That's only the first 10 of the 100, because Slashdot won't let me post the full list. Of the full hundred, only 13 numbers are represented exactly 1% of the 1000, 41 are less than 1%, and 46 are more. The maximum was 1.7% and the minimum was 0.3%.
Why not simply use POPFile? It has a very nice web interface that makes it very easy to reclassify false positives or false negatives. It also supports multiple email categories; unlike most other Bayesian filters, you can filter email as more than just "spam" or "not-spam". I myself have five classification groups: spam, mail from my university, two mailing lists I'm on but don't read actively, and "real" email. It works very well in concert with Mozilla, and my email is automatically directed to the appropriate folder based on POPFile's extra headers.
Seems to me like it isn't an artificial constraint, but merely a practical one. It sounds like he scripted the programs to run through his data all at once, so querying the online resources a thousand times an hour would not be feasible. The Bayesian filters were at a similar disadvantage because of the automated testing: normally, each false negative gets added to the spam corpus, which would haved improved their accuracy over time.
Later on, he criticized sa-learn's manual for indicating that you should have an even ratio of spam to "ham", which was not born out by his tests. Obviously, he did read the manual.
The licenses support the BBC, no? And you can watch the BBC with your TV card, no? So you should pay for watching the BBC, yes?
Otherwise you get PBS, which interrupts every 5 minutes to remind you that if you don't donate and receive this nifty travel mug, you're a scum-sucking thief.
All of the sweeping generalizations you make about state universities are exactly untrue at mine*. Nearly all of my professors at least know me by name and if they don't, it's because I haven't approached them. Even a little too often for my liking, they know who I am despite my best efforts. Among my worst professors are the incredibly bright and incredibly nice type who simply can't get thoughts out of their head fast enough. All of my professors though, are very much willing to go beyond the minimum requirements to help you understand the material.
As much noise as US News makes with these ratings, what's really important is choosing the right school for your major. When I got my admissions responses back, I had cheaper options, and I had more prestigious options, but Cal Poly won out because it's the right education at the right price. (Though recent hikes are pissing me off..)
* California Polytechnic State University, San Luis Obispo: ranked #1 "public largely undergraduate university in the West" for the 11th year in a row, which is a lot of hogwash. The more qualifiers you add, the less impressive that #1 becomes. But it really is a very good school, and I can't imagine where I'd rather go.
Alright, you can excuse me if I wasn't quite paying attention to the thread I was replying to, but software that acts just as I described is mentioned in several other comments. Obviously, port scanning on a legitimate network is more devious than simply copping out pirates, but my view holds true: If I don't want it, it shouldn't be there.
Whether or not it does anything in a legitimate installation, it's still there and checking each time you run the program. If for no other reason than it's an unnecessary waste of CPU cycles (assuming I'm using a legitimate copy, it should always return the same result) I don't want it running on my computer.
A Lojak system as you described (I suppose -- honestly I've never heard of it) would be something I want as a customer, because it benefits me. This extra "feature" benefits only the software company, and that is where your analogy breaks down. A more apt analogy would be a system which reports to the manufacturer if you install stolen car parts. Then again, the only truly apt analogy is a piece of software which reports home if it detects that it is stolen.
I'd say that it's certainly proper for the software publisher to do some sort of check to be sure that the software is properly licensed. What they do beyond that, though, is probably a matter best attended to by their legal staff.
My computer should never do anything I don't want it to do. Plain and simple. If I don't want you to scan my network for illicit copies, then don't do it. I don't really care about any legal "right" software companies have to do it. I don't want them to do it; I'm their customer; they shouldn't do it unless they feel like pissing me off and losing me as a customer.
Anyway, any cracker with enough skill to remove normal copy protection techniques (not that much) can remove this sort of protection too.
I'm gonna put on my psychic cap on right now and make two predictions:
Dean will win the Democratic primary. Dean will lose the general election.
