Two of the pairs are simply unused. There were some gigabit ethernet standards that used all four pairs, but I'm not sure if that is the standard that is in common use today.
There was also a really neat 100Mbps standard called 100BASE-T4 which used 4 pairs of Cat3: 2 pairs switching directions (half duplex only) and 1 pair each dedicated TX/RX. It also used a trinary (3 state) carrier wave. I believe it's also covered under RFC 1925 section 2.3, "With sufficient thrust, pigs fly just fine."
You do realize ethernet originally ran over coax, right? Google '10BASE2'
It also ran at 10Mbps at half duplex, with a maximum utilization of about 35% - delivering a best-case rate of 3.5mbps (~350kBps) for a lightly-loaded network.
WiFi is arguably faster. Heck, his DSL is arguably faster.
Interesting that your post ends with "YMMV". If you see the other response that I posted about my printer problem, MMDV (My Mileage DID Vary!) In short, Linux handles my printer fine, and Windows refuses to talk to it without a resource-intensive application which tries to circumvent the normal printing process.
My personal interpretation of the reason behind this is that someone at MS decided to create a revenue center behind "Certified for Windows". The unintended consequence is that, rather than pay MS for the privilege of having hardware "just work", some manufacturers are refusing to accept things "the MS" way, and are releasing drivers not merely non-certified, but encumbered with "value added" software whose main claim to value, as far as I can tell, is forcing users to operate in a way inconsistent with the normal Windows method, not to mention the overall Windows user experience. Rather than assuring a quality customer experience, MS has almost architected the opposite effect.
This isn't just "fringe" players. My printer is a Brother. My fiancee's netbook is a Toshiba - and it came loaded with crapware for setting up WiFi, Bluetooth, etc. The UI is "high concept" which means it makes sense only after you puzzle out the metaphor that the designed thought you'd intuitively understand. At least it had less crapware than came on my Dell, most of which was trial versions of software I'd never use for free, much less pay for, and most of which was tedious to remove.
It's ironic that when I pay for Windows, it comes laden with invasive pseudo-advertising. Linux, which is free, doesn't.
But I digress from your response, which, if I read it right, boils down to "manufacturers can't write drivers for Linux unless they release the source code, which they would be stupid to do, so I can't guarantee that any given piece of hardware will work with any given Linux release."
Here's the problem: you're arguing that MS has provided end users with quality control through their "Buy a sticker from MS" certification program. However, their stable ABI allows vendors to bypass that quality control and hijack my user experience. You further argue that the Linux quality control procedure, namely releasing source code, is bad for business, so vendors should have a similar option of bypassing linux's quality control.
The discussion exposes 3 types of hardware vendor: 1) provide true enhanced value through better hardware. 2) claim to provide enhanced value by shipping common hardware with non-standard software. 3) ship the same thing as everyone else.
Type 1 vendors shouldn't have to worry about releasing source code, since they provide value through hardware, so having their software copied doesn't hurt them.
Type 2 vendors don't really add value to my conmputing experience. They are software vendors who are hiding as hardware vendors. Their software often has as UI which can most favorably be classified as "experimental". If there is a standard way in the OS to use the hardware, I will spend time trying to remove their non-standard software. This software often runs only on Windows, but thankfully the commodity hardware often already has driver support in Linux.
Type 3 vendors admit they're selling commodity hardware, which is cheap enough and common enough that there has been Linux support for quite some time.
Who loses in the Linux model? Vendor type 2, the "value adder" / "crapware pusher".
What about in the Windows model? Windows, lacking a native source to access drivers except through vendor support (or by vendors paying MS), requires that vendors include drivers with their hardware. Vendor type 3 therefore has to source a driver, often from some 3rd party, reducing quality control, adding cost, decreasing actual value. It is my hypothesis that this software outsourcing is what originally bred vendor type 2. In the Windows model, the party who loses is the end user, whose choice is indirectly paying MS more money (via inceased vendor
My fiancee just had me print a document for her, because she hasn't figured out how to add the printer to her XP netbook. (No CD drive) OTOH, adding that same printer to my Linux netbook was quick & painless.
