First of all, Storrs-Hall just described a "mass driver", but one propped up on a 100km high tower for God knows what reason. Which proves:
He doesn't know anything about orbital mechanics, but just thinks you need to "point up to get into space". Well, that's probably unfair; he at least got orbital velocity about right.
He certainly doesn't know anything about mechanical engineering, or he'd realize that making a 100km structure that will support tension may be possible, but making one that won't buckle under compression is ridiculous. Buckytubes may have a hundred times the tensile strength of steel, but they still bend real easy...
You've got it all backwards
on
Going Up?
·
· Score: 4
Do you realize how much energy our Sun is spitting out every second? And how little of that energy happens to impinge on this tiny planet for us to use?
How about matter? I'll spare you the numbers, but it wouldn't take a very large asteroid to supply our entire civilization's structural metal requirements for centuries, and to provide enough mass of everything else to make the phrase "precious metal" an oxymoron.
There's a nice Kuro5hin discussion going on right now about overpopulation, including the question of what is a "sustainable" population for humanity. The answer isn't encouraging; our fossil fuels won't be around in a few centuries, our fissionable metals will give us a few centuries more... and then what? Solar power? Not concentrated enough, unless you've got a plan to reduce our population 10-fold, or pull in extra power from space. Fusion? That's better (assuming we get it working eventually), but then you run into the problem that the cleanest fusion fuel, He3, only exists in quantity on the Moon and outer planets. Even if you don't see the value of going into space to support life there, eventually we'll want to leave this planet to better support life here.
The solar system has the resources to support quintillions of people; unfortunately for us an insignificant fraction of those resources happen to be on Earth.
An elevator to nowhere. Imagine how silly it'd look.
So anyway, like I said, you've got it exactly backwards. It would be an elevator to everywhere.
It is only illegal to * create or distribute* such software.
It's also illegal to "import" (and there goes our offshore developer defense) or "otherwise traffic in" such software. Now maybe "downloading an (illegally distributed) copy for personal use" doesn't fit that final catch-all category, but do you want to be one of the defendants in the test case?
I've had a Geforce DDR since this summer, with mixed results.
With the default 2.2 kernel, XFree86 4.0 setup I had at the time, following instructions I got the NVidia drivers compiled and set up pretty easily. Hell, if you're running an untouched Red Hat (or some others) without even recompiling the kernel, then you can pretty much download their RPM drivers, change a line in XF86Config, and restart X.
Sure, there's something that made me uneasy with using binary-only drivers and loading a half-megabyte kernel module, but oh well. Oh, and did I mention the annoying bugs? There's the "X hosed if a 3D app gets killed suddenly bug", and the "horrible graphics glitches when switching between multiple X servers" bug, and the "if anything goes wrong, your text consoles will be permanently black until you reboot, even if restarting X works fine" bug, and well, you get the picture. Some of their problems got a little better in the 5 revisions of their Linux drivers I've seen. Some didn't. Will the rest of them get fixed, including the "we'll have this fixed in a few days messages" on their months old Linux FAQ? I don't know, but I don't like that the answer depends solely on the NVidia Marketing Department's appraisal of my value as a future customer.
But anyway, X was pretty stable, Quake III was damn fast, and life was good.
Then I decided to try 2.4, because of better support for my motherboard's onboard sound among other reasons. The Nvidia kernel module wouldn't compile against test7, test8,... I uninstalled all the binary stuff, switched to the open source 2D drivers, and forsake 3D support outside of Windows for months. Fortunately, the beneficial effect this had on my GPA prevented the loss from angering me greatly.
Finally, someone came out with a patch to compile the kernel module against 2.4.0-test11, and stuff ran great again with the NVidia drivers. Same niggling bugs, but 3D was golden.
Then came 2.4.0-test12. Now NVidia's drivers compile but don't work, so I'm back to the open source stuff. Of course, when I mean don't work, I mean *really* don't work: If I run X with the closed nvidia driver, it fails and thrashes the screen until the next (blind or remotely logged in) reboot. If I run X with the open nv driver, it works, but I don't get 3D, and if I were to foolishly run an app linked to the binary NVidia GLX drivers (like 3D Xscreensaver hacks, before I turned them off), it crashes the kernel.
That's right. Do you want blazing fast 3D support, or do you want months of uptime? NVidia's drivers remind me of software I had to fix at work last summer: use them in exactly the situations that the original coders tested with, and you're gold. Try something that should be within the design spec, but is still out of the ordinary, and down it goes. You can debug code by understanding what can go wrong and preventing it, or you can debug it by running it, fixing whatever leads to a crash/error, and repeating it until you stop seeing crashes. nvidia_drv.o smells like a product of the latter method.
