Actually, I'm not making a category mistake (I even noted in my original post that the things I was using as examples were not necessarily interchangeable), but you've made my point for me.
As you note, the real power of distributed/parallel computing comes from the message passing libraries, most commonly MPI or PVM. Beowulf per se is almost nothing more than a label for the generic concept of distributed computing on Linux. The same thing can be done with any other reasonably modern networked computer you have lying around, even those running Windows - you can even mix OSes in a cluster, although this introduces new and interesting problems. (There are a serious lot of underutilzed cycles sitting out there on the corporate world's desktops if they're not running OpenGL screensavers...)
BTW: If the phenomenal success of Sun's E10000 Starfire has taught us anything at all, it's that where I/O is important, a big honkin' SMP box kicks cluster butt! Seriously, the interconnect technology between boxes just *can't* be fast enough to compete effectively with a huge multi-level crossbar packet switch like the ones in the E10K. Sun and the other SMP vendors can win here because they own the domain in which the simpler problem resides...
Don't assume by this that I'm against Beowulf clusters at all - they are a great and amazing thing, but there's more than one way to skin a cat, and Beowulf isn't the only path to Linux distributed computing.
I'll bet you a big bag full of Helium-3 filled buckyballs that your PC or calculator is obsolete before the spiral notebook.
In fact, I'll bet the entire concept of PCs is obsolete before the spiral notebook.
Oh, and even acid-based paper spiral notebooks can be reasonably expected to preserve their information for a century or so with no power and little liklihood of the data format becoming unreadable.
Somehow, that spiral notebook doesn't look so obsolete after all, does it?
NO! If you really know the material, the time limit is never a problem. (See my post above - knowing where/how to find info is not at all the same as knowing it, which implies it is available for immediate recall and use.)
I always loved those exams (too few and far between, admittedly), where I knew the material cold, walked in, aced the test in 20 minutes, checked my answers for another 10 minutes, and walked out confident while those that didn't know and understand the material were sweating it through to the bell.
BTW, one of my better profs used the length of the exam, even in an open book situation, as the mechanism to separate those that knew from those that didn't. If you knew the material, you had time to finish - if you didn't, you soon found you didn't have enough time to look everything up. Excellent tests, excellent testing method, and they produced a class of students that did know the material.
Re:Calulators in math class
on
Laptop Exams?
·
· Score: 2
They are nifty inventions, but most of the people ere couldn't build one....
O.K., for all of you saying the calculators/computers are a good thing, try this challenge: Without using your calculator or computer, design (and build, if you feel this is too easy) a fully functioning slide rule. I don't think there are very many college students today with an understanding of logarithms (big hint there) adequate to that task...
Re:Curious fact that I have noticed
on
Laptop Exams?
·
· Score: 2
This is absolutely true. Without exception, every "really good" engineer, scientist, or other professional I've ever worked with had one thing in common: an understanding of the subject so comprehensive that they are at all times able to take any aspect of the discussion all the way back to first principles.
This is the kind of knowledge that allows one to *easily* derive the formulae they haven't memorized, and it's the kind of knowledge one will never develop using the crutch of a calculator to generate symbolic solutions to calculus problems. (BTW: I speak as someone who used such a crutch, and then had to learn the error of my ways once I saw what real engineers (actual rocket scientists, some of them) were expected to know and do.) Schools that allow students to use such aids are robbing their students of the real education they should be receiving.
Knowing how/where to find the answer is a very poor substitute for knowing the answer. Serious thinking is not possible without that knowing.
The article doesn't say, but despite what you'd think by reading the rantings of the ill-informed 3l337 d00dZ on slashdot, Beowulf isn't even a very good clustering technology for most problems.
There are far more serious, industrial-strength solutions out there, things like MPI, PVM, LINDA,and IBM's own HACMP. (Note these cover a lot of ground and are not necessarily even comparable to one another.)
Beowulf (or any of the others listed above) is not automatically the correct distributed computing methodology. Selecting the proper solution for the job at hand is far more complex than you might imagine. There is a lot more developer activity on some of these than there is on Beowulf - MPI in particular is maturing rapidly and is used for solving big/tough problems in many of the largest companies in the world. (No particular MPI advocacy or bias, it just seems like I run into it more often than the others...)
I'm really surprised at the number of people on/. that are so adamantly opposed to the use of personal electronics while driving. (You sound more like Luddites than geeks...)
Seriously, there was a huge debate early in this century about whether or not *radios* should be allowed in cars because they were so distracting to the driver and would surely lead to hair, teeth, and eyeballs all over the road. After all, who could possibly concentrate on driving while they were listening to music?
A recent issue of American Heritage's Invention and Technology magazine (required reading for geeks, IMHO) carries a story on the numerous technical problems involved in getting radios to work in cars and a reference to the anti-car-radio) forces. There were even laws passed in some locales prohibiting car radios in an effort to prevent the inevitable roadside carnage. Fortunately cooler heads and the eventual availabilty of factory car radios prevailed. (It was the radical idea of a factory car radio that Bill Lear (later the father of the Lear Jet, and the 8-track tape) used to start a company you may have heard of: Motorola!)
In spite of the doomsayers, we've managed to survive reasonably well with not only radios, but also tape and CD players in our cars, and few people consider this to be an undue hazard. Hopefully, this attitude that bashes cellphones and other personal electronics use in cars will die off just as did the bias against radios. In a few years, the idea of not being able to talk on the phone while driving will seem as ludicrous as not being able to listen to your tunes while driving. (And to many of us, it already does...)
Most people have no idea how difficult this is. In fact, this stuff really *is* rocket science!
The moon missions had to hit a date/time-specific moon corridor that was only about 10 miles wide with a velocity tolerance of only about 100 mph. There was one mid-course correction that they could use to finesse position within the corridor, but they had to have hit it in the first place. Think about it for a second, and you'll realize those are *very* tight tolerances, in context.
(BTW: the figures above are from the Ranger missions. Apollo may have used slightly different paths, but the physics is pretty much the same regardless. This is one of the reasons it was such a miracle that we got Apollo 13 back at all. I used to do a lot of work at JSC in Houston, and several of the oldtimers who worked on 13 told me that Mission Control intentionally aimed the capsule along the steepest allowable side of the flight path, so that if something went wrong, it would result in the astronauts being vaporized rather than skipped off into space to asphixiate, which was viewed as being as bad for PR as it would be for the astronauts.)
