It's easier to change 1 million ballots electronically than it is to change 1 million paper ballots. That's what computers are good at. You can have all that crypto signature crap for all you want but a doctored box will just sign the forged results as well as it signs legit results.
You might be able to make it harder to fake by adding a lot of complexity but that brings us to the most important bit of elections that many people miss:
Elections don't just have to be fair, they have to be _seen_ as fair.
Don't ever forget that. You can explain paper ballots and how they can be done securely to Joe Sixpack, and he might even understand it while drunk.
Go explain all that fancy crypto stuff to a drunk Joe Sixpack when his favourite candidate has just lost by 0.5%.
I'm an IT guy and I can tell you there is ZERO need for electronic voting. Counting can be done in parallel. The more voters a country has, the more volunteers you can get for counting and the more observers you can have to oversee stuff.
The more spread out the counting the harder it is to fake it - if the paper votes can be faked substantially, the your country is as screwed as Zimbabwe or Myanmar, in which case it doesn't really matter anymore what system you use.
I repeat: Elections don't just have to be fair, they have to be _seen_ as fair.
Otherwise you have a higher chance of riots and other nastiness (you can never avoid it totally - some people are just sore losers).
If the people from your favourite party observing the counting say - it sure looks like everyone was voting the other guy, a defeat is more likely to be accepted.
The difference between archon and battlechess is battlechess is just chess with 3d graphics animation of what happens when stuff is captured.
Whereas archon, just because you move the piece to a spot doesn't mean you get to eliminate the piece - what it does is it starts up an arcade battle between the two pieces.
The two pieces could be a knight versus a dragon, and if you are really good at the knight you could actually kill the dragon, it helps of course if the dragon was badly injured in previous battles and was not healed by the player (it costs a move to heal, so heals are rare).
Also in the first archon, white pieces and black pieces get health bonuses depending on whether the square is black, dark gray, gray, light gray or white.
Overall there was some strategy involved but it's typically overshadowed by arcade skills.
How does it even point in the right direction for exonerating Bush?
Try reading the article. I know it's a lot of words and all that, but persevere till the bitter middle and you will find:
"Israeli warplanes bombed a reactor project at the site in 1981. Later, U.N. inspectors documented and safeguarded the yellowcake, which had been stored in aging drums and containers since before the 1991 Gulf War. There was no evidence of any yellowcake dating from after 1991, the official said."
In case your memory needs refreshing, the first US vs Iraq war was in 1991 and there was great worldwide support for it. The next US vs Iraq war was in 2003 and there was not much support for it worldwide (I'm sure you still wonder why).
I bet the most US people seeing the headlines will think a similar way - "Hey Bush was right".
So it's going to be yet another wonderful "mission accomplished" by the "news people". Like shooting fish in the barrel.
RTFA: "Israeli warplanes bombed a reactor project at the site in 1981. Later, U.N. inspectors documented and safeguarded the yellowcake, which had been stored in aging drums and containers since before the 1991 Gulf War. There was no evidence of any yellowcake dating from after 1991, the official said."
But I guess many stupid/ignorant people will read the headlines and "understand" it the same way you did.
It's amazing how many people on this allegedly nerd site can't or won't do the math. Doesn't even need to be exact, just ballpark estimations will show how impractical the VW is (and many of the other "alternatives").
Guess how long it will take for the VW with its 8.5 HP engine to climb up a hill. I'm assuming that 600 pounds does not include the weight of passenger and driver. You're not going to break the laws of physics anytime soon.
Does it have an airconditioner or heater? If it doesn't, you might as well get a motorcycle. You can easily get 50-70mpg with a 250cc motorcycle and these bikes are 17-25HP and so can take you up a hill faster than the VW.
But yes it's true these bikes have a tendency of taking people to their graves much faster;). How much safer will this VW be in practice with 18 wheel trucks on the road? The car seems to be quite low.
"Solar vehicles are completely on the fringes even when in some markets they would be quite useful."
Do the math.
Assuming you're talking about solar vehicles as in vehicles with solar panels rather than electric vehicles with the solar panels elsewhere.
Say the car is efficient and uses the equivalent of 0.5 litres of petrol a day. Take the car in the article, add airconditioning/heating and you won't get 100km per litre. I'm assuming a car that can carry two people and somewhat isolate them from the environment.
1 litre of petrol has 34 megajoules. 9.5 hours * 1 kilowatt (think hair dryer power) gives you 34 megajoules.