I've had some problems (including crashes that resemble the black screen Linux crashes, but with no known cause) with the card under Windows too, but not too much more than the problems I have with everything under Windows. I had an epiphany the other day, realizing that if just one binary-only third party driver under Linux can suck like this, what happens in Windows where *every* driver is code that neither I nor Microsoft gets to see or debug? I almost pity the Windows OS developers now; they probably get blamed for three times the crashes they're actually responsible for.
Anyway: if you want blazing fast 3D, there's really no choice but NVidia. Their hardware is a generation ahead of ATI and 3dfx, and two ahead of everyone else out there. If you want a card that's excellent at everything else, and are willing to trade excellent Linux drivers for mediocre 3D acceleration, go with Matrox.
The guy who actually wrote this, Dean Pannell (who goes by "dinotrac", not "Maldivian") does say "Copying with proper attribution is expressly permitted" on his essay... but that copyright statement and attribution seems to be the only thing that wasn't cut and pasted here.
Please, people; some of us want legal Linux DVD players even more than we want DVD ripper software. Can you at least try to look like you care about copyright law?
But it's a repeated prisoner's dilemma game, every moment, as long as you both are working. In other words, if you decide to start goofing off, and your partner decides to work, he gets to change his mind as soon as he sees you slacking.
Sure, when you play PD once, there's no reason not to defect. If you play a thousand times against the same player, on the other hand, defecting even once is dangerous.
Of course, it breaks down a little... in PD, if you both defect, you both lose. In real life, if you both goof off, but not so long that you delay the project, you feel like you've both won.
At the time I bought my K6-II, it was as cheap as (lower clock speed) Celerons with similar FPU performance, and it blew away anything in it's price class for integer performance. Of course, Celerons got cheaper, and gamers stopped buying K6II chips, but they were always good for business apps. If you didn't need the highest Quake II framerate, and didn't want to upgrade the system later, you were set.
"Bad", of course, comes in when you try and stick a real AGP card in that AGP-shaped slot on most Super7 motherboards. Turns out they provided enough power through that slot for my Millenium II, but not nearly enough for even a Voodoo 3, let alone the GeForce I bought last summer. Feeling the undersized power transistors become excruciatingly hot was not encouraging. Having the computer crash, in Windows and Linux, after a few minutes use, was less encouraging. Having to upgrade my motherboard, processor, and case months before I'd intended to, to use a video card that was supposed to work in the original system, sucked.
I went to a Duron after checking out compatibility issues, BTW.
The original Athlon itself got a mostly-undeserved bad rep for this reason; there were a some early motherboards that had the same half-assed AGP compatibility problems.
1. Starting an uncontrolled nuclear chain reaction requires lower tolerances than controlling a chain reaction that is just subcritical. This means that fusion bombs are technologically simpler than nuclear rockets or fusion power plants.
2. Elliptical orbits with periapsis belows a planet's surface are lower energy than circular orbits at the ellipse apogee. This means that ICBMs are technologically simpler than orbital rockets.
...for posting related stories together so readers can get the whole picture. The "Will Linux Save Microsoft?" article seems inexplicable until you read the preceding "Crack for Sale" headline, and realize what Hal Plotkin was smoking.
Here's a clue: resize/.'s window - what do you notice?
I notice that at 640x480, that row of section category icons at the top of each page causes both Mozilla and IE5 to require a horizontal scrollbar to render the page correctly.
You see, while Slashdot avoids all the fixed table size travesties which make half the web look like a thin column (subtly phallic - does it imply that the web developer was a dick?) of text at my preferred high browsing resolution, it does succumb to the temptation of those fancy "graphics" which are all the rage among the kiddies.
And even though this example could be coded around (it wouldn't be hard to make those icons wrap when the available space is too small), there are others that won't. I had a graphical button bar on the top of an old website of mine, for instance, and while it only took a little work to make it happy with 640x480, trying to view it on a palm pilot (except with a text-only browser) would be disasterous.
The future solution is not to code "The One True HTML Subset", by the way. If you do that, then you avoid having 5% of users think your page sucks, at the expense of the 95% who now think it's mediocre. The ideal solution is to provide a second stylesheet to control the layout of your pages for the smallest web browsers. Of course, this requires you to get the content/presentation separation right from the beginning, so it won't be the popular near future solution...
Are they serious? That has such a Austin Powers, "hilariously frozen in time and don't know what's going on in the rest of the world" sound to it. I can just picture Gates in his silver Dr. Evil suit, explaining to his assistants the potential of the (making quote marks in air with fingers) "commands line" and the "laser".
Kiss The Blade's hilarious MO is to make straight-faced "devil's advocate" posts with just enough misinformation and logical fallacy to send the average intelligent-but-gullible slashdot user into a blindly raging counterattack. Normally, the only reason I don't mod him (-1, troll) is that unlike the hot grits morons he's so good at it. 23 replies, 6 replies, 39 replies, 11 replies... he picks up everything from flaming rants to philosophical theses. That earns respect.
But damnit, KTB, what are you doing now???