But then, the democratic race has always been a race to find out who is going to lose to Bush. The country has moved frightenly to the right in the past few years, and despite how many fucked up things Bush does, he's still popular. IMO, Dean has the best chance to win, but it's still not enough.
First and foremost, the democratic base likes to see someone with a backbone come out against Bush. Someone who didn't belly over after September 11th and vote for all these horrendous laws. Secondly, you talk about "Middle America", filled with people that you presume would never vote for an anti-war candidate. That may be true, but they'd never vote for a pro-war Democrat either. Why vote for Kerry or Lieberman instead of Bush, when all they do is talk about how great Bush is?
It is not sufficient to simply mimick your opponent. The Coke vs. Pepsi argument has lasted for decades because they're different and they appeal to different people. If one simply copied the other, they wouldn't be around for very long. Kerry and especially Liberman are trying very hard to be the Shasta Colas of the world. Dean, on the other hand, is more like 7-Up: the un-cola.
By making things easier for buyers and sellers, they've made it easier for scammers too.
The thing is, there are some very simple steps eBay could take to make it easy for legitimate buyers and sellers yet hard for scammers. For example, a "report suspected fraud" button, as mentioned previously.
It should be pointed out that the Volts are actually re-branded iRiver players. iRiver is a Korean corporation that had been manufacturing cheap CD players under a variety of names until Rio sold their player as the Rio Volt and actually made it successful. Now iRiver is selling the players under its own name and undercutting Rio's prices.
In Statistics 101, you hear about the "sample size". The sample size relates to the confidence interval, which is a range of values you expect the actual value to be. In this case, the actual value is the failure rate of all IBM or Company X drives, not just the ones you tested. With a hundred samples, your confidence interval is going to very small, meaning that your estimate is probably very close to the actual value. With only two samples, the confidence interval is going to be very wide, because you can't be sure it's not a fluke.
When you bring a time-scale into it, you either have to measure the failure rate over the same time period, or measure the mean time between failure (MTBF), which is more informative. Technically, in the scenario given, 100% of Company X drives did fail within the same period that IBM's drives were being tested. If they fail in the first five minutes or in the last five minutes of those five years, they still failed. But again, with a sample size that low, you can't use that to reach any kind of conclusion.
However, if you measure the MTBF, even with a small sample size, I think you can say with reasonable confidence that the MTBF of the IBM drives is higher than Company X's, simply because the wide confidence interval (the region of uncertainty) of Company X's MTBF still wouldn't quite make it out to five years.
Of course, none of this takes into consideration the fact that IBM's drives aren't made by IBM anymore, and that (from the words of an ex-IBM engineer) Hitachi doesn't know how to make hard drives worth shit.
Whenever the company's creditors figure it out, any open source project with the slightest bit of code is wearing a big fat sign that says "Sue Me!" For a non-existant company with no revenue, the potential pay-off from a suit could quickly cause them to re-appear from nowhere. No honest coder could touch the stuff, leaving the code to closed-source projects. They'd likely be fly-by-night companies that wouldn't maintain it properly anyway, not exactly the kind of adoptive parents you'd hope for.
Ignore Wal-Mart. Go to Costco. A 24-pack for $3 comes to 2 sodas for a quarter. If you want more variety, a 6-pack typically goes for $1, so 16.7 cents a soda.
No one said any of these prices are for Coke or Pepsi products, mind you.
Maybe the encryption you speak of would cause problems, but the way I was thinking it would let "hits" through, and only come into action when the box requests permission to view restricted channels.
... it'll call the head end to see if you've 'subscribed' to the channel, and will be polled regurally to see if you've ordered any Pay Per Views.
Now, I have no idea what I'm talking about, but if it prompts the cable company to see if you're subscribed, would it be terribly difficult to sit an intelligent box in between to trap those requests and say "yes" every time? This is along the lines of some of the more elaborate anti-piracy techniques in video games. Q3 and Black and White both reported in with home base, but both were compromised quickly.
