I don't know about China, but in Russia, education is generally on a high level. A student in Moscow told me that when she visited the United States while she was still going to school, she went to a school, and pupils there had great difficulties in solving maths problems where she could just write down the result. Education in Russia is a bit old-fashioned, people learn lots of facts (e.g. they know many history dates by heart) and a lot of rather specialized things in natural sciences, but it seems to be quite efficient. A student I met in St.Petersburg was a specialist for radio transmission - but she thought she probably wouldn't be able to use it professionally except in a very limited way in a radio shop. That could change. My impression is that education in Russia can easily compete with Western Europe and Northern America and that there is a huge potential. Probably, that isn't different in some countries in East Asia.
Oh yes, tax those US companies twice for their coding overseas!;-) The IT market will soon be dominated by Russian, Chinese and Indian companies. Or do you want to introduce high import duties on code?
A solution to this problem would be a "temporary" tax of say 500% on all code developed over seas, and any administration done as well. This tax could then be lowered over time. Say 20-30 years.
I doubt if that could work. To avoid companies buying code from (at least formally) independent companies instead of having their own employees in countries like China or Russia programming it, this approach would mean that high import duties on code would have to be introduced.
I don't think that would be a viable solution.
First, it would be very hard to solve the problem of code smuggling. Nothing can be moved across boundaries as easily as electronic data.
Secondly, that would mean that everything having to do with IT would be significantly more expensive in those countries that have such protectionist policies, which would have huge economic consequences. I don't think governments and voters would be ready to accept that to save domestic IT jobs.
I think it is important that negative social consequences of globalisation are alleviated, but globalisation can't be stopped in such a way.
In the long run, the price gap will become smaller. As far as I know, salaries of IT specialists in India have risen sharply in recent years. Cheaper IT products could also lead to even wider application of IT, and that could mean that the demand for certain jobs that have to be close to the customers or their headquarters (many people want to be able to talk with someone face to face in their native language) could rise.
At events like LinuxTag, most communication probably isn't technical work as such, rather people talk about the possibilities and consequences of these matters.
I work in IT in German-speaking Switzerland. English is, of course, important for reading documentation and finding information in the Internet, but that's mostly technical jargon and has relatively little to do with profound knowledge of the English language. However, when you talk to customers, you have to know their language well, and my experience has been that a good knowledge of French is more important for people working in IT here (France and the French-speaking part of Switzerland isn't far). While English is very important for technical matters, non-tech people who have to take decisions about IT investments or clarify something about details of a program don't want to use a foreign language (the matter is complicated enough for them already when they talk about it in their native language). Therefore, I think the significance of the English language shouldn't be exaggerated.
Yes, and therefore the fact that most of LinuxTag was in German rather than English shows that it was an event a broader public and not just for experts. In my view, that's a good sign.
You'll end up having to blackhole entire countries to staunch it.
Why not? I block mails from South Korea and China already, but many people cannot do that because they rely on receiving legitimate mails from there. When most big countries have laws that have the effect of most spammers using servers in a few small countries where spamming is legal, blocking whole country ranges is much more realistic, and it would certainly be done. When the EU has a good law against spamming, it is certainly an important first step, and I hope that at least the US, China, South Korea and Brazil will follow soon. Although most spam in Europe comes from elsewhere, it would already mean something if it could stop spam sent through interbusiness.it, tpnet.pl (Poland will soon be a member of the EU) and terra.es.
I wouldn't be so pessimistic about the cost of enforcing the laws. Spam samples are easy to obtain (e.g. many peopole report them via systems like SpamCop), and it isn't necessary to trace down every piece of spam, if spammers have to take a significant risk that they are caught and have to pay a hefty fine, this will make spamming much less profitable.
I wouldn't be so sure that it doesn't work technically. There is a congestion charge for an area in Central London, people who enter an 8-square-mile area during working hours have to pay (see e.g. IHT article on that subject). The number plates are read automatically, only when someone doesn't pay, the pictures are viewed by a human being before a fine is imposed. The system relies on automatic reading of number plates, as do other systems for tracking car drivers illegally using bus lanes or speeding.
Of course, all these systems only control a very limited area, building a system that controls "everything that moves" in a large area would be very expensive at the moment, but, judging from existing experience, it seems to be feasible technically, and it can well be that it becomes much cheaper in the future.