Adding this printer in Windows requires either:
a) installing the "enhanced" software that came with the printer, or
b) downloading an installer for the same bloated crapware from the vendor's website - if you can find it. If I want to download just the driver, that's not an option.
The infrastructure built into XP via Windows Update implies that the printer manufacturer could register the driver with MS and make it magically work, but oddly enough, manufacturers seem to prefer not to do this. Apparently, it's easier for them to ship dubious 3rd-party crapware than it is to get MS to host their driver.
("Crapware" in this case is defined as software which adds several seconds to my boot time, takes over 5% of my system memory, and _requires_ me to perform actions in a manner inconsistent with the standard ways Windows would do it for any other vendor. Extra credit if it crashes.)
In short, it is interesting to me that, going with Linux, I had less user effort, a more consistent user interface, and a more stable / faster system.
YMMV, this is anecdotal evidence, etc... but so is a lot of the MS-provided FUD. I bought the printer from Fry's (best price per feature - and it's Brother, which arguably isn't obscure), my fiancee's netbook from Best Buy, and my netbook from t3h internet.
You're right that drivers don't belong in the kernel. At the very least they belong in loadable modules. That way, if the driver's being flaky, it can be unloaded and a different driver loaded, just like in... huh, I guess Windows doesn't do that without a reboot.
Kidding aside, how does Windows address the critique you have of Linux in a way that provides more value to the end user or IT administrator?
I miss Robert L. Forward
on
Tetraktys
·
· Score: 1
Sci-fi written by a JPL scientist, well-written, with compelling stories and characters... and knowing that all of it, however fantastic it sounded (e.g. life on a neutron star), was completely plausible...
If they can compress your health and your life down into a dollar amount don't think they can't do it in this case.
By that measure, copyright is vastly more valuable than human health and life.
Since when has human life helped corporations make money?
The average corporation makes more money than the average individual. The average corporation can then afford to (through lobbying, etc) influence laws to protect the ability of corporations to make money - whereas the average individual has little influence to protect his/her ability to live, especially when that individual's life comes in conflict with the interests of a corporation.
There are a couple of commercial products which will handle the job. I'm most familiar with the Barracuda IM Firewall. For about $2k, you'll get everything you've listed - full logging of conversations and file xfers, plenty of capacity, integrated client, plus a few other nice features like keyword administrator notification & message blocking, LDAP integration, and reporting.
The biggest feature you might appreciate is its ability to BLOCK the public IM protocols. The larger models also connect to the public IM networks, so you can log & apply policy to those conversations on a per-user basis. Some people _insist_ on bypassing IT policies, so allowing those folks to connect in a way you control might make both you and them happier.
The factors I think need to be weighed are 1) the cost of your time 2) the cost of a HIPPA violation, and 3) your ability to set up something bulletproof (no offense intended - I wouldn't trust myself to do it right the first time!)
Disclaimer: I used to work for Barracuda a couple of years ago. Some of their technology is crap, but the IM firewall is IMHO one of the best things they've ever released.
Xargs is much more fun with complex data processing[...]
It's especially fun when 'find' finds nothing, and returns that to xargs, which happily applies the given command to the nothing it was given... I lost/var that way once.
But it doesn't REALLY get fun until you write shell scripts that are supposed to work on both Linux and Solaris... because one has a 'find' which smart-splits the list like 'xargs', and the other has a 'xargs' that won't do something to nothing... 2 complete, yet completely incompatible, solutions.
Ah, the 40% ideal is far more ideal than I'd realized. Is this specifically for the Otto cycle? (Reciprocating pistons, Wankel, etc.) Your explaination is phrased in the second law, which implies that even "weird" stuff like the Atkinson cycle can't bend the rules (third law) - although I found some numbers on Wikipedia that the Brayton (Joule) cycle could achieve up to 60%. There may be different definitions of "efficiency" in use.