There are something like 60-odd of them and they're not geostationary so they swoop by on a regular basis. (They were parked in LEO to save money.)
This is completely bogus. Does *anyone* here bother to understand the facts before posting thier ill-informed opinions.
There is a VERY GOOD technical reason why Iridium satellites are in LEO - and it's quite simple: You want them low to avoid tha time-of-flight problems that plague geostationary satellites. (I used to do geosync satellite protocol tuning - TCP/IP wasn't even capable of dealing with geosyncronous delays until the RFC 1323 enhancements became commonplace. (It took Sun forever to put these in Solaris.) The delays are on the order of thousands of milliseconds - that's right, whole seconds. More bandwidth only hurts the problem. The only things you can do that really help are: 1) shorten the communication path, or, 2) speed up the signal. Mr. Einstein says you can't do the second even if you want to.;-)
It's true that LEO birds are cheaper because they don't need the expensive boosters to geosync (23,000 miles is *way* the heck out there - nearly 3 Earth diameters!), but the chief reason Iridium birds are in LEO is to achieve reasonable signal latencies.
It's a shame Iridium got the moey first and dd it wrong, because the concept is excellent. Iridium just underestimated bandwidth requirements of the network by a couple of orders of magnitude, which put a serious crimp in their pricing model.
Actually, electric cars would be considerably worse for the environment than our current IC engines. Remember, they're not zero-emitting, they're just remote-emitting, and that's after losing most tof the power generated to the hideous inefficiencies of the electric's systems.
I know some of you folks love to bash gasoline engines, and no, I don't think we'll use them forever myself - although I'm betting it's >50 years before we see a dent in thier dominance.
But you have to recognize that technology has improved the performance, efficiency, and cleanliness of gasoline engines by several orders of magnitude, while electrics remain steadfastly immune to technological progress. Remember that gasoline engines weren't given a ghost of a chance at the turn of the century (that would be the 1901 one) - *everyone* knew that steam and/or electric would trounce the stinking, noisy IC motors. But what happened is instructive to all students of technology: the gasoline engine, for all its faults, became the best option - to the point that its competition quit trying. This happened not because of a conspiracy by Detroit and the Oils, but because unlike other powerplants, the ungainly IC engine continues to respond well to technological attention, and seems to show no sign of slowing down its progress anytime soon. At tthe same time, we have only proven that electric cars are as unviable as ever, and that even spending ridiculously high sums on them results in vehicles that are still seriously sub-par compared to their gasoline eqiuvalents. Electric cars are scarcely more viable today than they were in 1920.
Today's gasoline engines are remarkably efficient and clean, with emissions that are effectively zero after the catalyst is warmed up. Today's running emissions would be off the bottom of the scale for cars of just twenty-five years ago. Further, efficency is now remarkable - Honda's new (normally aspirated!) S2000 roadster has a specific output better than many *race cars* of the 1960's!
And don't forget that we don't have excess electricity generating capacity out there burning a hole in our pocket, either. Most utilities are seriously encouraging conservation as a method of delaying the construction of new power plants, which is becoming horrendously expensive due to new environmental rules. Even if the electric car side of the equation were solved, it would likely be a very long time before electrics became economically viable as an alternative, since sharply increased demand would soon cause skyrocketing electric rates. We've gotten very, very good at producing, distributing, and using fossil fuels, and you don't replace that infrastructure overnight.
Oh, and don't forget that there is very little energy lost in transporting fossil fuels, but considerable loss even in the best high-voltage transmission lines we can make...
Electric isn't likely to be the answer anytime soon. (Pure electric, that is - hybrids may be sort of viable sometime soon, although they'll always cost more.)
I am doing and have done several similar "conversion" projects with embedded computers and Linux.
My first conversion of this type (other than laptops, which I was running Linux on way before it was cool, and when video drivers were a really ugly problem) was an Epson IM-403 cash register computer. (Available from Timeline, the surplus guys, for $99.) It's got a lot less stones than the iOpener, and no screen, but it does have a nifty if tiny little UPS that lets it ride out most power glitches. I've posted about this project here on/. before, but in a nutshell: - 486SX33 CPU, pretty nice little BIOS - nice little Chips and Tech video controller will do 800x600 max to std VGA connector - four serial ports (great for control projects!) - a parallel port - socket/tray for a 2.5" IDE notebook hard disk (will accept even the tall 17 or 19mm drives, so you can put IBM's latest monster in it) - a single ISA half-card slot for the obligatory network card. - Socket for up to 32MB of RAM (plenty for a decent Linux system - I've run several versions of Caldera on the box, but recommend sticking to slim WMs for obvious reasons.) - Flat ribbon connector for Epson floppy (know where to order one, if you need it.) - Tiny little NiCad UPS built in. This is one of the coolest features, and the reason one of these is my primary file server at home. It will only keep the box alive for a few minutes, but that's always been enough so far.
These run Linux fine, or if you just need vNC, you can run the DOS vNC client with the packet drivers or a stripped Linux with the SVGALIB vNC, either of which should fit on a floppy. [As an aside, I have one extra IM-403 with a 3GB HDD that I'll sell for $120+S&H in the US, if anyone's interested.]
(BTW, anybody know where I can get a multiport network card (2 or more Ethers, 1 half-size ISA slot) that works under Linux? If so, I've got my new firewall...)
Also, I'm just starting to hack on a little ARM-based *touchscreen* webphone. These were built at enormous cost by AT&T and Philips before they decided they cost more than the market would bear. (They would have had to sell them for about $600.) They run AT&T's Inferno operating system, and are pretty darn nice speakerphones in addition to finally having the holy grail I've been searching for: a *touchscreen*. I think I can lay my hands on a couple hundred more of these at ~$200 each +S&H, if there's interest. They aren't speedy, but they have a built-in web browser and use one of those little WebTV-style wireless keyboards. Although I don't know of anyone that's replaced the OS (althogh ARM Linux is out there...), but there are hacking instructions available that show how to get root in Inferno. They don't have E-net connections, but do have two PCMCIA slots, so decent networking should be possible.