The max amount of sunlight at the surface of the earth is 1 kilowatt per square metre (clear sky, sun directly overhead - perpendicular). The typical is probably a lot lower (evening, morning, clouds, dust etc) - 500-700 watts.
If you have 10% efficient solar panels this means you need 10 kilowatts of light to generate 1 kilowatt of electricity. For example, 10 square metres of 10% efficiency panels will produce 1 kilowatt of electricity, in an equatorial desert at noon, when the sky is perfectly clear and the sun is right above.
So you need a few days of normal sunlight before you can move the car, or very high intensity light (e.g. run the car on some other planet ) or use so many solar panels that they won't fit on the car (that means it's not a solar vehicle).
In summary, solar cars are a silly idea on this planet. If solar cars could be practical, trees and grass would be running around (away from stuff trying to eat them). As it is, the solar energy has to be gathered separately, and then you feed the stuff that moves (cow, horse, car, whatever).
By the way, if you'd like a car that can take you and a passenger up a hill in a reasonable amount of time, you're going to need a fair bit more power.
420kg (300kg car + passengers ) * 9.82 * 100 metre high hill = 410 kilojoules. So to move 420kg up 100 metres in 1 minute = 7 kilowatts.
The volkswagen in the article has a 8.5 hp engine ( 6.3 kilowatts), so guess how long it'll take to climb a 100 metre hill in practice. Add all the inefficiencies (air, rolling, drivetrain resistance) and the fact that you are not going straight up but horizontal as well, you're talking about quite a while.
So this car problem is not so simple to solve. There are many good reasons why we are still using hydrocarbons. One litre of petrol has enough energy to run a hair dryer or microwave oven for 9 hours nonstop. But it won't take an SUV very far.
To me I think in the near term we really need to go nuclear in a big way and use breeder reactors, and supplement that with solar etc where it makes sense. That will buy us some time to think of what to do after that.
Otherwise it's bye bye modern life as we know it. Back to the old days of cars and planes being only for the very rich.
Yeah I'm not surprised a 9 buck wine can be pretty good.
I'm not too embarassed to prefer a cheaper wine to a more expensive one - I don't have a great sense of taste for wine or cheese - the Bega strong and bitey cheese is one of my favourites and it's relatively cheap;). So if cheap will do, that saves some money:)
We were tasting some cheap wines to get for my brother's wedding some time ago and a Bushman's Gully and a French wine ended up on the shortlist (can't remember the details unfortunately). I actually preferred the French wine - it was smoother - but the majority said that while it was smoother it didn't have as much kick - so the vote from the people who liked higher alcohol wines went to the Bushman's Gully.
If your claim that Kerry was so similar to Bush is true then the people obviously liked Bush or someone like him (Kerry) - given that 99% of the votes went to either Bush or Kerry.
People who didn't want a Bush style president could have voted for the other candidates. There were Nader and Badnarik and a few others.
Based on what you say, since only 1% wanted someone different and 99% preferred either Bush or his twin, it means that Bush in his previous term was doing what the US voters wanted.
And so either you are wrong or 99% of the US voters are stupid/ignorant/evil.
Maybe the rest (32%) figured the existing broadband packages are substandard - with all that throttling to dial up speeds.
That'll be hard to believe though;).
I think it's more likely they just can't be bothered - it's not important enough to them - if someone suddenly installed it for them without any fuss (so that "it works") maybe they'd keep it.
But in event of nonDiebolded elections, voters still can vote for whoever they want.
They don't have to vote for the most well-financed candidates.
My conclusion is that most voters either a) don't really care that much b) actually support the status quo
Of course the "first past the post" voting system does make things tend to "two party", but if people are that pissed off they could try to get more organized and then really vote for someone different.
Are they that upset? The fact that Bush actually got reelected should give you an indication about the reality despite all the loud complainers.
The reality is as long as there's Bread and Circuses most voters don't care, the Emperors and their Senate can do whatever they want.
Now with the increasing oil prices and recession there might be a bit of a problem with the Bread and Circuses supply.
I'm not a wine expert so maybe it's my crappy tastebuds, but some cheap wine can actually taste quite good.
I've had cheap wine tasting better than some crappy expensive wine 10 times the price.
But I've also had kilobux wines and yes those were good, but I wouldn't pay the extra - they weren't so much better to my low-end tastebuds. Good that someone rich and generous paid;).
So I wouldn't be surprised if there might be very good sounding cheap violins, maybe not the dirt cheap ones;).
Refocus on what? If you're just looking at the text at the same range why would there be a need to refocus?