Another question is, of course, how millions of information items of the kind 'X drove from A to B at 12:34' could be interpreted. If it is to get payment from X, it's clear what the aim is (which of course doesn't mean that the data could be used for something else, as well), but that's not the aim of DARPA. I think we shouldn't rely too much on such systems not being feasable technically, but think about possible abuse in time, before they are in place.
I think that's completely right. I think there are some misconceptions people who advocate the use of high-tech surveillance as a means against terrorism use.
Using data mining techniques as a means to detect terrorist activities is absolutely unfeasable. If there were many terrorist attacks every day, it might be possible to develop systems that find patterns automatically with methods similar to those used to analyze consumer behavior, but luckily, that's not the case. To find out about terrorist attacks, human intelligence is needed, people have to infiltrate networks of extremists and identify fields of potential danger (now, with hindsight, one thinks of flight schools, probably there are many others). Once a track has been found, technical means can be used to follow it, but it seems that they already exist - they can probably be improved, but the main problem seems to be finding the tracks, not following them.
Well-organized terrorist networks will also have the means to find out something about the kind of surveillance that is in place. In the worst case, they could get hold of surveillance information and use it (with such large systems it can be hard to avoid leaks) and in any case, they can take measures not to be conspicuous in the large flood of data. Systems that gather large masses of rather superficial information can probably be used against petty criminals and ordinary citizens (perhaps involved in political activities the government doesn't like), but hardly against terrorist or mafia organizations who know what to do not to be detected. Systems like Echelon can be used for economical espionage against unsuspecting companies, but they don't help much against people who think about surveillance.
With past terrorist attacks, it seems that there were warnings that weren't investigated. On one hand, there rather seems to be too much information than too little from systems like Echelon, it would be more important to organize what is there in a better way. Then, there is certainly information that is missing, but that's rather not the kind of information that can be gathered with electronic means. Infiltrating terrorist networks and their surroundings is certainly difficult, but it is much more efficient than gathering superficial information about millions of people and hoping that something helpful will show up.
What I don't know is how these projects should be interpreted. Are they just looking for lucrative orders from the government that won't help against terrorism, but make some people richer or is there really the wish of people in the government to gather information about everyone in order to get more power? Probably, it's both, to some degree.
The free market is good at some things, but it certainly isn't enough for everything - the increasing spam problem is just a case in point. Would you perhaps want to regulate the amount of pollution that is permitted by the free market (I don't mean ideas about vouchers that are bought for a permission to cause a limited amount of pollution, that's already a market regulated by the state and makes sense, in my view, but it certainly wouldn't make sense to allow any amount of pollution or spamming without regulation)? There are traffic rules, the market doesn't regulate that (what about bidding at the crossroads for the right to pass first?). The market isn't an almighty god or some magic that can solve everything, and in the case of spammers that cause lots of costs (mainly work time) for which they don't have to pay, it clearly doesn't provide a solution.
Comcast.net, Charter.net, Sprint.net, AOL and Swbell.com/sbc.com are mainly US-American ISPs, aren't they? Quite a significant part of the spam I receive at spamtraps and normal addresses is from their IP ranges.
It's true, some countries like China and Brazil rank higher as spam source than the US, but when you read these spam mails sent via Chinese or Brazilian servers or the spamvertized websites, it says things like "We ship to all 50 states" (especially with the majority of medical spam, sometimes they even ship to Canada, in addition), much of the insurance and mortgage spam is meant only for US Americans. The vast majority of spam I see is US business - whatever servers they use, and just as much as servers in China, South Korea, Brazil and Italy should be secured against spammers, I think it is up to the US to fight this way of doing business. If US members of parliament want, they can certainly make legislation that penalizes American companies that use spamming - whether through US or other servers.
By the way, one of the most frequent categories of my spam clearly not intended for US citizens is... well, green card lotteries and other ways of obtaining visas to the US. Certainly, other countries have a lot to do, as well, but the need for the US to do something against spam is most pressing. There isn't much Chinese spam for and by 'people in all provinces', French spam for 'people in all départements' or German spam for all Bundesländer, but loads of spam for 'all 50 states'.
So what does that mean? Something won't work 100%, so don't do anything at all? Do you suggest legalizing murder, rape and speeding because these crimes happen despite of the laws?