I've done a lot of hacking with laptop hard disks, and it's extraordianrily difficult to get the 44-pin cables for these things, especially if you want a cable to support two of the little drives. (Which I'd like to do for several machines I've got, including my FIC Sahara Databook, which uses a notebook CDROM on the secondary IDE channel that could give me a place to hang a backup drive. I can't even find anyplace to get a cable like that *made* anywhere here in Austin! You can get the connectors from Jameco, but I haven't found out where to get the dinkier ribon cable these need. (The 44-pin IDE 2.5" drive header uses 2mm pin spacing rather than the standard.1" used by pretty much all other header connectors - damn metric system!)
Also, a local computer shop once had sockets and little plastic cases to allow a 2.5" HDD to be plugged into a 3.5" drive bay socket. They don't have them anymore. Anybody stumble across these in your searches?
The points in my original submission or this story were edited by the/. staff - most of it was done quite well and added important content, but a couple of things were lost.
I made the point that the CNET article did not explicitly reference design patents, but I expected that Apple and others had them and could be expected to enforce them, sometimes to the detriment of consumers. The point was to open a more general discussion of alternative mechanisms which could be used to render technologies proprietary. (Remember the big flap a couple of years ago - Motorola sued Qualcomm for building a phone that folded, saying it violated their design patents on the StarTAC. Fortunately, the courts ruled that MOT couldn't own the generic idea of a folding phone (since everyone knew, and their product name even alluded to, the fact that they stole the idea from Star Trek.) Still, the MOT attack delayed QCOM's Q-phone until it was no longer a serious competitive threat. Not that that's hurt their stock price any...[grin])
You're right that the CNET article doesn't have anything to do with patents, but I was intentionally trying to raise the issue of some of these other mechanisms by which we can expect companies large and small to attempt to excercise control or block competitors. As I was quoted in Timothy's post, I'm particularly concerned about potential abuse of design patents. (BTW, I am on record here on/. as a heretic (being in general a supporter of our current patent system) and feel strongly that much of the "reform" wanted by the community here would simply remove any viable protections for small/startup/entrepreneurial companies while leaving the large ones protected by thier inherent economic/political might. For all its warts, the US patent system is one of the most effective systems in history to ensure a level playing field for all, so we should tamper with it only with extreme caution and certainty.)
Also, my article submission indicated that I thought this particular decision was well-justified, since there was a clear attempt to copy the iMac, even though I don't think for a minute that anyone would confuse these Windows boxes for real Macintoshes. Still, even if a good call in this case, this could be a slippery slope since it is necessarily subjective.
Finally, if this sort of limitation becomes viewed as effective, I raised the spectre of ID squatters protecting everything they can think of this way and then selling all the good and necessary ones off to the highest bidder. (I considered doing this myself a number of years ago, since PDA's still don't have nearly the functionality I described in writing nearly a decade ago. If I'm not going to get venture money myself, then I could at least sue someone who did - but that's really cheezy now isn't it? And yes, in that context, a "z" in "cheezy" seems just right...)
The point of discussion I was trying to provoke is, "Are there any effective safegaurds against abuse of other parts of our IP system that may not have recieved as much scrutiny as patents for software?"
In this context, in contrast to what some other posters seem to think, this topic is appropriate for/., and does, in fact, matter.
Oh, and for the record, I think most of these clear and whatever things are kind of ugly, although the ones that are plain clear (no other colors, like the Palm IIIe) are interesting. And as an aside, it's a shame Apple didn't learn from DEC's experience with puck mice - they used them on their VAXstations and DECstations at least up until the Alpha came out, and they also were a pain to grab and use without looking.
I'm sure he already has. It's funny, but a liberal arts degree, because it's based on more classical content, is one of the most conservative educations possible.
Remember, all conservatives are "classical liberals", so you need to keep time context in mind and avoid knee-jerk use of or reaction to the term "liberal"...
I'd second the fact that digital interfaces are important. Most of the problems I've encountered with flat screens and regular analog (*VGA) connections center around the difficulty inherent in redigitizing the analog signal that the RAMDAC on your video card just thoughtlessly created, and mapping that *cleanly* and solidly into the panel's native pixel grid.
I've seen some very expensive LCD screns that look awful because they aren't able to perfectly re-gen the digital signal. Most analog-input flat-panel monitors can be made to work well with a singel computer, but on the rare occasions I've tried to use a vid/kbd/mouse switch box, I've found that it's pretty much impossible to come up with a setting which will look good with the input from differing video cards. (It might be better if they were identical, or it might not...)
The problem is that there's not a lot of support for the digital interfaces yet even on the hardware side, and you may (although I haven't checked) have problems finding video drivers for XFree86 or any other non-MS display system.
This seems to be kind of a chickenand-egg problem - no one can buy digital interface flat screens until someone builds them, and no one will build them (and price them sanely) until there's a large proven market.
I've gone through this set of gyrations myself, to the point that I may just give up and buy a Gateway Profile just to avoid the hassle.
I wish there were more, better, and cheaper choices for flat screens. It galls me that it's nearly as cheap to buy an all-in-one like the Profile, or one of the new "suitcase-luggable" laptops like Dell's with the big screens than to add a good digital flatscreen and video card to my existing PC. Something's very wrong here...
Actually, radio *was* invented here. This was settled legally in a sensational lawsuit (which was appealed IIRC to or near the Supreme Court) between Nicola Tesla (who really invented it first), and Gulielmo Marconi (who independently invented it a short time later and proceeded to devlop it much faster than Tesla.)
Although Tesla was not an American by birth, he got here as fast as he could!
This is non-news if I ever heard it. CPU cores and even complete CPU designs have been free or near-free for years.
Sun started giving away SPARC uP designs last year - before that they were charging a stunning $99 for them.
Oh, there is the catch that in order to actually build the chips, you'll need a modern stae-of-the-art semiconductor fab facility, which should run you about US$3e9. Or you could spring for a few tens or hundreds of millions and buy a slice of some else's fab, if they have excess capacity. (I don't expect there's much of that around right now...)
CPU designs have been efectively free for a long time now, and that hasn't changed much. Ross was the last company to seriously try to capitalize on this idea and look where it got them!
A discussion of computer animation or even portrayals of future tech would have been appropriate, news of casting decisions is not. I stand by my original gripe.