Looking at dim stars in the night sky doesn't force the eye to constantly refocus.
I don't know about you, maybe I'm a nocturnal sort - I find looking at stars in a dark sky more relaxing than looking at black stuff (crows?) in a bright blue sky. Maybe everything is more focused, but it sure is more of a strain.
Yeah well, I'd rather they didn't have to contract so much - they start to feel a bit tired after a while - ever looked at a bright surface for a while? How about the bright blue sky? (And see your white blood cells zooming about like little comets:) ). Gets tiring after a while.
There's not much regulating the eye needs to do when the screen doesn't move and you stick to a sane colour scheme (stuff like pure red on green/blue is probably crazy - due to chromatic aberration). If you're lucky and your vision is good or correctable to good, even if your pupils aren't that contracted your vision will be fine.
Some of these new LCDs are crazy - just so they can advertise 2000:1 contrast ratios they make the whites so bright they sear your retinas. So the popular "modern" black on white scheme starts to feel like you need ski goggles.
I've got the brightness set to 0 out of 100 on my samsung 206BW and it's still kinda bright. Trouble is at that brightness level when you play games the darker areas are too dark.
Anyway each to their own - just try out various schemes and see which ones your eyes are fine with.
I'm fine with green on black (except when you alt tab to your browser your eyes might get a shock on some black on white or some crazy black on yellow schemes;) ).
Pastels on dark grey are OK fo me too. Contrast and colour difference enough to read, but not so much that the text and background appear to be floating at different depths. Maybe when I'm older I'll go for higher contrast.
Yeah maybe someday we can upgrade to something we can adjust.
90% efficiency when running (want to stay cool - stuff stops working well when the temperature goes up[1]), and 10% efficient when sitting on the couch watching TV - to stay warm and not get fat after eating all that junk food.
"while to others it would be like a crime against humanity and all that is decent"
Actually to me it's more of whether they LIED or not.
If they said they'd keep it a secret, or "you'd only get email from us", and spam starts showing up on your unique hard to guess email address reserved for them (e.g. brand@something.random.yourdomain.com, then it's likely they lied.
If they lied about something like that, I'd say they'd lie about other stuff too.
Whatever you want to call it, P2P, camfrog, if they notice you're actually eating a lot at the "all you can eat buffet" you paid for, they'll stop you.
You think comcast has a problem with P2P just because it's "Peer to Peer"? I don't think so.
It's horrifically flawed because it won't just stop P2P, but it'll stop all those who are getting in the way of Comcast providing less service for more money.
Of course, what those ISPs should do is actually make good on their promises they made in return for the USD 200 billion they took. But who is going to make them?
Say all the zillions of cores happen to cache the same memory area (running same compute program or something).
Then one of them _writes_ to that memory area.
The rest of the cores will have to know in order for all of them to be in the same "reality".
So that address in their caches will have to be invalidated - so the next time any of them tries to read that address they'll fetch it from wherever they should (probably cheaper to do that than to forcibly cause them to all fetch - it may turn out they never read from that area).
It's all very easy if the cores are treated separately. That's just like having thousands of PCs that aren't connected to each other and can run independently - and only submit their results at the end (just like those public SETI and protein folding stuff).
But many interesting problems aren't so easy. For these problems at some points you'll have to serialize stuff. I don't think you can ever get away from it.
If it's so easy to coordinate parallel computers to serialize stuff and not mess up AND be fast. Then whether you have 1000 cores on a chip or 500 computers with 2 cores each starts to matter less for the computation.
Now the advantage of 1000 cores on a chip of course is they're close to each other- so naturally the locking and stuff will be faster.
Slashdot Mirror
How about a religious president hearing a voice saying: "This is God, you should invade Iran".
It's easier to change 1 million ballots electronically than it is to change 1 million paper ballots. That's what computers are good at. You can have all that crypto signature crap for all you want but a doctored box will just sign the forged results as well as it signs legit results.
You might be able to make it harder to fake by adding a lot of complexity but that brings us to the most important bit of elections that many people miss:
Elections don't just have to be fair, they have to be _seen_ as fair.
Don't ever forget that. You can explain paper ballots and how they can be done securely to Joe Sixpack, and he might even understand it while drunk.
Go explain all that fancy crypto stuff to a drunk Joe Sixpack when his favourite candidate has just lost by 0.5%.
I'm an IT guy and I can tell you there is ZERO need for electronic voting. Counting can be done in parallel. The more voters a country has, the more volunteers you can get for counting and the more observers you can have to oversee stuff.