The other question is whether heat exchange engines could be used to capture the entropic "waste" heat. Just as turbochargers change the power game by tweaking the input, it may be possible to change the efficiency game by harnessing the output. (See the other branch in the comment tree re: BMW using a secondardy steam engine.) I know I've got a Carnot engine stashed around here somewhere...
My apologies for the accidental use of the terms "potential" and "energy" congruently: I meant (potential (energy capture)), not ((potential energy) capture). Your answer clearly answers the larger question by addressing free energy.
Re: the SAE entrants: agreed. (Although 3169 mpg is damn impressive - 2008 winner, Laval Universite)
Also, thanks - your replies are thoughtful, patient, and polite, and provide a good chance for me to learn something new.
With all the people on this topic posting that it's an "engineering feat" to go 1000mph - tops of wheels at Mach 2.8 relative to the surrounding air, rims subject to 173 G of force - your response is to say that 1000mpg is "impossible" - and OT?
I appreciate that your math makes certain generous assumptions, e.g. cd=.1, A=1m^2, and "maximum" efficiency for an internal combustion engine burning gasoline. I will also admit that your knowledge of the physics here is better than mine. However, I wonder if there are hidden assumptions that you are not factoring. I say this, knowing that the maximum theoretical speed of a modem is 9600 baud, but higher bps are possible, and knowing that the death of Moore's Law has been predicted for the last several years - both of which predictions are based in seemingly sound physics calculations.
I would also like to point out that the Bugatti Veyron's 1000+ HP apparently is also physically impossible in less than a 16L engine (@ 6000RPM), unless some game-changing technology is added.
Two other responses to your comment have already pointed out that a) the SAE Supermileage contest had several entrants whose mileage exceeded 1000mpg, and b) the methods you provide in your calculations may not be inclusive of all potential energy capture mechanisms. In short, YMMV. (Sorry, couldn't resist:-)
Reading this discussion, it seems to me that the 1000mpg problem requires much more subtlety than the 1000mph problem, and is, to me, the far more interesting one.
I do hope that Slashdot won't degrade to a point where 1000mph = exciting, 1000mpg = off-topic.
The problem with Gmail is that spam isn't the only problem.
The "perfect security" solution - apart from no Internet access - is never to let the kids on the Internet without a parent right there, actively involved.
So... is there a tool which lets the parents who can't PHYSICALLY be there still provide oversight?
Example: a few years ago, when MS was trying to get people to pay for the MSN service, there was a parental oversight feature which created a dynamic web whitelist: the kid wants to go to a new site, the service IMs the parent, the parent checks out the site, and approves or blocks. I would LOVE to find this kind of feature, as it would provide a sense of self-sufficiency to my kids, while still allowing me a measure of oversight, when I can't physically be there.
From an email perspective, I would want the ability to read every message my kids receive, before they receive it, plus the ability to read every email they write, before it's sent. I see it as comparable to making them keep the door to their rooms open whenever they have friends over - as a parent, it is my responsibility to monitor their actions.
The last sentence of TFA sums up the non-usefulness of the result: "Of course, you have to be judicious in your choice of the data to add to your data set."
I refer you to the question of training Bayesian data sets for anti-spam: should you classify every single email, or only the ones that are "clearly" well-defined? Without a good algorithm to extract the search terms, the additional data just poisons the data sets, reducing the effectiveness of the filter.
See also any decent physiological study, in which "extraneous" factors are "corrected". Without enough data pruning, you have a correlation like the study that showed that losing weight, and keeping it off, reduces life expectancy. They didn't correct for the terminally ill, who lost weight as a result of their conditions. However, do too much pruning, and you have the controversial Harvard study, which reached the "common sense" conclusion almost at the expense of the data.
For more examples of massaging data using a bad algorithm, see studies that demonstrate a better TCO by going Microsoft.
In short, adding additional data is no guarantee of good results. The students clearly got lucky in finding a similar data set on a well-researched topic, based on an established taxonomy rather than a murky preference rating.