As I've mentioned here before, I've got a copy of an old Interface Age magazine in my garage that has in it a "floppy ROM" - one of those cheesy flexible LPs they used to put in magazines and books that contains a BASIC ROM for an old Z-80 or 8080 computer.
Although I never used it, I know people that did. (I didn't have a computer then, I just read about them and dreamed about the cool stuff from Imsai, Altair, The Digital Group (the coolest looking computers), Compucolor (wow, color graphics!), and SW Technical Products (possibly history's ugliest terminal.))
Back then there was something called the "Kansas City Standard" which was a big deal at the time because it brought interoperability to audio cassette tape data storage. The folks who did Flexi-LPs like these just pressed the KC Std. audio onto the disc, then all you needed was the correct adapter to plug the audio output from the turntable (well, a pre-amp, actually, since you didn't get line-level output from most turntables) to the KC interface box that you would normally plug into the cassette player.
It wasn't fast, but it was pretty robust: there was enough slop in the spec that you could usually read copies of copies of copies of tapes. (We're not talking quality copies here, folks, just hooking together whatever two cassette decks you could find.)
1. Bluetooth is not meant to be a wireless LAN replacement (not without a gazillion "access points", anyway.) It's intended to be a cordless "desktop area network", and RF bubble that encircles Bluetooth devices allowing them to form ad hoc connections wiht one another for both data and isochronous (e.g., voice) connections. One node you connect to may optionally act as a gateway (access point) to the "real" network, if it's connected and wants to offer that service. As a transitory, ad hoc, solution, 10m is just about the most range you'd want - any more would create problems, and personally, I think they made the bubble too big. Bluetooth isn't meant for seamless raoming wireless connections, but to facilitate connections to and between things like telephone sets (wired or wireless), PDAs, desktop computers, etc.
2. 2.4 GHz doesn't belong to either the Bluetooth or the 802.11 guys. In the US, 2.4 is one of the ISM (Industrial, Scientific, and Medical) bands (others are at 900 MHz and 5.7 GHz.) These are for unlicensed use providing that the FCC's restrictions are met. These restrictions include accepting any interference without recourse - if you want recourse, you go for licensed spectrum! 2.4 GHz was chosen because with minor adjustments, it's available for unlicensed use in pretty much every country that matters. (And it's the only reasonably low frequency for which this is true...)
BTW - there are lots of other perfectly legal 2.4 GHz devices which can kill your wireless LAN, so don't single out Bluetooth unfairly...
I know there's been a history of abuse w.r.t. this topic by the/. owners, but Star Wars rumors are neither "News for Nerds" nor "Stuff that matters".
Save the Star Wars rumors for E!online, or maybe nationalenquirer.com. Apparently, now that the site is fully commercialized, the/. folks prefer to post whatever articles will result in hits and generate endless drivel-filled opinion posts than actually facilitate tech discussions with any real content.
I'm starting to seriously question whether Slashdot is worth my time anymore. Tragedy of the commons is taking its grisly toll and the S/N ratio continues to plummet. Does anyone care?
I would love to use voice recognition in an appropriate capacity, but it's not going to happen soon, if ever.
(Note also that using VoRec correctly requires a near total redesign of user interfaces to accomodate correlation of recognized voice commands and gestural input - notice how often you point or make another meaningful motion when you talk to someone, especially in trying to convey information, and that those gestures mean very different things from the way interfaces use pointing today.)
It's really sad, but today's best voice recognition is only marginally more useful than the CoVox voice recognition I had on my Commodore 64 thirteen years ago. Sure, vocabularies are larger now, and we can get somewhat closer to true connected speech recognition, but by and large, the lack of progress proves that despite an improvement of several orders of magnitude in processing power, (and memory speed and capacity, and storage space, not to mention DSPs) voice recognition is not a problem that can be effectively addressed by any amount of brute force processing using current methods.
I'm going to really date myself here: I've got a copy of Interface Age magazine (anybody else remember that?) in the garage (the one with the floppy ROM: a BASIC ROM on one of those floppy vinyl records they used to put in magazines) that has ads for several voice recognition systems that didn't work all that much worse that the ones we have now - I had friends that had some, back when everybody was using the S-100 bus. Speech recognition has been "just around the corner" in the computer/hacker community for as long as I can remember, and the VoRec prognosticators have been hoovering money for years out of suckers of either the investor or customer ilk who fell for "all it needs is next year's hardware and we'll be fine!"
I would love to believe in VoRec, but I see little evidence that it will exist in a significantly useful form for another ten years or so. Somebody prove me wrong - please.
This is sthe thing I hate most about everything that comes out of the FSF: they just have to throw out any prior art as unworthy because it wasn't developed by the free software community - witness the gnastiness of the perverted "--" options in GNU apps, and the bizarre insistence on using those stupid info pages in place of the well understood and accepted man page standard.
I'm not opposed to the FSF's ideas so much as I'm opposed to it's implementaiton philosophies, and I find that the more I can avoid GNU code, the better life is. Unfortunately, there aren't enough alternatives for some of this stuff, so stupidity wins by default...
Sure, your life as a developer is just dreamy now.
But recognize that what you have done is to shift at least part of that burden onto the shoulders of your customers, who now must ensure they've got the correct versions of all the non-standard (in the sense they're not part of the Sun distro) libraries required to run your product.
As a customer, I'd think very hard before accepting your app in a Sun environment, since your actions just significantly increased *my* TCO! Whether you like it or not, you have to admit that the uniformity of the Sun (or other commercial Unix) environment(s) has its advantages to customers.
I've hardly used the GTK stuff, so what you say comes as a complete shock to me.
Despite all the X-bashing present here, the ability to arbitrarily redirect the display to another place on the network is an incredibly valuable capability, and one that the Windows folks haven't figured out yet.
If you can't even use -display with GTK apps, then they're next to useless in a networked environment!
Also, I'm going to be contrarian and say that I think themability is a really, really, bad idea, and may ultimately be the single largeest contributing factor inhibiting Linux from making large-scale inroads against Windows. Why not do one interface really well, instead of 500 that are ugly, confusing, sick jokes? And I won't even mention the user training issues themeability presents...
Slashdot Mirror
Actually, I'm not making a category mistake (I even noted in my original post that the things I was using as examples were not necessarily interchangeable), but you've made my point for me.