Certainly, there will still be spam when spamming is illegal, but if there is a good law there will be less spam because spammers would take higher risks.
Even their language has nothing to do with Russian. It shares its roots with Swedish and Norwegian.
Yes, that's really nonsense.
- Estonian does not share any roots with Swedish and Norwegian (it does share roots with Finnish and Karelian).
- The fact that Estonia (like Latvia and Lithuania) was an independent state before it was occupied by the USSR does not have anything to do with 'even' theEstonian language not being related to Russian. The Ukraine and Belarus are the only former Soviet republics whose language is related to Russian (their languages are East-Slavic, as well). The language of Moldavia is very close to Romanian, Latvian and Lithuanian are Baltic languages (a separate group of Indoeuropean languages) and the languages of most former Soviet republics in Central Asia are related to Turkish.
Belarus perhaps or Moldavia? Certainly not. Also the Ukraine is worse off economically than Russia.
Or did you think of the republics in Central Asia (Uzbekistan, Turkmenistan, Kyrgystan,...), most of which are totalitarian dictatorships and very far away from prospering.
The only post-Soviet countries that can perhaps be said to prosper are the three small Baltic republics. It is hardly an accident that it's just the countries that are closest (culturally) to Western Europe. So, there's not really a contrast to the "Western World", they have just joined it with quite some success. They are not that different, there is also economic hardship in the Baltic states, especially in Latvia and Lithuania, although things are improving.
OK, they can't depend on other countries for military aggressions...
But they can have a huge trade deficit and depend on lost of money coming to the US from other countries?
Of course, trade wars usually affect both sides, so everyone is careful for the time being, but when US politics gets even further away of international law and becomes even more threatening, this will sooner or later trigger co-ordinated economic actions by the European Union, Japan and other countries that will bring the US to its knees economically. Of course, some Americans might then consider conquering these countries, after all, the military is the only area in which the US is stronger than the rest of the world if they work together, but an impoverished US could only drop bombs, a successful occupation to reap economic benefits would not be possible. Probably, US politicians will come to reason before that happens.
Well, at least, the more difficult cleaning the data is the more expensive address marketing gets...
Something that should work quite well is using different first names. Once it happened by accident, and then I could see who gave the address to whom. Both the snail mail service and the marketeers probably assume that it's another member of the same family.
Of course, the average windows user does not have his or her own domain, but I think services like Emailias, Sneakemail, SpamEx or SpamMotel are in principle suited for the masses. It's true that they are not very known, and most people seem to take it for granted that one has a limited and relatively small number of e-mail addresses, the idea to create a new address for every new contact (with a few mouse clicks) seems strange to them. I think that can change when mainstream e-mail providers (Hotmail, AOL, GMX etc) offer disposable e-mail addresses.
So, if they are serious about preventing spamming, these large mainstream mail providers should acquire a few hundred domains and let people create additional addresses for their existing accounts in an easy way on the web interface. That would be relatively easy to do.
I don't think the basic definition of crime implies injury, 'attempted murder' has a long tradition. No one is injured. I think that is a crime that should be punished.
Two things are important:
- planning is different from attempting. In many countries, planning to overthrow the government is a crime, the rationale behind this is that this must be prosecuted in an early stage because after the coup the current justice system won't work any more - that's of course prone to abuse -, but generally planning a crime is not a crime by itself. The important difference is that a person who planned, but didn't attempt to do something could still have given up the plan - someone who buys a gun and draws maps of the victim's house can still decide not to commit the murder. I don't know enough about the case, but my impression from the article is that the sale of the devices was still in the stage of planning and therefore shouldn't be punished.
- One cannot be prosecuted in general for attempting to do something illegal, only for severe crimes like murder. I don't know exactly about the US, but in Switzerland, people can be sent to prison or fined for attempted morder or attempted incurring of grievous bodily harm, but certainly not for attempting to cause material damage. I think that is justified, someone who shot at people to kill them and missed them must be punished, but not someone who attempted to smash a window and missed it (or developed devices for decoding satellite TV signals, but was stopped before he could sell them).
I think LyX, which has already been mentioned in another post, is a good compromise. It is based on LaTeX, so you can use LaTeX for formulae, tree diagrams etc, and on the other hand, it comes quite close to WYSIWYG (if you count previewing the DIV files, it is WYSIWYG). People who are used to LaTeX perhaps prefer writing LaTeX source code, but I think for many LyX is a good way to have the best of both worlds.