This sets a bad precedent. The DoD is acting to prevent totally unjustified anxiety on the part of millions of Americans. I thought the function of Government Departments was to do what is right, not what people think is right
I totally agree with this. What is going on? Did you get bored and decide to write something honest instead of baiting people? Did you realize that many of us were on to you weeks ago, and decide to throw in a monkey wrench by posting something reasonable? Is your karma dangerously close to the default score: -1 threshold?
I don't detect any hyperbole, feigned ignorance, or baiting here. And what scares me is the thought that maybe KTB hasn't changed, maybe I'm just in the select group of misinformed idiots being parodied by one of his posts, and don't realize it...
...it also cost 10 times what missions in the current Mars exploration program do. And not all the current missions have been failures. All we need is 1/10 to succeed and we're still coming out even.
When you boot Linux, if the root partition is marked as FAT32 in the partition table, but LILO and/etc/fstab know it to be an ext2 partition, then Linux doesn't care what the partition table says. Does BSD?
It was really using 100-200 MB
on
Nvidia's NV20
·
· Score: 2
The reason I noticed at all is that my machine was swapping like mad, even with 128MB of physical RAM, and even after turning the usual memory hog culprit (Netscape) off.
So do their Windows drivers suck, too?
on
Nvidia's NV20
·
· Score: 2
NVidia, on the other hand, uses the same codebase for both their Windows and Linux drivers.
Even for the enormous Linux kernel module that's required to use their drivers? Really?
Does their Windows driver, after less than a week of use, bloat to consume over 200MB of virtual memory? That's what their closed source XFree86 driver did with my GeForce DDR, on XFree86 4.0.1 and kernel 2.4.0-test9, even without using the 3D features at all. The open source nv.o driver that came with XFree86 isn't exactly a spartan RAM user either, but at least after it's sucked up a big chunk it stops asking for more.
Granted, they don't seem to care about keeping up with development kernels (their kernel module didn't even compile against 2.4-test for a while); I haven't exactly put much work into fixing the problem (but how can I, when I can't even recompile with debugging symbols?); and their drivers did seem to work OK with kernel 2.2.16.
Nevertheless, I don't intend to buy another NVidia card until I have an open source 3D driver to run it with. By contrast, my previous 3D acceleration in Linux came from Mesa on top of Voodoo2 glide; the frame rate may not have been as fast, but the rate of driver improvment certainly was faster.
Is this really an advantage of IRV? I'd think that any system which allows voters to do a complete preference ranking of candidates would allow multiple winners to be selected by one vote: Select first place winner; remove that candidate from everyone's vote; repeat until all seats are filled.
What you incorrectly call "lying", most people call setting priorities.
No, a ballot that said "I prefer Browne to Nader to Gore to Bush" would be setting priorities (plural). A ballot that says "I prefer Gore to everyone else" sets a priority (singular), and millions of voters regularly set that priority in a way that makes it an inaccurate statement, a "lie".
Voters in the U.S. here have more power because they can decide they dislike a candidate so much that they'll vote for a stronger candidate who would otherwise be their second or third choice just to be able to knock the guy they dislke off.
Did you read about approval voting? Instant runoff voting? Those systems let you express an "I don't want this person to be president" opinion just as strongly; they just don't force you to sacrifice your other opinions in a multicandidate race to do it.
This leads to candidates who not only worry about energizing their base, but also worry about being considerate enough of their opposition to not unduly piss them off.
No, it leads to candidates who cannot afford be considerate of their opposition, lest their least moderate "base" voters splinter off to a more extreme candidate and fracture their party. It leads to primary elections which drop candidates like McCain who have broad bipartisan appeal but less appeal among the party faithful.
On the other hand in most EU democracies,
In most EU demt EU democracies, don't they have proportional representation to ensure that the legislature is just as split as the voters? That has nothing to do with any of the voting methods being discussed; we can dump the plurality system without getting rid of the "one district, one representative" House.
This leads to extremist parties (willing to switch votes on national concerns) in a position to make or break governments
...and completely negate the idea of a secret ballot. great.
By "ID", I mean the same ID you get on ballots in many places today: a unique number, on both the ballot and a tear-off stub, which could (in theory) be used for you to verify that your vote was counted correctly.
Imagine how much harder election fraud would be, if, instead of each district reporting "16,835 people voted for Gore", each district put a text file online saying:
PLMBCH000001 went to Gore.
PLMBCH000002 went to Bush.
PLMBCH000003 went to Buchanan.
I just finished reading your recent article, "May the Best Man Lose".
The author unfortunately underestimates one of the greatest weaknesses of the plurality system, and so fails to realize that this weakness applies nearly as strongly to the Borda count: both voting systems encourage voters to lie!
Of course, the media today doesn't call strategic voting "lying", they call it "not wasting your vote". It is considered standard practice to give your vote not to the candidate you prefer, but to the poll-leading candidate you dislike least.
