Actually, all it needs to do is create a glowing plasma ball out of reach and leave it there. It would be a built-in bug zapper as it attracts the bugs and zaps them both in a single package. No need to try to track and shoot the things...
Then what is to stop somebody from making lots of promises and having a business partner do all the suing? That would essentially nullify promissory estoppel.
Re:Require login, forbid any subdirectory access.
on
Full-Disclosure Wins Again
·
· Score: 2, Informative
If you store your session IDs in a central database you'd be covered. Maybe under extremely high load this might be an issue, but often these bugs crop up in software that doesn't face these sorts of high-demand applications.
I would think so. If an end-user can argue that they relied on the promise and acted on it, then the promise becomes binding. Microsoft would inherit these sorts of obligations even as they inherit any assets should they buy Novell out. This is what often saves companies with massive problems from buyouts - nobody wants to buy the headaches even for a cheap price - a company with sufficient liabilities could have negative value.
There would be no way for a human to sense whether the computer is bluffing other than in its betting patterns, so if the computer can't sense the biological state of the human it isn't really at a disadvantage.
I don't see any reason why a well-designed algorithm couldn't defeat a human at poker. The key would be having sufficient randomness that a human couldn't tell what it was up to, or the ability to make a human think it has an idea of what is going on when in fact it doesn't.
There is more to poker than just looking for tells.
Are we certain that Kasperov didn't ask any friends for advice in defeating Big Blue between matches? I imagine that humans getting advice from each other is rather common in such matches - just not during play.
A Verizon tech drilling a hole through an electrical circuit is like a sysadmin who sets a HDD on fire by improperly wiring the drive. It's amusing.
How about ordering a $2 power cable splitter and cooking several hard drives before discovering that the leads were backwards on one of the connectors (the drives were new and were assumed faulty before anybody thought to check the connector)? So much for idiot-proof fits-one-way connectors...
I once worked at a major retail store in high school, and one day a guy walked in wearing a hard hat and indicated that if they didn't FAX a copy of a check to the electric company and hand it to him in 15 minutes he'd be cutting the power. You never saw so many managers scramble!
You'd be amazed at the little details that get missed by major corporations. Just look at microsoft forgetting to send in their $8 payment for hotmail.com...
Surely you overreact. They submitted the email on Sat the 11th and as of the 13th when they filed their request with the courts they had not gotten a reply (probably at 9AM).
I'm all for hammering the RIAA, but it would seem reasonable to give them at least a few business days to cough up a check for almost $100k. If the RIAA filed a complaint on the basis of not getting a response to an email over the weekend we'd all be first in line to complain about that.
It really depends on intended use - but a range of motherboards with IO options would be nice.
Personally, I'd like to see:
100MB ethernet TV-out (RCA or S-video) SPDIF DVI stereo audio out USB (a few) 1 DIMM slot
Of course, this is for a multimedia application - diskless mythtv front end. If you add an IDE interface for a compact-flash drive you expand the uses and it becomes a pretty generic multimedia front-end.
My point is that spending millions of dollars throwing people in jail for driving drowsy is unlikely to put much of a dent in the problem. Sure, it is both dangerous and illegal - the problem is that most of the people doing it don't see much of an alternative. Some people work multiple jobs and are tired pretty-much around the clock - the only way you're going to get them off the roads is by putting them in prison. There are quite a few of them around - and it is unlikely you'll be able to jail all of them. It is also very difficult to enforce - there is no breathalyzer for tiredness, and most likely those drowsy drivers would snap to attention if pulled over by an officer.
The only way to get rid of impaired drivers is by getting rid of drivers entirely! Anything else amounts to a feel-good measure - you can throw lots of people in jail but it won't fix the problem.
$5/month is a bit steep. That is what most commercial DVR services cost. However, Mythtv users often have to pay for multiple receivers if they want dual-tuners (which are pretty-much standard on commercial DVRs these days). That brings the price up to over $10/month in many cases with the extra $5 for TV listings.
At $20/yr, however, it isn't a big deal. It looks like schedulesdirect is non-profit, so that should help keep prices down (as long as salaries don't get too high - technically you can probably be a non-profit while paying your founders $200k/yr for services rendered).
Relax - while I'm sure you've never done it I'm guessing that 80% of the driving public has driven impaired in some way. Sure, it is VERY dangerous. The problem is that there isn't a practical solution at this point. Sure, you can just say "well, get more sleep" - but lots of things contribute to tiredness beyond lack of sleep. And as a society we've created a culture that frowns upon getting 8+ hours of sleep every night.
