Not sure how serious you are being, but I think there a number of problems with your suggestion.
* IT is far too market driven to support a union. If tech people went on strike, then they'd just pay someone else to do it. When so much work is ultimately contract based, we need to delver products, or our jobs get taken by someone willing to delver (maybe not directly).
* It's illegal in most countries to be destructive. The Fire Fighter union cannot hold a strike where they go around lighting fires. Furthermore, doing damage does little to build public support.
Finally, from an ethical point of view, what's worse: writing the software that will be used to invade privacy in line with the law; or writing software that intentionally returns bad data, increasing the chances of innocent people being investigated, while also increasing the chances of them missing someone they would have caught without the software? I know which one would give me the fuzzier feeling.
But how many open source data mining projects can compete (in the features sense) with these commercial systems?
These systems are extremely specialised and targetted at law enforcement and/or large corporations with huge databases.
Seccessful OSS projects tend to be the ones that are used by the people writing them, and are of use to a wide community. If the developers do not have a vested interest in the product, then development will tend to stagnate.
I think it is hard to argue that OSS has been successful in making products that are targetted at such specific (and wealthy) groups.
You may want to read up on the DMCA, which makes it illegal to make or distribute technology that may be used to circumvent copy protection. I know making irrelevant references to DMCA has been flavour de jour on Slashdot for the past 5 years, but it doesn't have much at all to do with encryption in gerenal (only certain kinds/uses of encryption).
Since it is unlikely that anyone would consider using Enigma for any kind of copy protection or DRM, breaking Enigma would not be a problem for DMCA. Even if someone were crazy enough to use Enigma for some form of copy protection, it probably wouldn't make a difference. Most distributed projects to crack encryption schemes (which are generally used for copy-protection) are fine, because they only crack the key for a limited set of messages, so cannot directly be used to break copy-protection.
Further to that, we already know how Enigma worked. They just don't know the "key" for these particular messages. This is an effort to crack those 3 messages, not to perform a cryptographic attack on Enigma.
What seems to be my problem with e-mail is that I send a message and I run the topic into the ground in that first e-mail (saying everything about it). Now, that's written in stone like a Slashdot comment. No backsies.
I don't have that problem, because I'm always right to begin with.
String theory is not falsifiable and therefore is unlikely to stand the test of time.
It most certainly is falsifiable - we just don't quite have the technology to test it yet, but by all measures, we appear to be pretty close (hence the article).
Assuming they produce a mini-blackhole with the LHC, if the observations do not match string theory's predictions, then it will have been falsified. They then need to either throw it out, or take it back to the drawing board.
How do they define "longest flight"?
on
Global Flyer Part 2
·
· Score: 3, Interesting
The reference on that site to "Virgin Galactic" got me thinking. How do they differentiate between "flight", and "orbit"?
Do they say that the record is only available to jet-powered flight? Or do you have to be under power the whole time? Why doesn't a space agency hold this record?
The button is probably not as bad as the tap at the sink. I don't want to get too graphical here, but at least when people go for the button, they haven't wiped yet:). There's a good chance the average person's hands are still clean at that point.
If you're really paranoid, I guess you can try using your foot or something, but that's an image I don't need to think about:).
What studies to date show is a compound that attacks HIV at its molecular membrane level, disrupting the virus from interacting with their primary targets, the "T-helper" class white blood cells that comprise and direct the human immune system. Further, CSAs appear to be deadly to all known strains of HIV.
Umm - What studies? Does anyone else find it strange that there is no reference anywhere in the article as to how they came to believe that CSAs had this effect? Or am I just blind?
Was it tested in a test-tube? In monkeys? In humans?
Alright - from this day forward, I am hereby boycotting the Verizon ads. Whenever I click on the day-pass link, I am going to close my eyes for 15 seconds. During that 15 seconds, I am going to make a conscious effort not to think about what Verizon can do for my business.
No longer is it about organizing information, it's whether or not they want you listed.
Their track record says otherwise. In 2004, they came under fire for not removing an anti-semitic website, jewwatch.com, which was coming up as the first hit when searching for "jew". Even today, it is second only to Wikipedia.
Their argument at the time was that they were not going to block sites from their index based on content. According to that site that I linked, it was blocked in countries where the content of the site was illegal.
It looks to me like they will not block based on (legal) content, but will block people who fsck with PageRanks.
It's not particularly creative to implment sliding window schemes.