This practice would not change under a Borda count system. Voters who prefer Nader to Browne to Gore to Bush will still be encouraged to vote for Gore above either Nader or Browne, because that way they will add 3 points to the separation between the leading candidates rather than 1 point, and their vote will have almost three times the impact in the election.
Strategic voting makes independent and third-party candidates nearly irrelevant, and gives the Democrats and Republicans a chokehold on politics. That bipartisan chokehold, by the way, is why we may very well soon see a constitutional amendment to eliminate the electoral college, but we will never see a superior system of voting replace the plurality system. The electoral college is much less damaging to our democracy (and had less of a detrimental effect in this presidential election) than the plurality system, but removing the electoral college will result in only a minor power shift from the Democrats towards the Republicans; changing the plurality system would result in a major power shift away from both.
I was surprised to see that one important voting system was not even mentioned in your article: Instant Runoff. In that system, voters rank their candidates by preference. If no candidate has a majority of first place votes, the candidate with the fewest first place votes is removed from consideration, and from the rankings of voters who voted for him. (i.e. if Alice's second place choice was dropped, then her third place choice becomes her new second, her fourth becomes her new third, etc.) Once a candidate has a majority of first place votes, that candidate wins. The results are not guaranteed to equal Borda count results, but they often will. I don't think strategic voting is impossible, but it's a lot harder.
Don't dismiss the possibility of computer error and fraud so lightly. This year, the Republican National Committee's webpage was defacedVolusia County is doing a manual recount now, after computer error injected 15,000 erroneous Socialist Workers and Constitutionalist Party votes into their first tally. Keep in mind: the people who will be administering our first computerized elections will be these people, not Linus and Alan.
Sure, we can have physical "backstops" to try and prevent these kinds of problems... but if these backstops have to be resorted to every time an election comes this close, how is it any improvement over our current situation? If a software glitch in your voting machine causes every voter whose last name begins with Z to get 10 votes, will that same glitch from causing your voting machine to print out 10 punchcards to backstop those votes?
Don't get me wrong, there's certainly a lot of room for technological improvement here. Some Slashdotter suggested a touchscreen voting machine which would give you a clearer GUI, prevent or check for any invalid double votes, and print out a sheet with your ID and only the names of the people you voted for, in easily machine-scannable form, so you could take that sheet and give it to the poll workers. That would prevent both the ballot and the discarded vote problems in Palm Beach, at least.
It's tempting to think that there could be something even better. A little smartcard (because I don't trust the nation who let Melissa and ILOVEYOU loose to maintain their PCs securely) with public key crypto could let your "vote" be a digitally signed statement that you could safely send over the net, and the collection of all those signed statements could be publically downloadable, to allow you to check and make sure nobody tampers with your vote or the vote count. But even in that case, who would we trust to distribute the private keys, and never have their systems compromised? Verisign? Even if you can check your own vote's integrity, how do you know that a 6 million vote list isn't actually 5 million real votes, plus 1 million fraudulent inserts?
Oh, yeah, I forgot; I titled this post "Optimizing Election Fraud" for a reason. Consider: Right now, tampering with n votes is an O(n) operation. A well designed computer system could make that an O(1) operation. In most programs this would be a fantastic optimization; in this particular case it is not an improvement.
I don't mean to sound like such a conspiracy theorist; I'm sure it was really just a computer glitch. But seriously, the fact that there were about 15,000 changing votes in a 300-vote difference race, and that the national media was almost completely ignoring this... well, it was quite disturbing. Add to this the fact that nobody on Slashdot seemed to care... and it was like dropping into the movie "Brazil..." Am I going nuts? Is everybody else already there?
Good to hear that the county's returns will be straightened out, and that the Democrats and Republicans are at least paying attention.
Harris went from 9,888 votes to 8 votes.
Philips went from 2,927 votes to 20 votes.
Browne went from 3,211 votes to 442 votes.
Nader went from 2,436 votes to 2,903 votes.
Buchanan went from 396 votes to 496 votes.
Hagelin went from 33 votes to 36 votes.
McReynolds went from 3 votes to 5 votes.
Moorehead went from 59 votes to 69 votes.
George W. Bush and Al Gore had returns that were completely unchanged.
That's right. The recount from Volusia claims in effect, "We screwed up the vote on every other candidate, reported 16,000 erroneous votes in total... and yet managed to get the count for the major candidates exactly right!"
I hope someone's keeping track of pre-recount and post-recount totals by county in Florida. CNN.com has only pre-recount data up by county, and doesn't list Harris in it's table. FOX has both pre and post recount data by county, but only for Gore and Bush. That Daytona Beach news website basically overwrote their prerecount returns page with the postrecount returns. As soon as CBS news updates it's page, I'll be out of links to records of the initial 10,000 Harris votes.
Slashdot Mirror
I don't know where to begin with this one.