As you note, the real power of distributed/parallel computing comes from the message passing libraries, most commonly MPI or PVM. Beowulf per se is almost nothing more than a label for the generic concept of distributed computing on Linux. The same thing can be done with any other reasonably modern networked computer you have lying around, even those running Windows - you can even mix OSes in a cluster, although this introduces new and interesting problems. (There are a serious lot of underutilzed cycles sitting out there on the corporate world's desktops if they're not running OpenGL screensavers...)
BTW: If the phenomenal success of Sun's E10000 Starfire has taught us anything at all, it's that where I/O is important, a big honkin' SMP box kicks cluster butt! Seriously, the interconnect technology between boxes just *can't* be fast enough to compete effectively with a huge multi-level crossbar packet switch like the ones in the E10K. Sun and the other SMP vendors can win here because they own the domain in which the simpler problem resides...
Don't assume by this that I'm against Beowulf clusters at all - they are a great and amazing thing, but there's more than one way to skin a cat, and Beowulf isn't the only path to Linux distributed computing.
I'll bet you a big bag full of Helium-3 filled buckyballs that your PC or calculator is obsolete before the spiral notebook.
In fact, I'll bet the entire concept of PCs is obsolete before the spiral notebook.
Oh, and even acid-based paper spiral notebooks can be reasonably expected to preserve their information for a century or so with no power and little liklihood of the data format becoming unreadable.
Somehow, that spiral notebook doesn't look so obsolete after all, does it?
NO! If you really know the material, the time limit is never a problem. (See my post above - knowing where/how to find info is not at all the same as knowing it, which implies it is available for immediate recall and use.)
I always loved those exams (too few and far between, admittedly), where I knew the material cold, walked in, aced the test in 20 minutes, checked my answers for another 10 minutes, and walked out confident while those that didn't know and understand the material were sweating it through to the bell.
BTW, one of my better profs used the length of the exam, even in an open book situation, as the mechanism to separate those that knew from those that didn't. If you knew the material, you had time to finish - if you didn't, you soon found you didn't have enough time to look everything up. Excellent tests, excellent testing method, and they produced a class of students that did know the material.
They are nifty inventions, but most of the people ere couldn't build one....
O.K., for all of you saying the calculators/computers are a good thing, try this challenge: Without using your calculator or computer, design (and build, if you feel this is too easy) a fully functioning slide rule. I don't think there are very many college students today with an understanding of logarithms (big hint there) adequate to that task...
This is absolutely true. Without exception, every "really good" engineer, scientist, or other professional I've ever worked with had one thing in common: an understanding of the subject so comprehensive that they are at all times able to take any aspect of the discussion all the way back to first principles.
This is the kind of knowledge that allows one to *easily* derive the formulae they haven't memorized, and it's the kind of knowledge one will never develop using the crutch of a calculator to generate symbolic solutions to calculus problems. (BTW: I speak as someone who used such a crutch, and then had to learn the error of my ways once I saw what real engineers (actual rocket scientists, some of them) were expected to know and do.) Schools that allow students to use such aids are robbing their students of the real education they should be receiving.
Knowing how/where to find the answer is a very poor substitute for knowing the answer. Serious thinking is not possible without that knowing.
The article doesn't say, but despite what you'd think by reading the rantings of the ill-informed 3l337 d00dZ on slashdot, Beowulf isn't even a very good clustering technology for most problems.
There are far more serious, industrial-strength solutions out there, things like MPI, PVM, LINDA,and IBM's own HACMP. (Note these cover a lot of ground and are not necessarily even comparable to one another.)
Beowulf (or any of the others listed above) is not automatically the correct distributed computing methodology. Selecting the proper solution for the job at hand is far more complex than you might imagine. There is a lot more developer activity on some of these than there is on Beowulf - MPI in particular is maturing rapidly and is used for solving big/tough problems in many of the largest companies in the world. (No particular MPI advocacy or bias, it just seems like I run into it more often than the others...)
I'm really surprised at the number of people on /. that are so adamantly opposed to the use of personal electronics while driving. (You sound more like Luddites than geeks...)
Seriously, there was a huge debate early in this century about whether or not *radios* should be allowed in cars because they were so distracting to the driver and would surely lead to hair, teeth, and eyeballs all over the road. After all, who could possibly concentrate on driving while they were listening to music?
A recent issue of American Heritage's Invention and Technology magazine (required reading for geeks, IMHO) carries a story on the numerous technical problems involved in getting radios to work in cars and a reference to the anti-car-radio) forces. There were even laws passed in some locales prohibiting car radios in an effort to prevent the inevitable roadside carnage. Fortunately cooler heads and the eventual availabilty of factory car radios prevailed. (It was the radical idea of a factory car radio that Bill Lear (later the father of the Lear Jet, and the 8-track tape) used to start a company you may have heard of: Motorola!)
In spite of the doomsayers, we've managed to survive reasonably well with not only radios, but also tape and CD players in our cars, and few people consider this to be an undue hazard. Hopefully, this attitude that bashes cellphones and other personal electronics use in cars will die off just as did the bias against radios. In a few years, the idea of not being able to talk on the phone while driving will seem as ludicrous as not being able to listen to your tunes while driving. (And to many of us, it already does...)
Most people have no idea how difficult this is. In fact, this stuff really *is* rocket science!
The moon missions had to hit a date/time-specific moon corridor that was only about 10 miles wide with a velocity tolerance of only about 100 mph. There was one mid-course correction that they could use to finesse position within the corridor, but they had to have hit it in the first place. Think about it for a second, and you'll realize those are *very* tight tolerances, in context.
(BTW: the figures above are from the Ranger missions. Apollo may have used slightly different paths, but the physics is pretty much the same regardless. This is one of the reasons it was such a miracle that we got Apollo 13 back at all. I used to do a lot of work at JSC in Houston, and several of the oldtimers who worked on 13 told me that Mission Control intentionally aimed the capsule along the steepest allowable side of the flight path, so that if something went wrong, it would result in the astronauts being vaporized rather than skipped off into space to asphixiate, which was viewed as being as bad for PR as it would be for the astronauts.)
There are something like 60-odd of them and they're not geostationary so they swoop by on a regular basis. (They were parked in LEO to save money.)