Of course not - the implementation itself is very rarely creative. Any monkey should be able to implement the chaining part of the solution I suggested. It is the problem solving that is creative; cross-pollinating solutions from different areas to solve a new problem, or an existing problem in a different way - That's creative. Now, take the matching part of my solution, where I've solved it with a bit of hand-waving (I haven't actually offered a solution to "fast fuzzy matching", and you're going to need a bit of creativity to determine a reasonable algorithm. Then, a code monkey can implement your solution.
There's very little 'science' in 'Computer Science' because most of it was figured out by mathematicians decades ago.
Not sure what that comment is in response to; the point of my post was that comp sci grads can do interesting stuff without staying at uni, but I'll bite anyway (I want to avoid the eternal "comp sci: science or engineering" side of things though). The reason there is very little 'science' in computer science, is not because it's all been figured out already. It is because 'computer science' is a big field; the term has been heavily overloaded in the past couple of decades.
There is ample 'applied comp sci research' going on at universities (well, the ones that focus on more than just churning out drones), and out in the real world (well, at the companies that focus on more than shipping the next product - they do exist!). The more theoretical stuff is harder to identify, because as you say it tends to be done by mathematicians, and it takes significant time to be adopted (I'm not aware of much theoretical computer science being done outside universities, either).
Industry needs to solve new problems all the time. Many of these problems can be solved by modifying an existing algorithm. Some require brand now algorithms. Whether the work be done at a university or in industry, it needs to be backed by scientific method (I agree that it often isn't, but there are good players out there). Almost no comp sci grads end up doing that kind of work, but if they want it, it *is* out there.
Yeah, it's just a shame that if you look through the replies to my post, you just see a bunch of poeple who missed my point. Everyone seemed to think that I was condoning the move, when all I was doing was accusing my parent poster of FUDing.
The/. community seems so quick to turn any issue into a dichotomy. It's a shame really.
Kind of. The good news is that the Computer Scientist dorks can go out into the real world, do stuff that lets them use their brains, and a little creativity, all while (generally) getting paid more than the drones.
It's basically a statistical clustering problem that has been solved in a number of different ways (and I alluded to in my post with the fuzzy reference). Just off the top of my head, If I were to be writing such a system, I would probably try applying plagurism detection techniques.
Take for example Markov Chains, which are frequently used to detect plagurism in Computer Science schools at universities: The basic idea is after possibly doing some basic preprocessing of the emails (perhaps you will remove punctuation), you break the emails up into bite sized chunks (probably words). Take some number of words in a row (usually 3), and hash them. Then slide a "window" along the email by a word, and generate another hash. Continue throughout the email. You end up with a list of hashes for each email.
It turns out you can do really fast fuzzy searches in a collection of these lists, especially if you don't mind getting a few false positives (which you can then rule out by directly comparing the "possibly matching" emails to get a "percentage similarity"). The lists themselves are quick to create (linear over the amount of text in the email), don't take up much space (they're usually a little smaller than the main-text of the email), and don't need to be held for long (you're only really interested in the rate that similar emails are arriving).
No this affects mailing lists, not spammers. There's a very fine line between your average mailing list and your average spammer. One man's mailing list is another man's spam.
they'll continue to send small volumes of mail If they're truely sending small volumes, then they're probably not a problem.
If there are millions of boxes each sending small numbers of the same email, then it is trivial to realise that they are the same emailer.
The basic gist is: hash every email that comes in. Count identical hashes. If the count goes up too fast, block the messages matching that hash. The only real trickery is in defining the hash to be fuzzy (to stop a few characters difference from changing the hash).
My understanding is that they have no blacklist, so it has nothing to do with reporting messages as spam. If you send the 'same' email to more than some number of AOL recipients, then you are a bulk mailer, and must currently get yourself onto their whitelist.
That's just FUD. AOL has not said that they are blocking all email that is not sent through GoodMail. They are replacing their whitelist with the service.
You don't need to be on the whitelist to send personal emails today, so you won't need to pay to send email tomorrow.
This only affects senders of bulk emails (mailing lists and spammers).
Slashdot Mirror
Not sure how serious you are being, but I think there a number of problems with your suggestion.
* IT is far too market driven to support a union. If tech people went on strike, then they'd just pay someone else to do it. When so much work is ultimately contract based, we need to delver products, or our jobs get taken by someone willing to delver (maybe not directly).