Like I said, I know nothing about this, but it seems to me that these systems could be broken fairly easily if you had the right hardware and knew what you were doing, unlike me.
They did the same thing during the world series if Barry Bonds managed to hit a home run into a target in the San Francisco Bay. Pac Bell park is right on the water, but the target was pretty far out.
Uhm, that seems a lot like the default username for Kazaa lite, which would be the same for hundreds of lazy users. Same goes for "Jim", "Jeff" and "Jessica". Good luck proving which jim@Kazaa was sharing copyrighted music.
The only times I ever saw BSODs in 9x were usually during hardware swaps, if then. One headless 98se box I had was up continuously for 80 days, and died only then because the power went out. Yes, it is easy to crash a Win9x box, mainly because any random app or faulty driver can take down everything and royally screw things up. But honestly, if you're not stupid, you learn quickly what causes instability, and you can get rid of it. (Insert crack about Windows/RIAA/Saddam Hussein.)
The stability of a system has more to do with the person running it than the OS it's running. Give a Linux machine to some people, and not only could they find ingenious and theoretically impossible ways to crash it, they'd also find a way to install Gator, Bonzai Buddy and that stupid "Pop-Up Blocker" program that creates more pop-ups.
Hmm.. says the algorithm used takes about 40 moves to solve, and at 30 seconds per move, that's 20 minutes to solve a cube. I can solve the cube in 3-5 minutes; I think I'd just get bored and rip the cube out to pass the time.
Slashdot Mirror
I don't understand why D can't peck F, or why F can peck D. Was this particular to the arrangement of chickens on the chart you refer to?
Using Excel's pseudo-random number generator to generate 1000 random numbers from 1 to 100:
1 0.5%
2 1.3%
3 1.2%
4 1.5%
5 0.9%
6 0.6%
7 0.6%
8 0.7%
9 1.3%
10 1.5%
That's only the first 10 of the 100, because Slashdot won't let me post the full list. Of the full hundred, only 13 numbers are represented exactly 1% of the 1000, 41 are less than 1%, and 46 are more. The maximum was 1.7% and the minimum was 0.3%.
Why not simply use POPFile? It has a very nice web interface that makes it very easy to reclassify false positives or false negatives. It also supports multiple email categories; unlike most other Bayesian filters, you can filter email as more than just "spam" or "not-spam". I myself have five classification groups: spam, mail from my university, two mailing lists I'm on but don't read actively, and "real" email. It works very well in concert with Mozilla, and my email is automatically directed to the appropriate folder based on POPFile's extra headers.
Seems to me like it isn't an artificial constraint, but merely a practical one. It sounds like he scripted the programs to run through his data all at once, so querying the online resources a thousand times an hour would not be feasible. The Bayesian filters were at a similar disadvantage because of the automated testing: normally, each false negative gets added to the spam corpus, which would haved improved their accuracy over time.
Later on, he criticized sa-learn's manual for indicating that you should have an even ratio of spam to "ham", which was not born out by his tests. Obviously, he did read the manual.
The licenses support the BBC, no? And you can watch the BBC with your TV card, no? So you should pay for watching the BBC, yes?
Otherwise you get PBS, which interrupts every 5 minutes to remind you that if you don't donate and receive this nifty travel mug, you're a scum-sucking thief.
All of the sweeping generalizations you make about state universities are exactly untrue at mine*. Nearly all of my professors at least know me by name and if they don't, it's because I haven't approached them. Even a little too often for my liking, they know who I am despite my best efforts. Among my worst professors are the incredibly bright and incredibly nice type who simply can't get thoughts out of their head fast enough. All of my professors though, are very much willing to go beyond the minimum requirements to help you understand the material.
As much noise as US News makes with these ratings, what's really important is choosing the right school for your major. When I got my admissions responses back, I had cheaper options, and I had more prestigious options, but Cal Poly won out because it's the right education at the right price. (Though recent hikes are pissing me off..)