The more spread out the counting the harder it is to fake it - if the paper votes can be faked substantially, the your country is as screwed as Zimbabwe or Myanmar, in which case it doesn't really matter anymore what system you use.
I repeat:
Elections don't just have to be fair, they have to be _seen_ as fair.
Otherwise you have a higher chance of riots and other nastiness (you can never avoid it totally - some people are just sore losers).
If the people from your favourite party observing the counting say - it sure looks like everyone was voting the other guy, a defeat is more likely to be accepted.
A bit like this:
http://www.youtube.com/watch?v=W6MoTUJiOiU
I'm not the one playing.
You can probably download one of those emulators and play it - C64, Apple 2 etc.
The difference between archon and battlechess is battlechess is just chess with 3d graphics animation of what happens when stuff is captured.
Whereas archon, just because you move the piece to a spot doesn't mean you get to eliminate the piece - what it does is it starts up an arcade battle between the two pieces.
The two pieces could be a knight versus a dragon, and if you are really good at the knight you could actually kill the dragon, it helps of course if the dragon was badly injured in previous battles and was not healed by the player (it costs a move to heal, so heals are rare).
Also in the first archon, white pieces and black pieces get health bonuses depending on whether the square is black, dark gray, gray, light gray or white.
Overall there was some strategy involved but it's typically overshadowed by arcade skills.
How does it even point in the right direction for exonerating Bush?
Try reading the article. I know it's a lot of words and all that, but persevere till the bitter middle and you will find:
"Israeli warplanes bombed a reactor project at the site in 1981. Later, U.N. inspectors documented and safeguarded the yellowcake, which had been stored in aging drums and containers since before the 1991 Gulf War. There was no evidence of any yellowcake dating from after 1991, the official said."
In case your memory needs refreshing, the first US vs Iraq war was in 1991 and there was great worldwide support for it. The next US vs Iraq war was in 2003 and there was not much support for it worldwide (I'm sure you still wonder why).
I bet the most US people seeing the headlines will think a similar way - "Hey Bush was right".
So it's going to be yet another wonderful "mission accomplished" by the "news people". Like shooting fish in the barrel.
Thanks media alright.
RTFA: "Israeli warplanes bombed a reactor project at the site in 1981. Later, U.N. inspectors documented and safeguarded the yellowcake, which had been stored in aging drums and containers since before the 1991 Gulf War. There was no evidence of any yellowcake dating from after 1991, the official said."
But I guess many stupid/ignorant people will read the headlines and "understand" it the same way you did.
No wonder Bush got re-elected.
I thought Monsanto owns their first born.
;).
With all those gene patents and "no unauthorized reproduction"
It's amazing how many people on this allegedly nerd site can't or won't do the math. Doesn't even need to be exact, just ballpark estimations will show how impractical the VW is (and many of the other "alternatives").
Guess how long it will take for the VW with its 8.5 HP engine to climb up a hill. I'm assuming that 600 pounds does not include the weight of passenger and driver. You're not going to break the laws of physics anytime soon.
Does it have an airconditioner or heater? If it doesn't, you might as well get a motorcycle. You can easily get 50-70mpg with a 250cc motorcycle and these bikes are 17-25HP and so can take you up a hill faster than the VW.
Example: http://en.wikipedia.org/wiki/Kawasaki_Ninja_250R
But yes it's true these bikes have a tendency of taking people to their graves much faster ;). How much safer will this VW be in practice with 18 wheel trucks on the road? The car seems to be quite low.
"Solar vehicles are completely on the fringes even when in some markets they would be quite useful."
Do the math.
Assuming you're talking about solar vehicles as in vehicles with solar panels rather than electric vehicles with the solar panels elsewhere.
Say the car is efficient and uses the equivalent of 0.5 litres of petrol a day. Take the car in the article, add airconditioning/heating and you won't get 100km per litre. I'm assuming a car that can carry two people and somewhat isolate them from the environment.
1 litre of petrol has 34 megajoules. 9.5 hours * 1 kilowatt (think hair dryer power) gives you 34 megajoules.
The max amount of sunlight at the surface of the earth is 1 kilowatt per square metre (clear sky, sun directly overhead - perpendicular). The typical is probably a lot lower (evening, morning, clouds, dust etc) - 500-700 watts.
If you have 10% efficient solar panels this means you need 10 kilowatts of light to generate 1 kilowatt of electricity. For example, 10 square metres of 10% efficiency panels will produce 1 kilowatt of electricity, in an equatorial desert at noon, when the sky is perfectly clear and the sun is right above.