Combine this in a single case with a CPU and a laser keyboard and it's the ultimate ultra-portable. On an airline tray table, take out the small box, which projects the image on the back of the seat in front of you, and the keyboard on the tray itself.
As component miniaturization continues, the limiting factors in ultraportables are increasingly the human interfaces - the screen and the keyboard.
It could be entertaining to "accidentally" tell the Chinese where it is, and LET THEM blow it up first. Then the Bush administration could pretend to be outraged... lots of fun the whole planet can enjoy.
The infamous Oregon Exploding Whale seemed like a good idea at the time, but instead coated all onlookers in rotting blubber - while still leaving significant portions of the carcass at ground zero.
Blowing something up, just because you can, doesn't make the problem go away. It leaves a lot of the original problem, PLUS a huge mess - and a big stink.
That same love of detail you mention combines with his personal wealth to ensure that, whatever stories he's trying to tell, Lucas is telling poorly.
Lucas makes a great producer, perhaps even a decent writer, but a lousy director. I can't bear to watch him squander the talent he had in the prequilogy. He spent so much time creating a "rich texture" visually that his actors came across completely flat.
The result is a distance between the audience and the movie, a decrease in satisfaction, and an increasing level of frustration. In our current cynical society, it's easy to discard Lucas as "greedy" for delivering lackluster material. Granted, it's equally easy for me to say he's "disconnected from his fanbase", or "full of himself." That doesn't mean that any of those statements are true, just easy to believe.
The real questions, IMHO, are: 1) Does he know that there are people who don't like what he's made? 2) Does he care?
This leads to a 2x2 grid where the corners are: 11: Lucas keeps trying to make something better, to make it up to us. 10: Lucas is just making more Star Wars stuff to make money. 01: Lucas truly believes that we love Star Wars and can't get enough. 00: Lucas just makes Star Wars for himself.
Anakin's Padawan? So that would have occurred... in Episode 3, after they promoted him _from_ padawan status, before he killed everybody... about 6 months?
This sounds like a short-lived series. 30 episodes? Did I read that right?
Instead of a beer, if it were a gun, would he be liable?
Regarding intent, if you are driving faster than the legal limit, but you don't know it, are you still breaking the law? (Yes)
If there are risks associated with actions, then there's liability. John Doe's can only try to shift the liability from himself to someone else. He would have to convince the court one of 2 things:
1) He was not informed of his liability, which his network provider (BU) or the file-sharing application provider should have done, so it's their fault. (No posted speed limit sign)
2) He acted in a manner so as to limit the redistribution of the files, but that some system flaw or 3rd party negated his action. (Mechanical fault causes car to exceed speed limit, or the passenger steps on the accelerator.)
Given rulings over the past few years, it's unlikely that there are holes in any EULA or in BU's Acceptible Use policy. The second one would also be very hard to prove (especially if it's not true.)
BBC has a show called Brainiac, but rather than "science for kids", it's "science for drunken college students." There was an episode where they tested the hypothesis that looking at women's breasts for 30 minutes was as aerobically stimulating as jogging. Episodes also tend to feature explosions, making it moderately comparable to Mythbusters.
The best candidate I've seen in recent years for "science for kids" was Bill Nye, Science Guy, which ran for 100 episodes during the 1990s. Although edutaining, it featured a few too many MTV-style elements for my taste. Mr. Wizard held my attention with nothing fancier than a lab full of cool stuff - does that mean I'm old?
You could try this: http://www.bartleby.com/66/40/63040.html That's an incomplete fragment from a letter, and, although similar in spirit, not the quote in question.
Slashdot Mirror
Two of the pairs are simply unused. There were some gigabit ethernet standards that used all four pairs, but I'm not sure if that is the standard that is in common use today.
There was also a really neat 100Mbps standard called 100BASE-T4 which used 4 pairs of Cat3: 2 pairs switching directions (half duplex only) and 1 pair each dedicated TX/RX. It also used a trinary (3 state) carrier wave. I believe it's also covered under RFC 1925 section 2.3, "With sufficient thrust, pigs fly just fine."