If one is looking for a technical solution a better one might be self-piloted cars, personal rapid transit, or some other system that takes human drivers out of the loop. Such a system would solve a lot of other problems as well (road congestion, parking, cheap transit for those who can't afford cars, almost all car accidents (assuming complete enforcement of maintenance - something achievable when the cars are computer-controlled), etc).
Simply throwing the book at anybody caught doing something that 80% of the population does is going to be about as effective as passing a death-penalty speeding law.
Actually, it might make a very powerful sniper. Just parachute these onto every hilltop in an area. 95% will never be used, but if combat erupts chances are that there are 3 sniperbots in perfect position. And there is no reason that it can't fire a burst of rounds in a pattern to compensate for less accurate targetting. Maybe with 3 rifles at a time.
Imagine putting these on rooftops overlooking most of an occupied city. When a gunshot goes off the sound is triangulated and every sniperbot in the area takes aim at the vicinity. A human operator looks at the wide-field spotting camera and designates targets which the various bots start picking off only seconds after the original gunshot. The same tactic is used for counterbattery fire but on a much larger scale.
The things you mention about snipers laying low for long periods of time and creeping around are exactly what makes a bot so well-suited to the task. A bot can remain completely motionless for a month and then roar into action with a lot accuracy than a human whose muscles have been unused for a long period of time.
Humans have their advantages as well and they will certainly continue to be employed. However, there will be missions for which a robotic sniper will be better-suited than a person. Just as there are missions where a robotic UAV that has 24 hour endurance without a yawn is ideal.
That's essentially the tactic used on Germany after WWI, and all it did was create an angry and resentful population which led to nationalism, racism and Hitler.
And what was done in Germany and Japan after WWII? That turned out fairly well in the long term.
However, it was essentially the same sort of thing (please feel free to point out what was different), and I imagine that the tactics were much more heavy-handed than in the present war. Additionally the die-hards were mostly killed off in the regular combat.
Hey, I'm not a big fan of Iraq, but you have to admit that the US is being a LOT nicer than any of the allies were in WWII...
And define incorrect - 99%-101% filled in? Stray marks up to 10 microns in size? No paper ballot is perfect, so my stray mark might be your printing defect.
With a machine-generated paper ballot you can define specs and apply specs and six-sigma the final product to near-perfection. If you can make a skyscraper with 10,000 I-beams that doesn't collapse you can make a voting system that doesn't miscount votes. But not when everything is hand-made...
And whose votes are most likely to get tossed under an ultra-strict reading of the rules? If the answer isn't nobody-in-particular good luck!
Actually, I believe the Prius uses a 3-way transmission - both the electric and gasoline motors are coupled to each other and to the drive train. So either can power the wheels in the absence of the other. The electric motor also doubles as a starter by cranking the gas engine. Transmission is continuously-variable eliminating the biggest issue with gas engines - limited ideal RPMs.
I agree that both systems have their advantages - I'd just be hesitant to call one solution "the best" without some empirical data. I'm sure the engineers at Toyota aren't idiots.
Hybrid locomotives are usually diesel->electric->motors->wheels, but that is probably because of the REALLY low RPMs and VERY high torques that would probably defy a practical transmission design. They might or might not have taken the same approach for a light passenger train as they take for the huge diesel freight trains.
I didn't notice this in the replies to this post, but there is a big problem with paper that has nothing to do with convenience/speed of counting votes - validation of votes.
A computer program can present a GUI that forces all ballots to be valid. Votes are binary yes/no and clearly indicated. Voting for a second candidate for the same office unselects the first candidate or gives a nasty error message. The resulting ballot (whether digital or paper) can potentially be unambiguously and accurately counted by hand or by computer.
A paper ballot can have stray marks. It can be checked in more than one place for an office. Suppose somebody makes a small check for Bush and a slightly larger check for Gore - who gets the vote? What happens if one check is circled. What happens if one is scratched out. What if the cross-out just makes it look like an X instead of a check? When you have millions of ballots cast you can find just about any imaginable scenario in sufficient quantity to effect a close election.
Personally I think we should use each technology for what it is best-suited for. Use the GUI to validate the ballot - it works great for that. Tally the vote electronically - it works great for that too. Print out a very clear paper audit trail that the voter verifies but cannot touch. Thorougly and randomly audit the results - if you're paranoid make it 100% or even 300% every time. In the event of discrepancy the audit trail is the official record.