How long was that name? The hotmail space is so overcrowded that not only dictionary attacks, but also brute force is used quite efficiently by spammers.
I haven't used Popfile, but I'm generally quite sceptical about Bayesian filtering. I tried out the Bayesian filtering system of Mozilla, it needed an awful lot of mails for training until it had somehow acceptable results, and even then, it had quite a lot of false positives. Maybe, I expect too much. I expect that an e-mail in a language the filtering system has not seen before will not be misrecognized as spam, but invariably, some of the first mails in a language not seen before were labeled as spam.
I think there is a more fundamental problem with Bayesian filtering. Of course, you can get very high percentages with them, but that doesn't mean so much. Quantitatively, most e-mails we usually receive follow certain patterns (mails from mailing lists on a certain subjects, in general mails on subjects we often deal with), and it's quite easy for Bayesian filters to learn their characteristics. However, I think, as a rule, unusual messages are more important than those that are very similar to thousands of others (rare and unexpected events have more informational value), and it's just these unusual mails that are more likely to end up as false positives. So, the nice percentages can be quite misleading.
Maybe, I'm too radical here, for people who only rely on receiving mails in few languages for which the Bayesian system soon has enough samples, their performance isn't that bad. But I strongly dislike the principle underlying Bayesian filters: "What conforms to the rest is good, what is unusual is suspicious."
An example to illustrate what I mean (it's not very realistic, of course): Person A daily receives e-mails from people of the opposite sex wanting a date. So, every new mail of that category are safe. However, A does not have a job and always waits in vain for e-mail with job offers. When finally one such mail arrives, it goes into the spam folder and is overlooked because it doesn't have the characteristics of the legitimate mails A usually receives (and has some distant similarity with work-at-home and MMF spam).
B daily receives e-mails with great job offerings and makes an unusual career. So, every new mail of that category is safe. However, B is lonely in private life and waits in vain for a date proposal. When finally one such mail arrives, it's put into the spam folder and overlooked because it doesn't have the characteristics of the legitimate mails B usually receives (and has some distant similarity with spam advertising dating sites and telephone numbers).
Both A and B have very low rates of false positives, so the Bayesian filters are working well...
I prefer systems that check for typical spam characteristics and mail source (Spamassassin, Blacklists), in my experience, they aren't less efficient than these hyped Bayesian filters, but in contrast to them, they do not promote conformity against diversity. Maybe, Bayesian criteria are useful as an add-on, but I would never want to base the decision whether something is spam solely on them. What should I do when Bayesian filtering becomes popular, thresholds are set lower and lower because of the rising spam problem and I want to write someone a message? Should I try to guess which wordings conform better to the usual correspondence of the person I write to - otherwise there's the risk that the message won't be seen?
That would already be total capitulation, spam would have defeated e-mail.
Sign up for a throwaway Hotmail address. Never give the address out to anyone. Never use it for registration. Just let it sit there for a month or so. Then log into it and see the mountains of spam it contains.
That's everyone's experience, I think. I used to create throwaway accounts at Hotmail, and spam came in right from beginning.
Since you never gave this address to anyone, the only possible way the spammers got the address is because Microsoft sold it to them.
Don't rush to such conclusions. May I guess who the addresses in question looked like? Probably like the ones I used to register - a letter for the first name combined with a given name or a word that can be found in any dictionary. Who knows, maybe someone else had the address previously, gave it away everywhere and let it expire - in any case, it's easy for spammers to guess such addresses. I think no one should use e-mail addresses with domains like hotmail.com or aol.com for anything else than throwaway accounts. With these domains, it is so likely that a relatively short, pronounceable address you make up or generate with a program corresponds to an actual e-mail address that dictionary attacks are very efficient and used very often. I experienced the difference with GMX - a mail provider that is probably at least as popular in central Europe as Hotmail -, I received much more spam to addresses at gmx.de than to addresses at gmx.ch (Switzerland is much smaller than Germany, the likelihood of a random address @gmx.ch existing is therefore smaller and dictionary attacks less likely).