You do realize ethernet originally ran over coax, right? Google '10BASE2'
It also ran at 10Mbps at half duplex, with a maximum utilization of about 35% - delivering a best-case rate of 3.5mbps (~350kBps) for a lightly-loaded network.
WiFi is arguably faster. Heck, his DSL is arguably faster.
> What next, Duke Nukem Forever?
What, no one told you that "Forever" is how long we have to wait for it?
Interesting that your post ends with "YMMV". If you see the other response that I posted about my printer problem, MMDV (My Mileage DID Vary!) In short, Linux handles my printer fine, and Windows refuses to talk to it without a resource-intensive application which tries to circumvent the normal printing process.
My personal interpretation of the reason behind this is that someone at MS decided to create a revenue center behind "Certified for Windows". The unintended consequence is that, rather than pay MS for the privilege of having hardware "just work", some manufacturers are refusing to accept things "the MS" way, and are releasing drivers not merely non-certified, but encumbered with "value added" software whose main claim to value, as far as I can tell, is forcing users to operate in a way inconsistent with the normal Windows method, not to mention the overall Windows user experience. Rather than assuring a quality customer experience, MS has almost architected the opposite effect.
This isn't just "fringe" players. My printer is a Brother. My fiancee's netbook is a Toshiba - and it came loaded with crapware for setting up WiFi, Bluetooth, etc. The UI is "high concept" which means it makes sense only after you puzzle out the metaphor that the designed thought you'd intuitively understand. At least it had less crapware than came on my Dell, most of which was trial versions of software I'd never use for free, much less pay for, and most of which was tedious to remove.
It's ironic that when I pay for Windows, it comes laden with invasive pseudo-advertising. Linux, which is free, doesn't.
But I digress from your response, which, if I read it right, boils down to "manufacturers can't write drivers for Linux unless they release the source code, which they would be stupid to do, so I can't guarantee that any given piece of hardware will work with any given Linux release."
Here's the problem: you're arguing that MS has provided end users with quality control through their "Buy a sticker from MS" certification program. However, their stable ABI allows vendors to bypass that quality control and hijack my user experience. You further argue that the Linux quality control procedure, namely releasing source code, is bad for business, so vendors should have a similar option of bypassing linux's quality control.
The discussion exposes 3 types of hardware vendor:
1) provide true enhanced value through better hardware.
2) claim to provide enhanced value by shipping common hardware with non-standard software.
3) ship the same thing as everyone else.
Type 1 vendors shouldn't have to worry about releasing source code, since they provide value through hardware, so having their software copied doesn't hurt them.
Type 2 vendors don't really add value to my conmputing experience. They are software vendors who are hiding as hardware vendors. Their software often has as UI which can most favorably be classified as "experimental". If there is a standard way in the OS to use the hardware, I will spend time trying to remove their non-standard software. This software often runs only on Windows, but thankfully the commodity hardware often already has driver support in Linux.
Type 3 vendors admit they're selling commodity hardware, which is cheap enough and common enough that there has been Linux support for quite some time.
Who loses in the Linux model? Vendor type 2, the "value adder" / "crapware pusher".
What about in the Windows model? Windows, lacking a native source to access drivers except through vendor support (or by vendors paying MS), requires that vendors include drivers with their hardware. Vendor type 3 therefore has to source a driver, often from some 3rd party, reducing quality control, adding cost, decreasing actual value. It is my hypothesis that this software outsourcing is what originally bred vendor type 2. In the Windows model, the party who loses is the end user, whose choice is indirectly paying MS more money (via inceased vendor
Speaking of printers in Linux:
My fiancee just had me print a document for her, because she hasn't figured out how to add the printer to her XP netbook. (No CD drive) OTOH, adding that same printer to my Linux netbook was quick & painless.
Adding this printer in Windows requires either:
a) installing the "enhanced" software that came with the printer, or
b) downloading an installer for the same bloated crapware from the vendor's website - if you can find it.
If I want to download just the driver, that's not an option.