So you need a few days of normal sunlight before you can move the car, or very high intensity light (e.g. run the car on some other planet ) or use so many solar panels that they won't fit on the car (that means it's not a solar vehicle).
In summary, solar cars are a silly idea on this planet. If solar cars could be practical, trees and grass would be running around (away from stuff trying to eat them). As it is, the solar energy has to be gathered separately, and then you feed the stuff that moves (cow, horse, car, whatever).
By the way, if you'd like a car that can take you and a passenger up a hill in a reasonable amount of time, you're going to need a fair bit more power.
420kg (300kg car + passengers ) * 9.82 * 100 metre high hill = 410 kilojoules. So to move 420kg up 100 metres in 1 minute = 7 kilowatts.
The volkswagen in the article has a 8.5 hp engine ( 6.3 kilowatts), so guess how long it'll take to climb a 100 metre hill in practice. Add all the inefficiencies (air, rolling, drivetrain resistance) and the fact that you are not going straight up but horizontal as well, you're talking about quite a while.
So this car problem is not so simple to solve. There are many good reasons why we are still using hydrocarbons. One litre of petrol has enough energy to run a hair dryer or microwave oven for 9 hours nonstop. But it won't take an SUV very far.
To me I think in the near term we really need to go nuclear in a big way and use breeder reactors, and supplement that with solar etc where it makes sense. That will buy us some time to think of what to do after that.
Otherwise it's bye bye modern life as we know it. Back to the old days of cars and planes being only for the very rich.
Yeah I'm not surprised a 9 buck wine can be pretty good.
;). So if cheap will do, that saves some money :)
I'm not too embarassed to prefer a cheaper wine to a more expensive one - I don't have a great sense of taste for wine or cheese - the Bega strong and bitey cheese is one of my favourites and it's relatively cheap
We were tasting some cheap wines to get for my brother's wedding some time ago and a Bushman's Gully and a French wine ended up on the shortlist (can't remember the details unfortunately). I actually preferred the French wine - it was smoother - but the majority said that while it was smoother it didn't have as much kick - so the vote from the people who liked higher alcohol wines went to the Bushman's Gully.
I'm not the confused one here.
If your claim that Kerry was so similar to Bush is true then the people obviously liked Bush or someone like him (Kerry) - given that 99% of the votes went to either Bush or Kerry.
People who didn't want a Bush style president could have voted for the other candidates. There were Nader and Badnarik and a few others.
Based on what you say, since only 1% wanted someone different and 99% preferred either Bush or his twin, it means that Bush in his previous term was doing what the US voters wanted.
And so either you are wrong or 99% of the US voters are stupid/ignorant/evil.
I'm not confused at all. I repeat: "The fact that Bush actually got reelected should give you an indication about the reality".
If people were that pissed off they'd vote for Kerry anyway.
With the "1st past the post" system, the problem is pissed off people need to get coordinated to agree to vote for the same "Not Bush" candidate.
Maybe the rest (32%) figured the existing broadband packages are substandard - with all that throttling to dial up speeds.
;).
That'll be hard to believe though
I think it's more likely they just can't be bothered - it's not important enough to them - if someone suddenly installed it for them without any fuss (so that "it works") maybe they'd keep it.
But in event of nonDiebolded elections, voters still can vote for whoever they want.
They don't have to vote for the most well-financed candidates.
My conclusion is that most voters either
a) don't really care that much
b) actually support the status quo
Of course the "first past the post" voting system does make things tend to "two party", but if people are that pissed off they could try to get more organized and then really vote for someone different.
Are they that upset? The fact that Bush actually got reelected should give you an indication about the reality despite all the loud complainers.
The reality is as long as there's Bread and Circuses most voters don't care, the Emperors and their Senate can do whatever they want.
Now with the increasing oil prices and recession there might be a bit of a problem with the Bread and Circuses supply.
I think it's more like wine :).
;).
;).
I'm not a wine expert so maybe it's my crappy tastebuds, but some cheap wine can actually taste quite good.
I've had cheap wine tasting better than some crappy expensive wine 10 times the price.
But I've also had kilobux wines and yes those were good, but I wouldn't pay the extra - they weren't so much better to my low-end tastebuds. Good that someone rich and generous paid
So I wouldn't be surprised if there might be very good sounding cheap violins, maybe not the dirt cheap ones
Refocus on what? If you're just looking at the text at the same range why would there be a need to refocus?
Looking at dim stars in the night sky doesn't force the eye to constantly refocus.