Every ballot is valid, clear, and unambiguous. Everything ends up on paper where it is hard to tamper with. There is no reason everything can't be open-source as well. However, if the process is well-designed the machine could just be a black box - ultimately as long as the official record is shown to the voter then it really don't matter how it gets generated.
And none of this is really that hard to do. I'm sure I could have written a text-mode program that was barely sufficient in middle school.
Oh yeah - when half the machines show up broken to a ward treat it like a homicide and nail the offender to the wall. Voting shenanigans shouldn't just be material for the 6PM news - they are an affront to the democratic process and need to be treated with the greatest serverity. I'm astonished about how brazen people are with tampering with elections - it is as if it is considered expected and as a result ignored.
I've had my tivo die a few times - took reimaging to get it working again.
That is one of the biggest reasons I went with Myth - I wanted some control over my data. Now I can make my video partition as big as I want, or shrink it at any time I feel like it. I can RAID-5 my storage and not worry about losing 250 hours of TV when a hard drive crashes (a matter of when and not if). I can skip commercials with one click 95% of the time, and jump arbitrary numbers of minutes when that doesn't work. The files are mostly just standard mpg files so I can do what I want with them easily enough (ok - minor rant - what possessed them to write their own player instead of just using an established one as the core? It is WAY to finicky about input file formats...).
Sure, it isn't perfect, but it beats a green screen of death with no easy way to recover and resorting to all kinds of shifty websites to download images.
Slashdot Mirror
Actually, all it needs to do is create a glowing plasma ball out of reach and leave it there. It would be a built-in bug zapper as it attracts the bugs and zaps them both in a single package. No need to try to track and shoot the things...
Then what is to stop somebody from making lots of promises and having a business partner do all the suing? That would essentially nullify promissory estoppel.
If you store your session IDs in a central database you'd be covered. Maybe under extremely high load this might be an issue, but often these bugs crop up in software that doesn't face these sorts of high-demand applications.
I would think so. If an end-user can argue that they relied on the promise and acted on it, then the promise becomes binding. Microsoft would inherit these sorts of obligations even as they inherit any assets should they buy Novell out. This is what often saves companies with massive problems from buyouts - nobody wants to buy the headaches even for a cheap price - a company with sufficient liabilities could have negative value.
There would be no way for a human to sense whether the computer is bluffing other than in its betting patterns, so if the computer can't sense the biological state of the human it isn't really at a disadvantage.
I don't see any reason why a well-designed algorithm couldn't defeat a human at poker. The key would be having sufficient randomness that a human couldn't tell what it was up to, or the ability to make a human think it has an idea of what is going on when in fact it doesn't.
There is more to poker than just looking for tells.
Are we certain that Kasperov didn't ask any friends for advice in defeating Big Blue between matches? I imagine that humans getting advice from each other is rather common in such matches - just not during play.
Yes, but do you really want AI researchers devising machines that blend into society and excel at stalking and hunting and eating prey? :)
A Verizon tech drilling a hole through an electrical circuit is like a sysadmin who sets a HDD on fire by improperly wiring the drive. It's amusing.
How about ordering a $2 power cable splitter and cooking several hard drives before discovering that the leads were backwards on one of the connectors (the drives were new and were assumed faulty before anybody thought to check the connector)? So much for idiot-proof fits-one-way connectors...
I once worked at a major retail store in high school, and one day a guy walked in wearing a hard hat and indicated that if they didn't FAX a copy of a check to the electric company and hand it to him in 15 minutes he'd be cutting the power. You never saw so many managers scramble!
You'd be amazed at the little details that get missed by major corporations. Just look at microsoft forgetting to send in their $8 payment for hotmail.com...
Surely you overreact. They submitted the email on Sat the 11th and as of the 13th when they filed their request with the courts they had not gotten a reply (probably at 9AM).
I'm all for hammering the RIAA, but it would seem reasonable to give them at least a few business days to cough up a check for almost $100k. If the RIAA filed a complaint on the basis of not getting a response to an email over the weekend we'd all be first in line to complain about that.
It really depends on intended use - but a range of motherboards with IO options would be nice.