I've just registered a Hotmail address with 30 random characters and numbers. I am quite certain that it won't receive any spam. If it does, I might be inclined to believe that Microsoft sells addresses (I'm certainly not the first person to conduct this experiment). Of course, Microsoft does a lot of evil;-), but it's hard to believe that they would sell addresses to spammers. It's not in their interest because spammers can collect addresses quite cheaply, so the amount of money Microsoft would receive would be rather small, and when people receive too much spam, they may abandon the account and not see Microsoft's ads any more.
I don't think there's a danger that e-mail as a whole is dominated by Microsoft. I don't know the exact figures about mail servers, but those who decided to receive only mails sent via Microsoft servers would be in a disadvantaged position, even if it's a whole country.
On the other hand, I wouldn't be so optimistic about people flocking to the cheap alternative. In my view, many GNU/Linux distributions and OpenOffice.org can compete directly with MS systems for the everyday user as to useability and ease of use even now, but a very small proportion of the population uses them. Many have probably heard about Linux, but they have hardly seen it in action and would be very surprised how easy it is for an average Windows user to work with Gnome or KDE. What I find more striking is that most people I know, even people working in IT, haven't ever heard about OpenOffice.org. Somehow, they take it for granted that for presentations one needs MS Powerpoint, for writing formatted texts MS Word etc (some know about WordPerfect and other non-MS proprietary programs, but I haven't encoutered many users of OpenOffice.org), and they seem to accept that they either have to pay lots of money for that or to infringe on Microsoft's copyrights. I don't know how this is going to change - a huge advertising campaign might be needed to make average consumers more aware about free software, but that's not the way OSS spreads - who would finance it? In many cases when someone without special interest in IT uses OSS, it is probably due to a personal recommendations - I guess it's hard to predict whether that way a critical mass can be reached.
GNU/Linux distributors like Redhat, SuSE, Mandrake etc. do play an important role, Linux does have some presence in software shops, but, of course, they are very far from being able to run advertizing campaigns for everyday computer users like Microsoft or Apple.
Yes, but it's not only the right to be heard the US constitution doesn't guarantee, it doesn't guarantee freedom of speech, either, and it doesn't even try to protect it - as long as it isn't the state that prevents freedom of speech. If you don't want to publish a reply, just don't, then the person who wants to have a reply published can go to court, and they decide whether the reply has to be published.
Yes, of course you'd have to take them to court, but just to demand that they publish the reply, not for some American-style multi-million dollar recompensation claim. Maybe that wouldn't work in the US because US lawyers and courts would make extremely expensive and complicated suits when someone demands that a reply is published, but it does work with European newspapers, and I assume it will also work for the Internet.
Is freedom of speech really that strong in the US? One aspect of it, namely prohibiting that the state limits freedom of speech is stricter in the US than in other countries. But that is only one of several aspects of freedom of speech. Freedom of speech can, for example, be restricted by the concentration of mainstream media and their business interest or the threat of multi-million damage claim. In these respects, the protection of freedom of speech is more developed in Western Europe. The right to get a rebuttal published diminishes the limitation of freedom of speech that arises from the power of large mainstream media, and the more moderate amounts of compensation payments and costs of litigation diminishes self-censorship.
Slashdot Mirror
I don't know about China, but in Russia, education is generally on a high level. A student in Moscow told me that when she visited the United States while she was still going to school, she went to a school, and pupils there had great difficulties in solving maths problems where she could just write down the result. Education in Russia is a bit old-fashioned, people learn lots of facts (e.g. they know many history dates by heart) and a lot of rather specialized things in natural sciences, but it seems to be quite efficient. A student I met in St.Petersburg was a specialist for radio transmission - but she thought she probably wouldn't be able to use it professionally except in a very limited way in a radio shop. That could change. My impression is that education in Russia can easily compete with Western Europe and Northern America and that there is a huge potential. Probably, that isn't different in some countries in East Asia.
Oh yes, tax those US companies twice for their coding overseas! ;-) The IT market will soon be dominated by Russian, Chinese and Indian companies. Or do you want to introduce high import duties on code?
A solution to this problem would be a "temporary" tax of say 500% on all code developed over seas, and any administration done as well. This tax could then be lowered over time. Say 20-30 years.
I doubt if that could work. To avoid companies buying code from (at least formally) independent companies instead of having their own employees in countries like China or Russia programming it, this approach would mean that high import duties on code would have to be introduced.