I don't know about you, maybe I'm a nocturnal sort - I find looking at stars in a dark sky more relaxing than looking at black stuff (crows?) in a bright blue sky. Maybe everything is more focused, but it sure is more of a strain.
"Brightness will make the pupils contract"
:) ). Gets tiring after a while.
;) ).
Yeah well, I'd rather they didn't have to contract so much - they start to feel a bit tired after a while - ever looked at a bright surface for a while? How about the bright blue sky? (And see your white blood cells zooming about like little comets
There's not much regulating the eye needs to do when the screen doesn't move and you stick to a sane colour scheme (stuff like pure red on green/blue is probably crazy - due to chromatic aberration). If you're lucky and your vision is good or correctable to good, even if your pupils aren't that contracted your vision will be fine.
Some of these new LCDs are crazy - just so they can advertise 2000:1 contrast ratios they make the whites so bright they sear your retinas. So the popular "modern" black on white scheme starts to feel like you need ski goggles.
I've got the brightness set to 0 out of 100 on my samsung 206BW and it's still kinda bright. Trouble is at that brightness level when you play games the darker areas are too dark.
Anyway each to their own - just try out various schemes and see which ones your eyes are fine with.
I'm fine with green on black (except when you alt tab to your browser your eyes might get a shock on some black on white or some crazy black on yellow schemes
Pastels on dark grey are OK fo me too. Contrast and colour difference enough to read, but not so much that the text and background appear to be floating at different depths. Maybe when I'm older I'll go for higher contrast.
Yeah maybe someday we can upgrade to something we can adjust.
90% efficiency when running (want to stay cool - stuff stops working well when the temperature goes up[1]), and 10% efficient when sitting on the couch watching TV - to stay warm and not get fat after eating all that junk food.
[1] "muscles tire because they get too hot"
http://www.wired.com/wired/archive/15.03/bemore_pr.html
http://news.bbc.co.uk/1/hi/health/2354135.stm
Which is part of the parallizing process you're talking about?
If you talk about removing the serialization bits, I'd say it's practically impossible (unless there's a way of doing time travel or something).
If it was so simple, you'd be able to run everything on separate isolated computers in parallel - no need for low latency interconnects.
"while to others it would be like a crime against humanity and all that is decent"
Actually to me it's more of whether they LIED or not.
If they said they'd keep it a secret, or "you'd only get email from us", and spam starts showing up on your unique hard to guess email address reserved for them (e.g. brand@something.random.yourdomain.com, then it's likely they lied.
If they lied about something like that, I'd say they'd lie about other stuff too.
Whatever you want to call it, P2P, camfrog, if they notice you're actually eating a lot at the "all you can eat buffet" you paid for, they'll stop you.
You think comcast has a problem with P2P just because it's "Peer to Peer"? I don't think so.
It's horrifically flawed because it won't just stop P2P, but it'll stop all those who are getting in the way of Comcast providing less service for more money.
Of course, what those ISPs should do is actually make good on their promises they made in return for the USD 200 billion they took. But who is going to make them?
Try this:
http://www.vmware.com/appliances/directory/
I doubt most have even got around to reading the paper. They're too busy thinking of ways to display boobies on their calculators.
Just look at the above threads.
I'm not a CPU guy.
Say all the zillions of cores happen to cache the same memory area (running same compute program or something).
Then one of them _writes_ to that memory area.
The rest of the cores will have to know in order for all of them to be in the same "reality".
So that address in their caches will have to be invalidated - so the next time any of them tries to read that address they'll fetch it from wherever they should (probably cheaper to do that than to forcibly cause them to all fetch - it may turn out they never read from that area).
It's all very easy if the cores are treated separately. That's just like having thousands of PCs that aren't connected to each other and can run independently - and only submit their results at the end (just like those public SETI and protein folding stuff).
But many interesting problems aren't so easy. For these problems at some points you'll have to serialize stuff. I don't think you can ever get away from it.
If it's so easy to coordinate parallel computers to serialize stuff and not mess up AND be fast. Then whether you have 1000 cores on a chip or 500 computers with 2 cores each starts to matter less for the computation.
Now the advantage of 1000 cores on a chip of course is they're close to each other- so naturally the locking and stuff will be faster.
Yep like I said: http://news.slashdot.org/comments.pl?sid=601309&cid=24034091
Summary:
How many free market economists does it take to change a lightbulb?
None.
Free market economists don't change light bulbs. They write their papers in the darkness while waiting for Adam Smith's Invisible Hand to change it.