Personally, I'd like to see:
100MB ethernet
TV-out (RCA or S-video)
SPDIF
DVI
stereo audio out
USB (a few)
1 DIMM slot
Of course, this is for a multimedia application - diskless mythtv front end. If you add an IDE interface for a compact-flash drive you expand the uses and it becomes a pretty generic multimedia front-end.
My point is that spending millions of dollars throwing people in jail for driving drowsy is unlikely to put much of a dent in the problem. Sure, it is both dangerous and illegal - the problem is that most of the people doing it don't see much of an alternative. Some people work multiple jobs and are tired pretty-much around the clock - the only way you're going to get them off the roads is by putting them in prison. There are quite a few of them around - and it is unlikely you'll be able to jail all of them. It is also very difficult to enforce - there is no breathalyzer for tiredness, and most likely those drowsy drivers would snap to attention if pulled over by an officer.
The only way to get rid of impaired drivers is by getting rid of drivers entirely! Anything else amounts to a feel-good measure - you can throw lots of people in jail but it won't fix the problem.
$5/month is a bit steep. That is what most commercial DVR services cost. However, Mythtv users often have to pay for multiple receivers if they want dual-tuners (which are pretty-much standard on commercial DVRs these days). That brings the price up to over $10/month in many cases with the extra $5 for TV listings.
At $20/yr, however, it isn't a big deal. It looks like schedulesdirect is non-profit, so that should help keep prices down (as long as salaries don't get too high - technically you can probably be a non-profit while paying your founders $200k/yr for services rendered).
Relax - while I'm sure you've never done it I'm guessing that 80% of the driving public has driven impaired in some way. Sure, it is VERY dangerous. The problem is that there isn't a practical solution at this point. Sure, you can just say "well, get more sleep" - but lots of things contribute to tiredness beyond lack of sleep. And as a society we've created a culture that frowns upon getting 8+ hours of sleep every night.
If one is looking for a technical solution a better one might be self-piloted cars, personal rapid transit, or some other system that takes human drivers out of the loop. Such a system would solve a lot of other problems as well (road congestion, parking, cheap transit for those who can't afford cars, almost all car accidents (assuming complete enforcement of maintenance - something achievable when the cars are computer-controlled), etc).
Simply throwing the book at anybody caught doing something that 80% of the population does is going to be about as effective as passing a death-penalty speeding law.
Actually, it might make a very powerful sniper. Just parachute these onto every hilltop in an area. 95% will never be used, but if combat erupts chances are that there are 3 sniperbots in perfect position. And there is no reason that it can't fire a burst of rounds in a pattern to compensate for less accurate targetting. Maybe with 3 rifles at a time.
Imagine putting these on rooftops overlooking most of an occupied city. When a gunshot goes off the sound is triangulated and every sniperbot in the area takes aim at the vicinity. A human operator looks at the wide-field spotting camera and designates targets which the various bots start picking off only seconds after the original gunshot. The same tactic is used for counterbattery fire but on a much larger scale.
The things you mention about snipers laying low for long periods of time and creeping around are exactly what makes a bot so well-suited to the task. A bot can remain completely motionless for a month and then roar into action with a lot accuracy than a human whose muscles have been unused for a long period of time.
Humans have their advantages as well and they will certainly continue to be employed. However, there will be missions for which a robotic sniper will be better-suited than a person. Just as there are missions where a robotic UAV that has 24 hour endurance without a yawn is ideal.
That's essentially the tactic used on Germany after WWI, and all it did was create an angry and resentful population which led to nationalism, racism and Hitler.
And what was done in Germany and Japan after WWII? That turned out fairly well in the long term.
However, it was essentially the same sort of thing (please feel free to point out what was different), and I imagine that the tactics were much more heavy-handed than in the present war. Additionally the die-hards were mostly killed off in the regular combat.
Hey, I'm not a big fan of Iraq, but you have to admit that the US is being a LOT nicer than any of the allies were in WWII...
Good luck with that in a close election!
And define incorrect - 99%-101% filled in? Stray marks up to 10 microns in size? No paper ballot is perfect, so my stray mark might be your printing defect.
With a machine-generated paper ballot you can define specs and apply specs and six-sigma the final product to near-perfection. If you can make a skyscraper with 10,000 I-beams that doesn't collapse you can make a voting system that doesn't miscount votes. But not when everything is hand-made...
And whose votes are most likely to get tossed under an ultra-strict reading of the rules? If the answer isn't nobody-in-particular good luck!
Only issue with this is reliability.