The infrastructure built into XP via Windows Update implies that the printer manufacturer could register the driver with MS and make it magically work, but oddly enough, manufacturers seem to prefer not to do this. Apparently, it's easier for them to ship dubious 3rd-party crapware than it is to get MS to host their driver.
("Crapware" in this case is defined as software which adds several seconds to my boot time, takes over 5% of my system memory, and _requires_ me to perform actions in a manner inconsistent with the standard ways Windows would do it for any other vendor. Extra credit if it crashes.)
In short, it is interesting to me that, going with Linux, I had less user effort, a more consistent user interface, and a more stable / faster system.
YMMV, this is anecdotal evidence, etc... but so is a lot of the MS-provided FUD. I bought the printer from Fry's (best price per feature - and it's Brother, which arguably isn't obscure), my fiancee's netbook from Best Buy, and my netbook from t3h internet.
You're right that drivers don't belong in the kernel. At the very least they belong in loadable modules. That way, if the driver's being flaky, it can be unloaded and a different driver loaded, just like in... huh, I guess Windows doesn't do that without a reboot.
Kidding aside, how does Windows address the critique you have of Linux in a way that provides more value to the end user or IT administrator?
Sci-fi written by a JPL scientist, well-written, with compelling stories and characters... and knowing that all of it, however fantastic it sounded (e.g. life on a neutron star), was completely plausible...
If they can compress your health and your life down into a dollar amount don't think they can't do it in this case.
By that measure, copyright is vastly more valuable than human health and life.
Since when has human life helped corporations make money?
The average corporation makes more money than the average individual. The average corporation can then afford to (through lobbying, etc) influence laws to protect the ability of corporations to make money - whereas the average individual has little influence to protect his/her ability to live, especially when that individual's life comes in conflict with the interests of a corporation.
There are a couple of commercial products which will handle the job. I'm most familiar with the Barracuda IM Firewall. For about $2k, you'll get everything you've listed - full logging of conversations and file xfers, plenty of capacity, integrated client, plus a few other nice features like keyword administrator notification & message blocking, LDAP integration, and reporting.
The biggest feature you might appreciate is its ability to BLOCK the public IM protocols. The larger models also connect to the public IM networks, so you can log & apply policy to those conversations on a per-user basis. Some people _insist_ on bypassing IT policies, so allowing those folks to connect in a way you control might make both you and them happier.
The factors I think need to be weighed are 1) the cost of your time 2) the cost of a HIPPA violation, and 3) your ability to set up something bulletproof (no offense intended - I wouldn't trust myself to do it right the first time!)
Disclaimer: I used to work for Barracuda a couple of years ago. Some of their technology is crap, but the IM firewall is IMHO one of the best things they've ever released.
Xargs is much more fun with complex data processing[...]
It's especially fun when 'find' finds nothing, and returns that to xargs, which happily applies the given command to the nothing it was given... I lost /var that way once.
But it doesn't REALLY get fun until you write shell scripts that are supposed to work on both Linux and Solaris... because one has a 'find' which smart-splits the list like 'xargs', and the other has a 'xargs' that won't do something to nothing... 2 complete, yet completely incompatible, solutions.
What's fun is adding my admin account's ssh public key to root's authorized-keys, then aliasing something to:
ssh -X -f -l root localhost [favorite terminal]
Instant su, in a new terminal window, without using su. Who needs root passwords when you have keys?
Ah, the 40% ideal is far more ideal than I'd realized. Is this specifically for the Otto cycle? (Reciprocating pistons, Wankel, etc.) Your explaination is phrased in the second law, which implies that even "weird" stuff like the Atkinson cycle can't bend the rules (third law) - although I found some numbers on Wikipedia that the Brayton (Joule) cycle could achieve up to 60%. There may be different definitions of "efficiency" in use.
The other question is whether heat exchange engines could be used to capture the entropic "waste" heat. Just as turbochargers change the power game by tweaking the input, it may be possible to change the efficiency game by harnessing the output. (See the other branch in the comment tree re: BMW using a secondardy steam engine.) I know I've got a Carnot engine stashed around here somewhere...