First of all, Storrs-Hall just described a "mass driver", but one propped up on a 100km high tower for God knows what reason. Which proves:
He doesn't know anything about orbital mechanics, but just thinks you need to "point up to get into space". Well, that's probably unfair; he at least got orbital velocity about right.
He certainly doesn't know anything about mechanical engineering, or he'd realize that making a 100km structure that will support tension may be possible, but making one that won't buckle under compression is ridiculous. Buckytubes may have a hundred times the tensile strength of steel, but they still bend real easy...
Do you realize how much energy our Sun is spitting out every second? And how little of that energy happens to impinge on this tiny planet for us to use?
How about matter? I'll spare you the numbers, but it wouldn't take a very large asteroid to supply our entire civilization's structural metal requirements for centuries, and to provide enough mass of everything else to make the phrase "precious metal" an oxymoron.
There's a nice Kuro5hin discussion going on right now about overpopulation, including the question of what is a "sustainable" population for humanity. The answer isn't encouraging; our fossil fuels won't be around in a few centuries, our fissionable metals will give us a few centuries more... and then what? Solar power? Not concentrated enough, unless you've got a plan to reduce our population 10-fold, or pull in extra power from space. Fusion? That's better (assuming we get it working eventually), but then you run into the problem that the cleanest fusion fuel, He3, only exists in quantity on the Moon and outer planets. Even if you don't see the value of going into space to support life there, eventually we'll want to leave this planet to better support life here.
The solar system has the resources to support quintillions of people; unfortunately for us an insignificant fraction of those resources happen to be on Earth.
An elevator to nowhere. Imagine how silly it'd look.
So anyway, like I said, you've got it exactly backwards. It would be an elevator to everywhere.
It is only illegal to * create or distribute* such software.
It's also illegal to "import" (and there goes our offshore developer defense) or "otherwise traffic in" such software. Now maybe "downloading an (illegally distributed) copy for personal use" doesn't fit that final catch-all category, but do you want to be one of the defendants in the test case?
I've had a Geforce DDR since this summer, with mixed results.
With the default 2.2 kernel, XFree86 4.0 setup I had at the time, following instructions I got the NVidia drivers compiled and set up pretty easily. Hell, if you're running an untouched Red Hat (or some others) without even recompiling the kernel, then you can pretty much download their RPM drivers, change a line in XF86Config, and restart X.
Sure, there's something that made me uneasy with using binary-only drivers and loading a half-megabyte kernel module, but oh well. Oh, and did I mention the annoying bugs? There's the "X hosed if a 3D app gets killed suddenly bug", and the "horrible graphics glitches when switching between multiple X servers" bug, and the "if anything goes wrong, your text consoles will be permanently black until you reboot, even if restarting X works fine" bug, and well, you get the picture. Some of their problems got a little better in the 5 revisions of their Linux drivers I've seen. Some didn't. Will the rest of them get fixed, including the "we'll have this fixed in a few days messages" on their months old Linux FAQ? I don't know, but I don't like that the answer depends solely on the NVidia Marketing Department's appraisal of my value as a future customer.
But anyway, X was pretty stable, Quake III was damn fast, and life was good.
Then I decided to try 2.4, because of better support for my motherboard's onboard sound among other reasons. The Nvidia kernel module wouldn't compile against test7, test8,... I uninstalled all the binary stuff, switched to the open source 2D drivers, and forsake 3D support outside of Windows for months. Fortunately, the beneficial effect this had on my GPA prevented the loss from angering me greatly.
Finally, someone came out with a patch to compile the kernel module against 2.4.0-test11, and stuff ran great again with the NVidia drivers. Same niggling bugs, but 3D was golden.
Then came 2.4.0-test12. Now NVidia's drivers compile but don't work, so I'm back to the open source stuff. Of course, when I mean don't work, I mean *really* don't work: If I run X with the closed nvidia driver, it fails and thrashes the screen until the next (blind or remotely logged in) reboot. If I run X with the open nv driver, it works, but I don't get 3D, and if I were to foolishly run an app linked to the binary NVidia GLX drivers (like 3D Xscreensaver hacks, before I turned them off), it crashes the kernel.
That's right. Do you want blazing fast 3D support, or do you want months of uptime? NVidia's drivers remind me of software I had to fix at work last summer: use them in exactly the situations that the original coders tested with, and you're gold. Try something that should be within the design spec, but is still out of the ordinary, and down it goes. You can debug code by understanding what can go wrong and preventing it, or you can debug it by running it, fixing whatever leads to a crash/error, and repeating it until you stop seeing crashes. nvidia_drv.o smells like a product of the latter method.
I've had some problems (including crashes that resemble the black screen Linux crashes, but with no known cause) with the card under Windows too, but not too much more than the problems I have with everything under Windows. I had an epiphany the other day, realizing that if just one binary-only third party driver under Linux can suck like this, what happens in Windows where *every* driver is code that neither I nor Microsoft gets to see or debug? I almost pity the Windows OS developers now; they probably get blamed for three times the crashes they're actually responsible for.