What percentage of ballots will validate OK in the first count, and have a stray mark in the recount?
I'm sure it isn't 0.0000000%
A computer-printed ballot is more likely to be accurate on all passes, and I can't see any issues with this solution security-wise.
No solution is perfect - but we can aim for the best solution possible.
Actually, I believe the Prius uses a 3-way transmission - both the electric and gasoline motors are coupled to each other and to the drive train. So either can power the wheels in the absence of the other. The electric motor also doubles as a starter by cranking the gas engine. Transmission is continuously-variable eliminating the biggest issue with gas engines - limited ideal RPMs.
I agree that both systems have their advantages - I'd just be hesitant to call one solution "the best" without some empirical data. I'm sure the engineers at Toyota aren't idiots.
Hybrid locomotives are usually diesel->electric->motors->wheels, but that is probably because of the REALLY low RPMs and VERY high torques that would probably defy a practical transmission design. They might or might not have taken the same approach for a light passenger train as they take for the huge diesel freight trains.
What do you do with half-filled circles and stray marks? What if they fill in Bush but write in Gore?
That is the big problem with paper ballots - they can be incorrectly filled out.
I'm for a hybrid approach - GUI for ballot validation, secure voter-verified paper audit trail for counting...
I didn't notice this in the replies to this post, but there is a big problem with paper that has nothing to do with convenience/speed of counting votes - validation of votes.
A computer program can present a GUI that forces all ballots to be valid. Votes are binary yes/no and clearly indicated. Voting for a second candidate for the same office unselects the first candidate or gives a nasty error message. The resulting ballot (whether digital or paper) can potentially be unambiguously and accurately counted by hand or by computer.
A paper ballot can have stray marks. It can be checked in more than one place for an office. Suppose somebody makes a small check for Bush and a slightly larger check for Gore - who gets the vote? What happens if one check is circled. What happens if one is scratched out. What if the cross-out just makes it look like an X instead of a check? When you have millions of ballots cast you can find just about any imaginable scenario in sufficient quantity to effect a close election.
Personally I think we should use each technology for what it is best-suited for. Use the GUI to validate the ballot - it works great for that. Tally the vote electronically - it works great for that too. Print out a very clear paper audit trail that the voter verifies but cannot touch. Thorougly and randomly audit the results - if you're paranoid make it 100% or even 300% every time. In the event of discrepancy the audit trail is the official record.
Every ballot is valid, clear, and unambiguous. Everything ends up on paper where it is hard to tamper with. There is no reason everything can't be open-source as well. However, if the process is well-designed the machine could just be a black box - ultimately as long as the official record is shown to the voter then it really don't matter how it gets generated.
And none of this is really that hard to do. I'm sure I could have written a text-mode program that was barely sufficient in middle school.
Oh yeah - when half the machines show up broken to a ward treat it like a homicide and nail the offender to the wall. Voting shenanigans shouldn't just be material for the 6PM news - they are an affront to the democratic process and need to be treated with the greatest serverity. I'm astonished about how brazen people are with tampering with elections - it is as if it is considered expected and as a result ignored.
But less mechanical->electrical->mechanical conversion during cruising. Not sure which is the worse trade-off (no transmission vs no conversion).
I've had my tivo die a few times - took reimaging to get it working again.
That is one of the biggest reasons I went with Myth - I wanted some control over my data. Now I can make my video partition as big as I want, or shrink it at any time I feel like it. I can RAID-5 my storage and not worry about losing 250 hours of TV when a hard drive crashes (a matter of when and not if). I can skip commercials with one click 95% of the time, and jump arbitrary numbers of minutes when that doesn't work. The files are mostly just standard mpg files so I can do what I want with them easily enough (ok - minor rant - what possessed them to write their own player instead of just using an established one as the core? It is WAY to finicky about input file formats...).
Sure, it isn't perfect, but it beats a green screen of death with no easy way to recover and resorting to all kinds of shifty websites to download images.
Yup - same at work - less than 1/20th the quota of gmail and we keep hearing about how disk usage is doubling every couple of years...
It is just being penny-wise and pound-foolish, but there is legal value to not keeping stuff around forever...
Perhaps it is a staff meeting, or a training session? There are legitimate business purposes for inviting 20 people to a meeting...
And yes - a more typical case might be 5 people. That doesn't really change the value of Exchange.
And for the record I use Thunderbird at home for most of my mail over IMAP. It just would never fly in a corporate setting...