;-)
This is completely bogus. Does *anyone* here bother to understand the facts before posting thier ill-informed opinions.
There is a VERY GOOD technical reason why Iridium satellites are in LEO - and it's quite simple: You want them low to avoid tha time-of-flight problems that plague geostationary satellites. (I used to do geosync satellite protocol tuning - TCP/IP wasn't even capable of dealing with geosyncronous delays until the RFC 1323 enhancements became commonplace. (It took Sun forever to put these in Solaris.) The delays are on the order of thousands of milliseconds - that's right, whole seconds. More bandwidth only hurts the problem. The only things you can do that really help are: 1) shorten the communication path, or, 2) speed up the signal. Mr. Einstein says you can't do the second even if you want to.
It's true that LEO birds are cheaper because they don't need the expensive boosters to geosync (23,000 miles is *way* the heck out there - nearly 3 Earth diameters!), but the chief reason Iridium birds are in LEO is to achieve reasonable signal latencies.
It's a shame Iridium got the moey first and dd it wrong, because the concept is excellent. Iridium just underestimated bandwidth requirements of the network by a couple of orders of magnitude, which put a serious crimp in their pricing model.
Actually, electric cars would be considerably worse for the environment than our current IC engines. Remember, they're not zero-emitting, they're just remote-emitting, and that's after losing most tof the power generated to the hideous inefficiencies of the electric's systems.
I know some of you folks love to bash gasoline engines, and no, I don't think we'll use them forever myself - although I'm betting it's >50 years before we see a dent in thier dominance.
But you have to recognize that technology has improved the performance, efficiency, and cleanliness of gasoline engines by several orders of magnitude, while electrics remain steadfastly immune to technological progress. Remember that gasoline engines weren't given a ghost of a chance at the turn of the century (that would be the 1901 one) - *everyone* knew that steam and/or electric would trounce the stinking, noisy IC motors. But what happened is instructive to all students of technology: the gasoline engine, for all its faults, became the best option - to the point that its competition quit trying. This happened not because of a conspiracy by Detroit and the Oils, but because unlike other powerplants, the ungainly IC engine continues to respond well to technological attention, and seems to show no sign of slowing down its progress anytime soon. At tthe same time, we have only proven that electric cars are as unviable as ever, and that even spending ridiculously high sums on them results in vehicles that are still seriously sub-par compared to their gasoline eqiuvalents. Electric cars are scarcely more viable today than they were in 1920.
Today's gasoline engines are remarkably efficient and clean, with emissions that are effectively zero after the catalyst is warmed up. Today's running emissions would be off the bottom of the scale for cars of just twenty-five years ago. Further, efficency is now remarkable - Honda's new (normally aspirated!) S2000 roadster has a specific output better than many *race cars* of the 1960's!
And don't forget that we don't have excess electricity generating capacity out there burning a hole in our pocket, either. Most utilities are seriously encouraging conservation as a method of delaying the construction of new power plants, which is becoming horrendously expensive due to new environmental rules. Even if the electric car side of the equation were solved, it would likely be a very long time before electrics became economically viable as an alternative, since sharply increased demand would soon cause skyrocketing electric rates. We've gotten very, very good at producing, distributing, and using fossil fuels, and you don't replace that infrastructure overnight.
Oh, and don't forget that there is very little energy lost in transporting fossil fuels, but considerable loss even in the best high-voltage transmission lines we can make...
Electric isn't likely to be the answer anytime soon. (Pure electric, that is - hybrids may be sort of viable sometime soon, although they'll always cost more.)
I am doing and have done several similar "conversion" projects with embedded computers and Linux.
/. before, but in a nutshell:
.1" used by pretty much all other header connectors - damn metric system!)
My first conversion of this type (other than laptops, which I was running Linux on way before it was cool, and when video drivers were a really ugly problem) was an Epson IM-403 cash register computer. (Available from Timeline, the surplus guys, for $99.) It's got a lot less stones than the iOpener, and no screen, but it does have a nifty if tiny little UPS that lets it ride out most power glitches. I've posted about this project here on
- 486SX33 CPU, pretty nice little BIOS
- nice little Chips and Tech video controller will do 800x600 max to std VGA connector
- four serial ports (great for control projects!)
- a parallel port
- socket/tray for a 2.5" IDE notebook hard disk (will accept even the tall 17 or 19mm drives, so you can put IBM's latest monster in it)
- a single ISA half-card slot for the obligatory network card.
- Socket for up to 32MB of RAM (plenty for a decent Linux system - I've run several versions of Caldera on the box, but recommend sticking to slim WMs for obvious reasons.)
- Flat ribbon connector for Epson floppy (know where to order one, if you need it.)
- Tiny little NiCad UPS built in. This is one of the coolest features, and the reason one of these is my primary file server at home. It will only keep the box alive for a few minutes, but that's always been enough so far.
These run Linux fine, or if you just need vNC, you can run the DOS vNC client with the packet drivers or a stripped Linux with the SVGALIB vNC, either of which should fit on a floppy. [As an aside, I have one extra IM-403 with a 3GB HDD that I'll sell for $120+S&H in the US, if anyone's interested.]
(BTW, anybody know where I can get a multiport network card (2 or more Ethers, 1 half-size ISA slot) that works under Linux? If so, I've got my new firewall...)
Also, I'm just starting to hack on a little ARM-based *touchscreen* webphone. These were built at enormous cost by AT&T and Philips before they decided they cost more than the market would bear. (They would have had to sell them for about $600.) They run AT&T's Inferno operating system, and are pretty darn nice speakerphones in addition to finally having the holy grail I've been searching for: a *touchscreen*. I think I can lay my hands on a couple hundred more of these at ~$200 each +S&H, if there's interest. They aren't speedy, but they have a built-in web browser and use one of those little WebTV-style wireless keyboards. Although I don't know of anyone that's replaced the OS (althogh ARM Linux is out there...), but there are hacking instructions available that show how to get root in Inferno. They don't have E-net connections, but do have two PCMCIA slots, so decent networking should be possible.