Anyway: if you want blazing fast 3D, there's really no choice but NVidia. Their hardware is a generation ahead of ATI and 3dfx, and two ahead of everyone else out there. If you want a card that's excellent at everything else, and are willing to trade excellent Linux drivers for mediocre 3D acceleration, go with Matrox.
The guy who actually wrote this, Dean Pannell (who goes by "dinotrac", not "Maldivian") does say "Copying with proper attribution is expressly permitted" on his essay... but that copyright statement and attribution seems to be the only thing that wasn't cut and pasted here.
Please, people; some of us want legal Linux DVD players even more than we want DVD ripper software. Can you at least try to look like you care about copyright law?
But it's a repeated prisoner's dilemma game, every moment, as long as you both are working. In other words, if you decide to start goofing off, and your partner decides to work, he gets to change his mind as soon as he sees you slacking.
Sure, when you play PD once, there's no reason not to defect. If you play a thousand times against the same player, on the other hand, defecting even once is dangerous.
Of course, it breaks down a little... in PD, if you both defect, you both lose. In real life, if you both goof off, but not so long that you delay the project, you feel like you've both won.
At the time I bought my K6-II, it was as cheap as (lower clock speed) Celerons with similar FPU performance, and it blew away anything in it's price class for integer performance. Of course, Celerons got cheaper, and gamers stopped buying K6II chips, but they were always good for business apps. If you didn't need the highest Quake II framerate, and didn't want to upgrade the system later, you were set.
"Bad", of course, comes in when you try and stick a real AGP card in that AGP-shaped slot on most Super7 motherboards. Turns out they provided enough power through that slot for my Millenium II, but not nearly enough for even a Voodoo 3, let alone the GeForce I bought last summer. Feeling the undersized power transistors become excruciatingly hot was not encouraging. Having the computer crash, in Windows and Linux, after a few minutes use, was less encouraging. Having to upgrade my motherboard, processor, and case months before I'd intended to, to use a video card that was supposed to work in the original system, sucked.
I went to a Duron after checking out compatibility issues, BTW.
The original Athlon itself got a mostly-undeserved bad rep for this reason; there were a some early motherboards that had the same half-assed AGP compatibility problems.
1. Starting an uncontrolled nuclear chain reaction requires lower tolerances than controlling a chain reaction that is just subcritical. This means that fusion bombs are technologically simpler than nuclear rockets or fusion power plants.
2. Elliptical orbits with periapsis belows a planet's surface are lower energy than circular orbits at the ellipse apogee. This means that ICBMs are technologically simpler than orbital rockets.
...for posting related stories together so readers can get the whole picture. The "Will Linux Save Microsoft?" article seems inexplicable until you read the preceding "Crack for Sale" headline, and realize what Hal Plotkin was smoking.
Here's a clue: resize /.'s window - what do you notice?
I notice that at 640x480, that row of section category icons at the top of each page causes both Mozilla and IE5 to require a horizontal scrollbar to render the page correctly.
You see, while Slashdot avoids all the fixed table size travesties which make half the web look like a thin column (subtly phallic - does it imply that the web developer was a dick?) of text at my preferred high browsing resolution, it does succumb to the temptation of those fancy "graphics" which are all the rage among the kiddies.
And even though this example could be coded around (it wouldn't be hard to make those icons wrap when the available space is too small), there are others that won't. I had a graphical button bar on the top of an old website of mine, for instance, and while it only took a little work to make it happy with 640x480, trying to view it on a palm pilot (except with a text-only browser) would be disasterous.
The future solution is not to code "The One True HTML Subset", by the way. If you do that, then you avoid having 5% of users think your page sucks, at the expense of the 95% who now think it's mediocre. The ideal solution is to provide a second stylesheet to control the layout of your pages for the smallest web browsers. Of course, this requires you to get the content/presentation separation right from the beginning, so it won't be the popular near future solution...
This was an MSNBC link, right? Not Segfault?
Are they serious? That has such a Austin Powers, "hilariously frozen in time and don't know what's going on in the rest of the world" sound to it. I can just picture Gates in his silver Dr. Evil suit, explaining to his assistants the potential of the (making quote marks in air with fingers) "commands line" and the "laser".
Kiss The Blade's hilarious MO is to make straight-faced "devil's advocate" posts with just enough misinformation and logical fallacy to send the average intelligent-but-gullible slashdot user into a blindly raging counterattack. Normally, the only reason I don't mod him (-1, troll) is that unlike the hot grits morons he's so good at it. 23 replies, 6 replies, 39 replies, 11 replies... he picks up everything from flaming rants to philosophical theses. That earns respect.
But damnit, KTB, what are you doing now???
