...
Mir was designed in the 1970s. Spacelab was no different.
The fSovs rarely had different design philosophies. What they had was 1/10th to 1/100th's the money. They didn't have the luxury of being able to try three things and find out which one worked, they had to pick one and make it work the first time. This led to a low-risk approach and longer development times in many situations, but the starting-point technology was largely identical to the west until the 1980s.
"Infrasound is also produced by storms, seasonal winds and weather patterns and some types of earthquakes. Animals such as elephants also use infrasound to communicate over long distances or as weapons to repel foes."
So now we just have to explain how the elephants got into the haunted houses. Or how it is we don't see ghosts every time there's a thundershower.
Seriously, trying to come up with a physical explaination of ghost stories that doesn't include the mind of the person is dumb. The range of reported phenomina is so wide as to be clearly "made up".