I've done a lot of hacking with laptop hard disks, and it's extraordianrily difficult to get the 44-pin cables for these things, especially if you want a cable to support two of the little drives. (Which I'd like to do for several machines I've got, including my FIC Sahara Databook, which uses a notebook CDROM on the secondary IDE channel that could give me a place to hang a backup drive. I can't even find anyplace to get a cable like that *made* anywhere here in Austin! You can get the connectors from Jameco, but I haven't found out where to get the dinkier ribon cable these need. (The 44-pin IDE 2.5" drive header uses 2mm pin spacing rather than the standard
Also, a local computer shop once had sockets and little plastic cases to allow a 2.5" HDD to be plugged into a 3.5" drive bay socket. They don't have them anymore. Anybody stumble across these in your searches?
The points in my original submission or this story were edited by the /. staff - most of it was done quite well and added important content, but a couple of things were lost.
/. as a heretic (being in general a supporter of our current patent system) and feel strongly that much of the "reform" wanted by the community here would simply remove any viable protections for small/startup/entrepreneurial companies while leaving the large ones protected by thier inherent economic/political might. For all its warts, the US patent system is one of the most effective systems in history to ensure a level playing field for all, so we should tamper with it only with extreme caution and certainty.)
/., and does, in fact, matter.
I made the point that the CNET article did not explicitly reference design patents, but I expected that Apple and others had them and could be expected to enforce them, sometimes to the detriment of consumers. The point was to open a more general discussion of alternative mechanisms which could be used to render technologies proprietary. (Remember the big flap a couple of years ago - Motorola sued Qualcomm for building a phone that folded, saying it violated their design patents on the StarTAC. Fortunately, the courts ruled that MOT couldn't own the generic idea of a folding phone (since everyone knew, and their product name even alluded to, the fact that they stole the idea from Star Trek.) Still, the MOT attack delayed QCOM's Q-phone until it was no longer a serious competitive threat. Not that that's hurt their stock price any...[grin])
You're right that the CNET article doesn't have anything to do with patents, but I was intentionally trying to raise the issue of some of these other mechanisms by which we can expect companies large and small to attempt to excercise control or block competitors. As I was quoted in Timothy's post, I'm particularly concerned about potential abuse of design patents. (BTW, I am on record here on
Also, my article submission indicated that I thought this particular decision was well-justified, since there was a clear attempt to copy the iMac, even though I don't think for a minute that anyone would confuse these Windows boxes for real Macintoshes. Still, even if a good call in this case, this could be a slippery slope since it is necessarily subjective.
Finally, if this sort of limitation becomes viewed as effective, I raised the spectre of ID squatters protecting everything they can think of this way and then selling all the good and necessary ones off to the highest bidder. (I considered doing this myself a number of years ago, since PDA's still don't have nearly the functionality I described in writing nearly a decade ago. If I'm not going to get venture money myself, then I could at least sue someone who did - but that's really cheezy now isn't it? And yes, in that context, a "z" in "cheezy" seems just right...)
The point of discussion I was trying to provoke is, "Are there any effective safegaurds against abuse of other parts of our IP system that may not have recieved as much scrutiny as patents for software?"
In this context, in contrast to what some other posters seem to think, this topic is appropriate for
Oh, and for the record, I think most of these clear and whatever things are kind of ugly, although the ones that are plain clear (no other colors, like the Palm IIIe) are interesting. And as an aside, it's a shame Apple didn't learn from DEC's experience with puck mice - they used them on their VAXstations and DECstations at least up until the Alpha came out, and they also were a pain to grab and use without looking.
I'm sure he already has. It's funny, but a liberal arts degree, because it's based on more classical content, is one of the most conservative educations possible.
Remember, all conservatives are "classical liberals", so you need to keep time context in mind and avoid knee-jerk use of or reaction to the term "liberal"...
I'd second the fact that digital interfaces are important. Most of the problems I've encountered with flat screens and regular analog (*VGA) connections center around the difficulty inherent in redigitizing the analog signal that the RAMDAC on your video card just thoughtlessly created, and mapping that *cleanly* and solidly into the panel's native pixel grid.
I've seen some very expensive LCD screns that look awful because they aren't able to perfectly re-gen the digital signal. Most analog-input flat-panel monitors can be made to work well with a singel computer, but on the rare occasions I've tried to use a vid/kbd/mouse switch box, I've found that it's pretty much impossible to come up with a setting which will look good with the input from differing video cards. (It might be better if they were identical, or it might not...)
The problem is that there's not a lot of support for the digital interfaces yet even on the hardware side, and you may (although I haven't checked) have problems finding video drivers for XFree86 or any other non-MS display system.
This seems to be kind of a chickenand-egg problem - no one can buy digital interface flat screens until someone builds them, and no one will build them (and price them sanely) until there's a large proven market.
I've gone through this set of gyrations myself, to the point that I may just give up and buy a Gateway Profile just to avoid the hassle.
I wish there were more, better, and cheaper choices for flat screens. It galls me that it's nearly as cheap to buy an all-in-one like the Profile, or one of the new "suitcase-luggable" laptops like Dell's with the big screens than to add a good digital flatscreen and video card to my existing PC. Something's very wrong here...
Actually, radio *was* invented here. This was settled legally in a sensational lawsuit (which was appealed IIRC to or near the Supreme Court) between Nicola Tesla (who really invented it first), and Gulielmo Marconi (who independently invented it a short time later and proceeded to devlop it much faster than Tesla.)
Although Tesla was not an American by birth, he got here as fast as he could!
This is non-news if I ever heard it. CPU cores and even complete CPU designs have been free or near-free for years.
Sun started giving away SPARC uP designs last year - before that they were charging a stunning $99 for them.
Oh, there is the catch that in order to actually build the chips, you'll need a modern stae-of-the-art semiconductor fab facility, which should run you about US$3e9. Or you could spring for a few tens or hundreds of millions and buy a slice of some else's fab, if they have excess capacity. (I don't expect there's much of that around right now...)
CPU designs have been efectively free for a long time now, and that hasn't changed much. Ross was the last company to seriously try to capitalize on this idea and look where it got them!
Several posters have commented on the watery stuff that passes for beer among much of the American populace.
The following link explains how American beer got so wimpy - it was the feminists, who deviously eliminated real man's beer:
http://www.credenda.org/issues/vol1 1/reci11-2.htm
A discussion of computer animation or even portrayals of future tech would have been appropriate, news of casting decisions is not. I stand by my original gripe.