This sets a bad precedent. The DoD is acting to prevent totally unjustified anxiety on the part of millions of Americans. I thought the function of Government Departments was to do what is right, not what people think is right
I totally agree with this. What is going on? Did you get bored and decide to write something honest instead of baiting people? Did you realize that many of us were on to you weeks ago, and decide to throw in a monkey wrench by posting something reasonable? Is your karma dangerously close to the default score: -1 threshold?
I don't detect any hyperbole, feigned ignorance, or baiting here. And what scares me is the thought that maybe KTB hasn't changed, maybe I'm just in the select group of misinformed idiots being parodied by one of his posts, and don't realize it...
(Score -1: Confused)
...it also cost 10 times what missions in the current Mars exploration program do. And not all the current missions have been failures. All we need is 1/10 to succeed and we're still coming out even.
When you boot Linux, if the root partition is marked as FAT32 in the partition table, but LILO and /etc/fstab know it to be an ext2 partition, then Linux doesn't care what the partition table says. Does BSD?
The reason I noticed at all is that my machine was swapping like mad, even with 128MB of physical RAM, and even after turning the usual memory hog culprit (Netscape) off.
NVidia, on the other hand, uses the same codebase for both their Windows and Linux drivers.
Even for the enormous Linux kernel module that's required to use their drivers? Really?
Does their Windows driver, after less than a week of use, bloat to consume over 200MB of virtual memory? That's what their closed source XFree86 driver did with my GeForce DDR, on XFree86 4.0.1 and kernel 2.4.0-test9, even without using the 3D features at all. The open source nv.o driver that came with XFree86 isn't exactly a spartan RAM user either, but at least after it's sucked up a big chunk it stops asking for more.
Granted, they don't seem to care about keeping up with development kernels (their kernel module didn't even compile against 2.4-test for a while); I haven't exactly put much work into fixing the problem (but how can I, when I can't even recompile with debugging symbols?); and their drivers did seem to work OK with kernel 2.2.16.
Nevertheless, I don't intend to buy another NVidia card until I have an open source 3D driver to run it with. By contrast, my previous 3D acceleration in Linux came from Mesa on top of Voodoo2 glide; the frame rate may not have been as fast, but the rate of driver improvment certainly was faster.
Is this really an advantage of IRV? I'd think that any system which allows voters to do a complete preference ranking of candidates would allow multiple winners to be selected by one vote: Select first place winner; remove that candidate from everyone's vote; repeat until all seats are filled.
Someone want to moderate me down, and The Pim up? Finally, a chance for someone to use "Overrated" fairly...
What you incorrectly call "lying", most people call setting priorities.
No, a ballot that said "I prefer Browne to Nader to Gore to Bush" would be setting priorities (plural). A ballot that says "I prefer Gore to everyone else" sets a priority (singular), and millions of voters regularly set that priority in a way that makes it an inaccurate statement, a "lie".
Voters in the U.S. here have more power because they can decide they dislike a candidate so much that they'll vote for a stronger candidate who would otherwise be their second or third choice just to be able to knock the guy they dislke off.
Did you read about approval voting? Instant runoff voting? Those systems let you express an "I don't want this person to be president" opinion just as strongly; they just don't force you to sacrifice your other opinions in a multicandidate race to do it.
This leads to candidates who not only worry about energizing their base, but also worry about being considerate enough of their opposition to not unduly piss them off.
No, it leads to candidates who cannot afford be considerate of their opposition, lest their least moderate "base" voters splinter off to a more extreme candidate and fracture their party. It leads to primary elections which drop candidates like McCain who have broad bipartisan appeal but less appeal among the party faithful.
On the other hand in most EU democracies,
In most EU demt EU democracies, don't they have proportional representation to ensure that the legislature is just as split as the voters? That has nothing to do with any of the voting methods being discussed; we can dump the plurality system without getting rid of the "one district, one representative" House.
This leads to extremist parties (willing to switch votes on national concerns) in a position to make or break governments
You mean like the Green Party just did?
...and completely negate the idea of a secret ballot. great.
By "ID", I mean the same ID you get on ballots in many places today: a unique number, on both the ballot and a tear-off stub, which could (in theory) be used for you to verify that your vote was counted correctly.
Imagine how much harder election fraud would be, if, instead of each district reporting "16,835 people voted for Gore", each district put a text file online saying:
PLMBCH000001 went to Gore.
PLMBCH000002 went to Bush.
PLMBCH000003 went to Buchanan.
How do you intend to vote in the next election?
* In person
* By absentee ballot
* Online
* By writing a Libertarian macrovirus
A letter to the editor I wrote to Discover:
I just finished reading your recent article, "May the Best Man Lose".
The author unfortunately underestimates one of the greatest weaknesses of the plurality system, and so fails to realize that this weakness applies nearly as strongly to the Borda count: both voting systems encourage voters to lie!
Of course, the media today doesn't call strategic voting "lying", they call it "not wasting your vote". It is considered standard practice to give your vote not to the candidate you prefer, but to the poll-leading candidate you dislike least.