* California Polytechnic State University, San Luis Obispo: ranked #1 "public largely undergraduate university in the West" for the 11th year in a row, which is a lot of hogwash. The more qualifiers you add, the less impressive that #1 becomes. But it really is a very good school, and I can't imagine where I'd rather go.
GOD DAMN that was the funniest thing I've ever read on Slashdot. Almost makes me wish I didn't give up my mod priviledges... almost.
Alright, you can excuse me if I wasn't quite paying attention to the thread I was replying to, but software that acts just as I described is mentioned in several other comments. Obviously, port scanning on a legitimate network is more devious than simply copping out pirates, but my view holds true: If I don't want it, it shouldn't be there.
Whether or not it does anything in a legitimate installation, it's still there and checking each time you run the program. If for no other reason than it's an unnecessary waste of CPU cycles (assuming I'm using a legitimate copy, it should always return the same result) I don't want it running on my computer.
A Lojak system as you described (I suppose -- honestly I've never heard of it) would be something I want as a customer, because it benefits me. This extra "feature" benefits only the software company, and that is where your analogy breaks down. A more apt analogy would be a system which reports to the manufacturer if you install stolen car parts. Then again, the only truly apt analogy is a piece of software which reports home if it detects that it is stolen.
I'd say that it's certainly proper for the software publisher to do some sort of check to be sure that the software is properly licensed. What they do beyond that, though, is probably a matter best attended to by their legal staff.
My computer should never do anything I don't want it to do. Plain and simple. If I don't want you to scan my network for illicit copies, then don't do it. I don't really care about any legal "right" software companies have to do it. I don't want them to do it; I'm their customer; they shouldn't do it unless they feel like pissing me off and losing me as a customer.
Anyway, any cracker with enough skill to remove normal copy protection techniques (not that much) can remove this sort of protection too.
I'm gonna put on my psychic cap on right now and make two predictions:
Dean will win the Democratic primary. Dean will lose the general election.
But then, the democratic race has always been a race to find out who is going to lose to Bush. The country has moved frightenly to the right in the past few years, and despite how many fucked up things Bush does, he's still popular. IMO, Dean has the best chance to win, but it's still not enough.
First and foremost, the democratic base likes to see someone with a backbone come out against Bush. Someone who didn't belly over after September 11th and vote for all these horrendous laws. Secondly, you talk about "Middle America", filled with people that you presume would never vote for an anti-war candidate. That may be true, but they'd never vote for a pro-war Democrat either. Why vote for Kerry or Lieberman instead of Bush, when all they do is talk about how great Bush is?
It is not sufficient to simply mimick your opponent. The Coke vs. Pepsi argument has lasted for decades because they're different and they appeal to different people. If one simply copied the other, they wouldn't be around for very long. Kerry and especially Liberman are trying very hard to be the Shasta Colas of the world. Dean, on the other hand, is more like 7-Up: the un-cola.
By making things easier for buyers and sellers, they've made it easier for scammers too.
The thing is, there are some very simple steps eBay could take to make it easy for legitimate buyers and sellers yet hard for scammers. For example, a "report suspected fraud" button, as mentioned previously.
My semi-exhaustive collection of Iceman resources
... has a boat-load of broken links.
It should be pointed out that the Volts are actually re-branded iRiver players. iRiver is a Korean corporation that had been manufacturing cheap CD players under a variety of names until Rio sold their player as the Rio Volt and actually made it successful. Now iRiver is selling the players under its own name and undercutting Rio's prices.
In Statistics 101, you hear about the "sample size". The sample size relates to the confidence interval, which is a range of values you expect the actual value to be. In this case, the actual value is the failure rate of all IBM or Company X drives, not just the ones you tested. With a hundred samples, your confidence interval is going to very small, meaning that your estimate is probably very close to the actual value. With only two samples, the confidence interval is going to be very wide, because you can't be sure it's not a fluke.