* It's illegal in most countries to be destructive. The Fire Fighter union cannot hold a strike where they go around lighting fires. Furthermore, doing damage does little to build public support.
Finally, from an ethical point of view, what's worse: writing the software that will be used to invade privacy in line with the law; or writing software that intentionally returns bad data, increasing the chances of innocent people being investigated, while also increasing the chances of them missing someone they would have caught without the software? I know which one would give me the fuzzier feeling.
But how many open source data mining projects can compete (in the features sense) with these commercial systems?
These systems are extremely specialised and targetted at law enforcement and/or large corporations with huge databases.
Seccessful OSS projects tend to be the ones that are used by the people writing them, and are of use to a wide community. If the developers do not have a vested interest in the product, then development will tend to stagnate.
I think it is hard to argue that OSS has been successful in making products that are targetted at such specific (and wealthy) groups.
You may want to read up on the DMCA, which makes it illegal to make or distribute technology that may be used to circumvent copy protection. I know making irrelevant references to DMCA has been flavour de jour on Slashdot for the past 5 years, but it doesn't have much at all to do with encryption in gerenal (only certain kinds/uses of encryption).
Since it is unlikely that anyone would consider using Enigma for any kind of copy protection or DRM, breaking Enigma would not be a problem for DMCA. Even if someone were crazy enough to use Enigma for some form of copy protection, it probably wouldn't make a difference. Most distributed projects to crack encryption schemes (which are generally used for copy-protection) are fine, because they only crack the key for a limited set of messages, so cannot directly be used to break copy-protection.
Further to that, we already know how Enigma worked. They just don't know the "key" for these particular messages. This is an effort to crack those 3 messages, not to perform a cryptographic attack on Enigma.
+1, Funny
What seems to be my problem with e-mail is that I send a message and I run the topic into the ground in that first e-mail (saying everything about it). Now, that's written in stone like a Slashdot comment. No backsies.
I don't have that problem, because I'm always right to begin with.
So what if they have to work 80+ hours per week. Back in my day, I had to work 80 hours per day!
Are there even five successful MMOG companies in existence now? I can think of Sony, Blizzard, NCSoft, maybe whoever it is that makes FF.
Hmm.. I think it's about time you cleaned the dried up coffee and lint out of your keyboard.
String theory is not falsifiable and therefore is unlikely to stand the test of time.
It most certainly is falsifiable - we just don't quite have the technology to test it yet, but by all measures, we appear to be pretty close (hence the article).
Assuming they produce a mini-blackhole with the LHC, if the observations do not match string theory's predictions, then it will have been falsified. They then need to either throw it out, or take it back to the drawing board.
The reference on that site to "Virgin Galactic" got me thinking. How do they differentiate between "flight", and "orbit"?
Do they say that the record is only available to jet-powered flight? Or do you have to be under power the whole time? Why doesn't a space agency hold this record?
The button is probably not as bad as the tap at the sink. I don't want to get too graphical here, but at least when people go for the button, they haven't wiped yet :). There's a good chance the average person's hands are still clean at that point.
:).
If you're really paranoid, I guess you can try using your foot or something, but that's an image I don't need to think about
I tried sitting in them perfectly still so as it thinks it is vacant and cleans itself.
Saves having to wipe, huh?
What studies to date show is a compound that attacks HIV at its molecular membrane level, disrupting the virus from interacting with their primary targets, the "T-helper" class white blood cells that comprise and direct the human immune system. Further, CSAs appear to be deadly to all known strains of HIV.
Umm - What studies? Does anyone else find it strange that there is no reference anywhere in the article as to how they came to believe that CSAs had this effect? Or am I just blind?
Was it tested in a test-tube? In monkeys? In humans?
Some people have ten friends...
On Slashdot?!? You must be new here.
Alright - from this day forward, I am hereby boycotting the Verizon ads. Whenever I click on the day-pass link, I am going to close my eyes for 15 seconds. During that 15 seconds, I am going to make a conscious effort not to think about what Verizon can do for my business.
Nothing for you to see here. Please move along.
Since these guys are all for freedom of information, and all that good stuff, do you reckon the research team would run Linux? ;)
No longer is it about organizing information, it's whether or not they want you listed.
Their track record says otherwise. In 2004, they came under fire for not removing an anti-semitic website, jewwatch.com, which was coming up as the first hit when searching for "jew". Even today, it is second only to Wikipedia.