My apologies for the accidental use of the terms "potential" and "energy" congruently: I meant (potential (energy capture)), not ((potential energy) capture). Your answer clearly answers the larger question by addressing free energy.
Re: the SAE entrants: agreed. (Although 3169 mpg is damn impressive - 2008 winner, Laval Universite)
Also, thanks - your replies are thoughtful, patient, and polite, and provide a good chance for me to learn something new.
With all the people on this topic posting that it's an "engineering feat" to go 1000mph - tops of wheels at Mach 2.8 relative to the surrounding air, rims subject to 173 G of force - your response is to say that 1000mpg is "impossible" - and OT?
:-)
I appreciate that your math makes certain generous assumptions, e.g. cd=.1, A=1m^2, and "maximum" efficiency for an internal combustion engine burning gasoline. I will also admit that your knowledge of the physics here is better than mine. However, I wonder if there are hidden assumptions that you are not factoring. I say this, knowing that the maximum theoretical speed of a modem is 9600 baud, but higher bps are possible, and knowing that the death of Moore's Law has been predicted for the last several years - both of which predictions are based in seemingly sound physics calculations.
I would also like to point out that the Bugatti Veyron's 1000+ HP apparently is also physically impossible in less than a 16L engine (@ 6000RPM), unless some game-changing technology is added.
Two other responses to your comment have already pointed out that a) the SAE Supermileage contest had several entrants whose mileage exceeded 1000mpg, and b) the methods you provide in your calculations may not be inclusive of all potential energy capture mechanisms. In short, YMMV. (Sorry, couldn't resist
Reading this discussion, it seems to me that the 1000mpg problem requires much more subtlety than the 1000mph problem, and is, to me, the far more interesting one.
I do hope that Slashdot won't degrade to a point where 1000mph = exciting, 1000mpg = off-topic.
The problem with Gmail is that spam isn't the only problem.
The "perfect security" solution - apart from no Internet access - is never to let the kids on the Internet without a parent right there, actively involved.
So... is there a tool which lets the parents who can't PHYSICALLY be there still provide oversight?
Example: a few years ago, when MS was trying to get people to pay for the MSN service, there was a parental oversight feature which created a dynamic web whitelist: the kid wants to go to a new site, the service IMs the parent, the parent checks out the site, and approves or blocks. I would LOVE to find this kind of feature, as it would provide a sense of self-sufficiency to my kids, while still allowing me a measure of oversight, when I can't physically be there.
From an email perspective, I would want the ability to read every message my kids receive, before they receive it, plus the ability to read every email they write, before it's sent. I see it as comparable to making them keep the door to their rooms open whenever they have friends over - as a parent, it is my responsibility to monitor their actions.
The last sentence of TFA sums up the non-usefulness of the result: "Of course, you have to be judicious in your choice of the data to add to your data set."
I refer you to the question of training Bayesian data sets for anti-spam: should you classify every single email, or only the ones that are "clearly" well-defined? Without a good algorithm to extract the search terms, the additional data just poisons the data sets, reducing the effectiveness of the filter.
See also any decent physiological study, in which "extraneous" factors are "corrected". Without enough data pruning, you have a correlation like the study that showed that losing weight, and keeping it off, reduces life expectancy. They didn't correct for the terminally ill, who lost weight as a result of their conditions. However, do too much pruning, and you have the controversial Harvard study, which reached the "common sense" conclusion almost at the expense of the data.
For more examples of massaging data using a bad algorithm, see studies that demonstrate a better TCO by going Microsoft.
In short, adding additional data is no guarantee of good results. The students clearly got lucky in finding a similar data set on a well-researched topic, based on an established taxonomy rather than a murky preference rating.
Combine this in a single case with a CPU and a laser keyboard and it's the ultimate ultra-portable. On an airline tray table, take out the small box, which projects the image on the back of the seat in front of you, and the keyboard on the tray itself.
As component miniaturization continues, the limiting factors in ultraportables are increasingly the human interfaces - the screen and the keyboard.