When you bring a time-scale into it, you either have to measure the failure rate over the same time period, or measure the mean time between failure (MTBF), which is more informative. Technically, in the scenario given, 100% of Company X drives did fail within the same period that IBM's drives were being tested. If they fail in the first five minutes or in the last five minutes of those five years, they still failed. But again, with a sample size that low, you can't use that to reach any kind of conclusion.
However, if you measure the MTBF, even with a small sample size, I think you can say with reasonable confidence that the MTBF of the IBM drives is higher than Company X's, simply because the wide confidence interval (the region of uncertainty) of Company X's MTBF still wouldn't quite make it out to five years.
Of course, none of this takes into consideration the fact that IBM's drives aren't made by IBM anymore, and that (from the words of an ex-IBM engineer) Hitachi doesn't know how to make hard drives worth shit.
That's really naive. Why do we still rely on cars? They're over 100 years old; haven't we moved to something more modern?
I completely agree. Doesn't it bother anyone else that our world is built around a single mode of transportation?
Battery, antenna, and radio and you're "online".
The only noticeable difference is the absence of a pre-existing network. How does this contradict the parent post?
Whenever the company's creditors figure it out, any open source project with the slightest bit of code is wearing a big fat sign that says "Sue Me!" For a non-existant company with no revenue, the potential pay-off from a suit could quickly cause them to re-appear from nowhere. No honest coder could touch the stuff, leaving the code to closed-source projects. They'd likely be fly-by-night companies that wouldn't maintain it properly anyway, not exactly the kind of adoptive parents you'd hope for.
Ignore Wal-Mart. Go to Costco. A 24-pack for $3 comes to 2 sodas for a quarter. If you want more variety, a 6-pack typically goes for $1, so 16.7 cents a soda.
No one said any of these prices are for Coke or Pepsi products, mind you.
Maybe the encryption you speak of would cause problems, but the way I was thinking it would let "hits" through, and only come into action when the box requests permission to view restricted channels.
... it'll call the head end to see if you've 'subscribed' to the channel, and will be polled regurally to see if you've ordered any Pay Per Views.
Now, I have no idea what I'm talking about, but if it prompts the cable company to see if you're subscribed, would it be terribly difficult to sit an intelligent box in between to trap those requests and say "yes" every time? This is along the lines of some of the more elaborate anti-piracy techniques in video games. Q3 and Black and White both reported in with home base, but both were compromised quickly.
Like I said, I know nothing about this, but it seems to me that these systems could be broken fairly easily if you had the right hardware and knew what you were doing, unlike me.
They did the same thing during the world series if Barry Bonds managed to hit a home run into a target in the San Francisco Bay. Pac Bell park is right on the water, but the target was pretty far out.
2 1/1449087.html
Link: http://msn.espn.go.com/mlb/playoffs2002/s/2002/10
www.k_lite.tk_Kazaa_Lite@Kazaa
Uhm, that seems a lot like the default username for Kazaa lite, which would be the same for hundreds of lazy users. Same goes for "Jim", "Jeff" and "Jessica". Good luck proving which jim@Kazaa was sharing copyrighted music.
The only times I ever saw BSODs in 9x were usually during hardware swaps, if then. One headless 98se box I had was up continuously for 80 days, and died only then because the power went out. Yes, it is easy to crash a Win9x box, mainly because any random app or faulty driver can take down everything and royally screw things up. But honestly, if you're not stupid, you learn quickly what causes instability, and you can get rid of it. (Insert crack about Windows/RIAA/Saddam Hussein.)
The stability of a system has more to do with the person running it than the OS it's running. Give a Linux machine to some people, and not only could they find ingenious and theoretically impossible ways to crash it, they'd also find a way to install Gator, Bonzai Buddy and that stupid "Pop-Up Blocker" program that creates more pop-ups.
Hmm.. says the algorithm used takes about 40 moves to solve, and at 30 seconds per move, that's 20 minutes to solve a cube. I can solve the cube in 3-5 minutes; I think I'd just get bored and rip the cube out to pass the time.