As I've mentioned here before, I've got a copy of an old Interface Age magazine in my garage that has in it a "floppy ROM" - one of those cheesy flexible LPs they used to put in magazines and books that contains a BASIC ROM for an old Z-80 or 8080 computer.
Although I never used it, I know people that did. (I didn't have a computer then, I just read about them and dreamed about the cool stuff from Imsai, Altair, The Digital Group (the coolest looking computers), Compucolor (wow, color graphics!), and SW Technical Products (possibly history's ugliest terminal.))
Back then there was something called the "Kansas City Standard" which was a big deal at the time because it brought interoperability to audio cassette tape data storage. The folks who did Flexi-LPs like these just pressed the KC Std. audio onto the disc, then all you needed was the correct adapter to plug the audio output from the turntable (well, a pre-amp, actually, since you didn't get line-level output from most turntables) to the KC interface box that you would normally plug into the cassette player.
It wasn't fast, but it was pretty robust: there was enough slop in the spec that you could usually read copies of copies of copies of tapes. (We're not talking quality copies here, folks, just hooking together whatever two cassette decks you could find.)
Wow, I sound a lot older than 37, don't I?
Several points to clear up misconceptions:
1. Bluetooth is not meant to be a wireless LAN replacement (not without a gazillion "access points", anyway.) It's intended to be a cordless "desktop area network", and RF bubble that encircles Bluetooth devices allowing them to form ad hoc connections wiht one another for both data and isochronous (e.g., voice) connections. One node you connect to may optionally act as a gateway (access point) to the "real" network, if it's connected and wants to offer that service. As a transitory, ad hoc, solution, 10m is just about the most range you'd want - any more would create problems, and personally, I think they made the bubble too big. Bluetooth isn't meant for seamless raoming wireless connections, but to facilitate connections to and between things like telephone sets (wired or wireless), PDAs, desktop computers, etc.
2. 2.4 GHz doesn't belong to either the Bluetooth or the 802.11 guys. In the US, 2.4 is one of the ISM (Industrial, Scientific, and Medical) bands (others are at 900 MHz and 5.7 GHz.) These are for unlicensed use providing that the FCC's restrictions are met. These restrictions include accepting any interference without recourse - if you want recourse, you go for licensed spectrum! 2.4 GHz was chosen because with minor adjustments, it's available for unlicensed use in pretty much every country that matters. (And it's the only reasonably low frequency for which this is true...)
BTW - there are lots of other perfectly legal 2.4 GHz devices which can kill your wireless LAN, so don't single out Bluetooth unfairly...
I know there's been a history of abuse w.r.t. this topic by the /. owners, but Star Wars rumors are neither "News for Nerds" nor "Stuff that matters".
/. folks prefer to post whatever articles will result in hits and generate endless drivel-filled opinion posts than actually facilitate tech discussions with any real content.
Save the Star Wars rumors for E!online, or maybe nationalenquirer.com. Apparently, now that the site is fully commercialized, the
I'm starting to seriously question whether Slashdot is worth my time anymore. Tragedy of the commons is taking its grisly toll and the S/N ratio continues to plummet. Does anyone care?
I would love to use voice recognition in an appropriate capacity, but it's not going to happen soon, if ever.
(Note also that using VoRec correctly requires a near total redesign of user interfaces to accomodate correlation of recognized voice commands and gestural input - notice how often you point or make another meaningful motion when you talk to someone, especially in trying to convey information, and that those gestures mean very different things from the way interfaces use pointing today.)
It's really sad, but today's best voice recognition is only marginally more useful than the CoVox voice recognition I had on my Commodore 64 thirteen years ago. Sure, vocabularies are larger now, and we can get somewhat closer to true connected speech recognition, but by and large, the lack of progress proves that despite an improvement of several orders of magnitude in processing power, (and memory speed and capacity, and storage space, not to mention DSPs) voice recognition is not a problem that can be effectively addressed by any amount of brute force processing using current methods.
I'm going to really date myself here: I've got a copy of Interface Age magazine (anybody else remember that?) in the garage (the one with the floppy ROM: a BASIC ROM on one of those floppy vinyl records they used to put in magazines) that has ads for several voice recognition systems that didn't work all that much worse that the ones we have now - I had friends that had some, back when everybody was using the S-100 bus. Speech recognition has been "just around the corner" in the computer/hacker community for as long as I can remember, and the VoRec prognosticators have been hoovering money for years out of suckers of either the investor or customer ilk who fell for "all it needs is next year's hardware and we'll be fine!"
I would love to believe in VoRec, but I see little evidence that it will exist in a significantly useful form for another ten years or so. Somebody prove me wrong - please.
This is sthe thing I hate most about everything that comes out of the FSF: they just have to throw out any prior art as unworthy because it wasn't developed by the free software community - witness the gnastiness of the perverted "--" options in GNU apps, and the bizarre insistence on using those stupid info pages in place of the well understood and accepted man page standard.
I'm not opposed to the FSF's ideas so much as I'm opposed to it's implementaiton philosophies, and I find that the more I can avoid GNU code, the better life is. Unfortunately, there aren't enough alternatives for some of this stuff, so stupidity wins by default...
Sure, your life as a developer is just dreamy now.
But recognize that what you have done is to shift at least part of that burden onto the shoulders of your customers, who now must ensure they've got the correct versions of all the non-standard (in the sense they're not part of the Sun distro) libraries required to run your product.
As a customer, I'd think very hard before accepting your app in a Sun environment, since your actions just significantly increased *my* TCO! Whether you like it or not, you have to admit that the uniformity of the Sun (or other commercial Unix) environment(s) has its advantages to customers.
I've hardly used the GTK stuff, so what you say comes as a complete shock to me.
Despite all the X-bashing present here, the ability to arbitrarily redirect the display to another place on the network is an incredibly valuable capability, and one that the Windows folks haven't figured out yet.
If you can't even use -display with GTK apps, then they're next to useless in a networked environment!
Also, I'm going to be contrarian and say that I think themability is a really, really, bad idea, and may ultimately be the single largeest contributing factor inhibiting Linux from making large-scale inroads against Windows. Why not do one interface really well, instead of 500 that are ugly, confusing, sick jokes? And I won't even mention the user training issues themeability presents...