This practice would not change under a Borda count system. Voters who prefer Nader to Browne to Gore to Bush will still be encouraged to vote for Gore above either Nader or Browne, because that way they will add 3 points to the separation between the leading candidates rather than 1 point, and their vote will have almost three times the impact in the election.
Strategic voting makes independent and third-party candidates nearly irrelevant, and gives the Democrats and Republicans a chokehold on politics. That bipartisan chokehold, by the way, is why we may very well soon see a constitutional amendment to eliminate the electoral college, but we will never see a superior system of voting replace the plurality system. The electoral college is much less damaging to our democracy (and had less of a detrimental effect in this presidential election) than the plurality system, but removing the electoral college will result in only a minor power shift from the Democrats towards the Republicans; changing the plurality system would result in a major power shift away from both.
I was surprised to see that one important voting system was not even mentioned in your article: Instant Runoff. In that system, voters rank their candidates by preference. If no candidate has a majority of first place votes, the candidate with the fewest first place votes is removed from consideration, and from the rankings of voters who voted for him. (i.e. if Alice's second place choice was dropped, then her third place choice becomes her new second, her fourth becomes her new third, etc.) Once a candidate has a majority of first place votes, that candidate wins. The results are not guaranteed to equal Borda count results, but they often will. I don't think strategic voting is impossible, but it's a lot harder.
Don't dismiss the possibility of computer error and fraud so lightly. This year, the Republican National Committee's webpage was defacedVolusia County is doing a manual recount now, after computer error injected 15,000 erroneous Socialist Workers and Constitutionalist Party votes into their first tally. Keep in mind: the people who will be administering our first computerized elections will be these people, not Linus and Alan.
Sure, we can have physical "backstops" to try and prevent these kinds of problems... but if these backstops have to be resorted to every time an election comes this close, how is it any improvement over our current situation? If a software glitch in your voting machine causes every voter whose last name begins with Z to get 10 votes, will that same glitch from causing your voting machine to print out 10 punchcards to backstop those votes?
Don't get me wrong, there's certainly a lot of room for technological improvement here. Some Slashdotter suggested a touchscreen voting machine which would give you a clearer GUI, prevent or check for any invalid double votes, and print out a sheet with your ID and only the names of the people you voted for, in easily machine-scannable form, so you could take that sheet and give it to the poll workers. That would prevent both the ballot and the discarded vote problems in Palm Beach, at least.
It's tempting to think that there could be something even better. A little smartcard (because I don't trust the nation who let Melissa and ILOVEYOU loose to maintain their PCs securely) with public key crypto could let your "vote" be a digitally signed statement that you could safely send over the net, and the collection of all those signed statements could be publically downloadable, to allow you to check and make sure nobody tampers with your vote or the vote count. But even in that case, who would we trust to distribute the private keys, and never have their systems compromised? Verisign? Even if you can check your own vote's integrity, how do you know that a 6 million vote list isn't actually 5 million real votes, plus 1 million fraudulent inserts?
Oh, yeah, I forgot; I titled this post "Optimizing Election Fraud" for a reason. Consider: Right now, tampering with n votes is an O(n) operation. A well designed computer system could make that an O(1) operation. In most programs this would be a fantastic optimization; in this particular case it is not an improvement.
Finally, a reply.
I don't mean to sound like such a conspiracy theorist; I'm sure it was really just a computer glitch. But seriously, the fact that there were about 15,000 changing votes in a 300-vote difference race, and that the national media was almost completely ignoring this... well, it was quite disturbing. Add to this the fact that nobody on Slashdot seemed to care... and it was like dropping into the movie "Brazil..." Am I going nuts? Is everybody else already there?
Good to hear that the county's returns will be straightened out, and that the Democrats and Republicans are at least paying attention.
Check out the returns from Volusia after the recount, and compare them to the initial Volusia returns
In the recount:
Harris went from 9,888 votes to 8 votes.
Philips went from 2,927 votes to 20 votes.
Browne went from 3,211 votes to 442 votes.
Nader went from 2,436 votes to 2,903 votes.
Buchanan went from 396 votes to 496 votes.
Hagelin went from 33 votes to 36 votes.
McReynolds went from 3 votes to 5 votes.
Moorehead went from 59 votes to 69 votes.
George W. Bush and Al Gore had returns that were completely unchanged.
That's right. The recount from Volusia claims in effect, "We screwed up the vote on every other candidate, reported 16,000 erroneous votes in total... and yet managed to get the count for the major candidates exactly right!"
I hope someone's keeping track of pre-recount and post-recount totals by county in Florida. CNN.com has only pre-recount data up by county, and doesn't list Harris in it's table. FOX has both pre and post recount data by county, but only for Gore and Bush. That Daytona Beach news website basically overwrote their prerecount returns page with the postrecount returns. As soon as CBS news updates it's page, I'll be out of links to records of the initial 10,000 Harris votes.