I don't think that would be a viable solution.
First, it would be very hard to solve the problem of code smuggling. Nothing can be moved across boundaries as easily as electronic data.
Secondly, that would mean that everything having to do with IT would be significantly more expensive in those countries that have such protectionist policies, which would have huge economic consequences. I don't think governments and voters would be ready to accept that to save domestic IT jobs.
I think it is important that negative social consequences of globalisation are alleviated, but globalisation can't be stopped in such a way.
In the long run, the price gap will become smaller. As far as I know, salaries of IT specialists in India have risen sharply in recent years. Cheaper IT products could also lead to even wider application of IT, and that could mean that the demand for certain jobs that have to be close to the customers or their headquarters (many people want to be able to talk with someone face to face in their native language) could rise.
At events like LinuxTag, most communication probably isn't technical work as such, rather people talk about the possibilities and consequences of these matters.
I work in IT in German-speaking Switzerland. English is, of course, important for reading documentation and finding information in the Internet, but that's mostly technical jargon and has relatively little to do with profound knowledge of the English language. However, when you talk to customers, you have to know their language well, and my experience has been that a good knowledge of French is more important for people working in IT here (France and the French-speaking part of Switzerland isn't far). While English is very important for technical matters, non-tech people who have to take decisions about IT investments or clarify something about details of a program don't want to use a foreign language (the matter is complicated enough for them already when they talk about it in their native language). Therefore, I think the significance of the English language shouldn't be exaggerated.
Yes, and therefore the fact that most of LinuxTag was in German rather than English shows that it was an event a broader public and not just for experts. In my view, that's a good sign.
You'll end up having to blackhole entire countries to staunch it.
Why not? I block mails from South Korea and China already, but many people cannot do that because they rely on receiving legitimate mails from there. When most big countries have laws that have the effect of most spammers using servers in a few small countries where spamming is legal, blocking whole country ranges is much more realistic, and it would certainly be done. When the EU has a good law against spamming, it is certainly an important first step, and I hope that at least the US, China, South Korea and Brazil will follow soon. Although most spam in Europe comes from elsewhere, it would already mean something if it could stop spam sent through interbusiness.it, tpnet.pl (Poland will soon be a member of the EU) and terra.es.
I wouldn't be so pessimistic about the cost of enforcing the laws. Spam samples are easy to obtain (e.g. many peopole report them via systems like SpamCop), and it isn't necessary to trace down every piece of spam, if spammers have to take a significant risk that they are caught and have to pay a hefty fine, this will make spamming much less profitable.
I wouldn't be so sure that it doesn't work technically. There is a congestion charge for an area in Central London, people who enter an 8-square-mile area during working hours have to pay (see e.g. IHT article on that subject). The number plates are read automatically, only when someone doesn't pay, the pictures are viewed by a human being before a fine is imposed. The system relies on automatic reading of number plates, as do other systems for tracking car drivers illegally using bus lanes or speeding.
Of course, all these systems only control a very limited area, building a system that controls "everything that moves" in a large area would be very expensive at the moment, but, judging from existing experience, it seems to be feasible technically, and it can well be that it becomes much cheaper in the future.
Another question is, of course, how millions of information items of the kind 'X drove from A to B at 12:34' could be interpreted. If it is to get payment from X, it's clear what the aim is (which of course doesn't mean that the data could be used for something else, as well), but that's not the aim of DARPA. I think we shouldn't rely too much on such systems not being feasable technically, but think about possible abuse in time, before they are in place.
- Using data mining techniques as a means to detect terrorist activities is absolutely unfeasable. If there were many terrorist attacks every day, it might be possible to develop systems that find patterns automatically with methods similar to those used to analyze consumer behavior, but luckily, that's not the case. To find out about terrorist attacks, human intelligence is needed, people have to infiltrate networks of extremists and identify fields of potential danger (now, with hindsight, one thinks of flight schools, probably there are many others). Once a track has been found, technical means can be used to follow it, but it seems that they already exist - they can probably be improved, but the main problem seems to be finding the tracks, not following them.