It could be entertaining to "accidentally" tell the Chinese where it is, and LET THEM blow it up first. Then the Bush administration could pretend to be outraged... lots of fun the whole planet can enjoy.
The infamous Oregon Exploding Whale seemed like a good idea at the time, but instead coated all onlookers in rotting blubber - while still leaving significant portions of the carcass at ground zero.
Blowing something up, just because you can, doesn't make the problem go away. It leaves a lot of the original problem, PLUS a huge mess - and a big stink.
Ok, that makes a lot more sense, thanks.
That same love of detail you mention combines with his personal wealth to ensure that, whatever stories he's trying to tell, Lucas is telling poorly.
Lucas makes a great producer, perhaps even a decent writer, but a lousy director. I can't bear to watch him squander the talent he had in the prequilogy. He spent so much time creating a "rich texture" visually that his actors came across completely flat.
The result is a distance between the audience and the movie, a decrease in satisfaction, and an increasing level of frustration. In our current cynical society, it's easy to discard Lucas as "greedy" for delivering lackluster material. Granted, it's equally easy for me to say he's "disconnected from his fanbase", or "full of himself." That doesn't mean that any of those statements are true, just easy to believe.
The real questions, IMHO, are:
1) Does he know that there are people who don't like what he's made?
2) Does he care?
This leads to a 2x2 grid where the corners are:
11: Lucas keeps trying to make something better, to make it up to us.
10: Lucas is just making more Star Wars stuff to make money.
01: Lucas truly believes that we love Star Wars and can't get enough.
00: Lucas just makes Star Wars for himself.
Any of which I'm willing to believe.
Anakin's Padawan? So that would have occurred... in Episode 3, after they promoted him _from_ padawan status, before he killed everybody... about 6 months? This sounds like a short-lived series. 30 episodes? Did I read that right?
ob.IANAL
Instead of a beer, if it were a gun, would he be liable?
Regarding intent, if you are driving faster than the legal limit, but you don't know it, are you still breaking the law? (Yes)
If there are risks associated with actions, then there's liability. John Doe's can only try to shift the liability from himself to someone else. He would have to convince the court one of 2 things:
1) He was not informed of his liability, which his network provider (BU) or the file-sharing application provider should have done, so it's their fault. (No posted speed limit sign)
2) He acted in a manner so as to limit the redistribution of the files, but that some system flaw or 3rd party negated his action. (Mechanical fault causes car to exceed speed limit, or the passenger steps on the accelerator.)
Given rulings over the past few years, it's unlikely that there are holes in any EULA or in BU's Acceptible Use policy. The second one would also be very hard to prove (especially if it's not true.)
BBC has a show called Brainiac, but rather than "science for kids", it's "science for drunken college students." There was an episode where they tested the hypothesis that looking at women's breasts for 30 minutes was as aerobically stimulating as jogging. Episodes also tend to feature explosions, making it moderately comparable to Mythbusters.
The best candidate I've seen in recent years for "science for kids" was Bill Nye, Science Guy, which ran for 100 episodes during the 1990s. Although edutaining, it featured a few too many MTV-style elements for my taste. Mr. Wizard held my attention with nothing fancier than a lab full of cool stuff - does that mean I'm old?
You're absolutely right, one should always check one's sources.
i ct/?view=uk
e d
Start here: http://www.askoxford.com/dictionaries/quotation_d
Search for "Voltaire" and see result #1.
You could try this: http://www.bartleby.com/66/40/63040.html
That's an incomplete fragment from a letter, and, although similar in spirit, not the quote in question.
Anything Wiki is a questionable authority, but even so, here's one on quotes, which was your proposed criterion:
http://en.wikiquote.org/wiki/Voltaire#Misattribut
More details from the same site:
http://en.wikiquote.org/wiki/Evelyn_Beatrice_Hall
I'd be interested to know of any trustworthy sources which attribute the actual quote to Voltaire.
The proofs are still being developed - the bridge appears much nicer on film than via digital media.