Their argument at the time was that they were not going to block sites from their index based on content. According to that site that I linked, it was blocked in countries where the content of the site was illegal.
It looks to me like they will not block based on (legal) content, but will block people who fsck with PageRanks.
It's not particularly creative to implment sliding window schemes.
Of course not - the implementation itself is very rarely creative. Any monkey should be able to implement the chaining part of the solution I suggested. It is the problem solving that is creative; cross-pollinating solutions from different areas to solve a new problem, or an existing problem in a different way - That's creative. Now, take the matching part of my solution, where I've solved it with a bit of hand-waving (I haven't actually offered a solution to "fast fuzzy matching", and you're going to need a bit of creativity to determine a reasonable algorithm. Then, a code monkey can implement your solution.
There's very little 'science' in 'Computer Science' because most of it was figured out by mathematicians decades ago.
Not sure what that comment is in response to; the point of my post was that comp sci grads can do interesting stuff without staying at uni, but I'll bite anyway (I want to avoid the eternal "comp sci: science or engineering" side of things though). The reason there is very little 'science' in computer science, is not because it's all been figured out already. It is because 'computer science' is a big field; the term has been heavily overloaded in the past couple of decades.
There is ample 'applied comp sci research' going on at universities (well, the ones that focus on more than just churning out drones), and out in the real world (well, at the companies that focus on more than shipping the next product - they do exist!). The more theoretical stuff is harder to identify, because as you say it tends to be done by mathematicians, and it takes significant time to be adopted (I'm not aware of much theoretical computer science being done outside universities, either).
Industry needs to solve new problems all the time. Many of these problems can be solved by modifying an existing algorithm. Some require brand now algorithms. Whether the work be done at a university or in industry, it needs to be backed by scientific method (I agree that it often isn't, but there are good players out there). Almost no comp sci grads end up doing that kind of work, but if they want it, it *is* out there.
Yeah, it's just a shame that if you look through the replies to my post, you just see a bunch of poeple who missed my point. Everyone seemed to think that I was condoning the move, when all I was doing was accusing my parent poster of FUDing.
/. community seems so quick to turn any issue into a dichotomy. It's a shame really.
The
Kind of. The good news is that the Computer Scientist dorks can go out into the real world, do stuff that lets them use their brains, and a little creativity, all while (generally) getting paid more than the drones.
It's basically a statistical clustering problem that has been solved in a number of different ways (and I alluded to in my post with the fuzzy reference). Just off the top of my head, If I were to be writing such a system, I would probably try applying plagurism detection techniques.
Take for example Markov Chains, which are frequently used to detect plagurism in Computer Science schools at universities:
The basic idea is after possibly doing some basic preprocessing of the emails (perhaps you will remove punctuation), you break the emails up into bite sized chunks (probably words). Take some number of words in a row (usually 3), and hash them. Then slide a "window" along the email by a word, and generate another hash. Continue throughout the email. You end up with a list of hashes for each email.
It turns out you can do really fast fuzzy searches in a collection of these lists, especially if you don't mind getting a few false positives (which you can then rule out by directly comparing the "possibly matching" emails to get a "percentage similarity"). The lists themselves are quick to create (linear over the amount of text in the email), don't take up much space (they're usually a little smaller than the main-text of the email), and don't need to be held for long (you're only really interested in the rate that similar emails are arriving).
No this affects mailing lists, not spammers.
There's a very fine line between your average mailing list and your average spammer. One man's mailing list is another man's spam.
they'll continue to send small volumes of mail
If they're truely sending small volumes, then they're probably not a problem.
If there are millions of boxes each sending small numbers of the same email, then it is trivial to realise that they are the same emailer.
The basic gist is: hash every email that comes in. Count identical hashes. If the count goes up too fast, block the messages matching that hash. The only real trickery is in defining the hash to be fuzzy (to stop a few characters difference from changing the hash).
My understanding is that they have no blacklist, so it has nothing to do with reporting messages as spam. If you send the 'same' email to more than some number of AOL recipients, then you are a bulk mailer, and must currently get yourself onto their whitelist.
That's just FUD. AOL has not said that they are blocking all email that is not sent through GoodMail. They are replacing their whitelist with the service.
You don't need to be on the whitelist to send personal emails today, so you won't need to pay to send email tomorrow.
This only affects senders of bulk emails (mailing lists and spammers).