- Well-organized terrorist networks will also have the means to find out something about the kind of surveillance that is in place. In the worst case, they could get hold of surveillance information and use it (with such large systems it can be hard to avoid leaks) and in any case, they can take measures not to be conspicuous in the large flood of data. Systems that gather large masses of rather superficial information can probably be used against petty criminals and ordinary citizens (perhaps involved in political activities the government doesn't like), but hardly against terrorist or mafia organizations who know what to do not to be detected. Systems like Echelon can be used for economical espionage against unsuspecting companies, but they don't help much against people who think about surveillance.
With past terrorist attacks, it seems that there were warnings that weren't investigated. On one hand, there rather seems to be too much information than too little from systems like Echelon, it would be more important to organize what is there in a better way. Then, there is certainly information that is missing, but that's rather not the kind of information that can be gathered with electronic means. Infiltrating terrorist networks and their surroundings is certainly difficult, but it is much more efficient than gathering superficial information about millions of people and hoping that something helpful will show up.What I don't know is how these projects should be interpreted. Are they just looking for lucrative orders from the government that won't help against terrorism, but make some people richer or is there really the wish of people in the government to gather information about everyone in order to get more power? Probably, it's both, to some degree.
The free market is good at some things, but it certainly isn't enough for everything - the increasing spam problem is just a case in point. Would you perhaps want to regulate the amount of pollution that is permitted by the free market (I don't mean ideas about vouchers that are bought for a permission to cause a limited amount of pollution, that's already a market regulated by the state and makes sense, in my view, but it certainly wouldn't make sense to allow any amount of pollution or spamming without regulation)? There are traffic rules, the market doesn't regulate that (what about bidding at the crossroads for the right to pass first?). The market isn't an almighty god or some magic that can solve everything, and in the case of spammers that cause lots of costs (mainly work time) for which they don't have to pay, it clearly doesn't provide a solution.
Comcast.net, Charter.net, Sprint.net, AOL and Swbell.com/sbc.com are mainly US-American ISPs, aren't they? Quite a significant part of the spam I receive at spamtraps and normal addresses is from their IP ranges.
It's true, some countries like China and Brazil rank higher as spam source than the US, but when you read these spam mails sent via Chinese or Brazilian servers or the spamvertized websites, it says things like "We ship to all 50 states" (especially with the majority of medical spam, sometimes they even ship to Canada, in addition), much of the insurance and mortgage spam is meant only for US Americans. The vast majority of spam I see is US business - whatever servers they use, and just as much as servers in China, South Korea, Brazil and Italy should be secured against spammers, I think it is up to the US to fight this way of doing business. If US members of parliament want, they can certainly make legislation that penalizes American companies that use spamming - whether through US or other servers.
By the way, one of the most frequent categories of my spam clearly not intended for US citizens is... well, green card lotteries and other ways of obtaining visas to the US. Certainly, other countries have a lot to do, as well, but the need for the US to do something against spam is most pressing. There isn't much Chinese spam for and by 'people in all provinces', French spam for 'people in all départements' or German spam for all Bundesländer, but loads of spam for 'all 50 states'.
So what does that mean? Something won't work 100%, so don't do anything at all? Do you suggest legalizing murder, rape and speeding because these crimes happen despite of the laws?
Certainly, there will still be spam when spamming is illegal, but if there is a good law there will be less spam because spammers would take higher risks.
Even their language has nothing to do with Russian. It shares its roots with Swedish and Norwegian.
Yes, that's really nonsense.
- Estonian does not share any roots with Swedish and Norwegian (it does share roots with Finnish and Karelian).
- The fact that Estonia (like Latvia and Lithuania) was an independent state before it was occupied by the USSR does not have anything to do with 'even' theEstonian language not being related to Russian. The Ukraine and Belarus are the only former Soviet republics whose language is related to Russian (their languages are East-Slavic, as well). The language of Moldavia is very close to Romanian, Latvian and Lithuanian are Baltic languages (a separate group of Indoeuropean languages) and the languages of most former Soviet republics in Central Asia are related to Turkish.
All these post-Soviet countries are prospering??
...), most of which are totalitarian dictatorships and very far away from prospering.
Belarus perhaps or Moldavia? Certainly not. Also the Ukraine is worse off economically than Russia.
Or did you think of the republics in Central Asia (Uzbekistan, Turkmenistan, Kyrgystan,
The only post-Soviet countries that can perhaps be said to prosper are the three small Baltic republics. It is hardly an accident that it's just the countries that are closest (culturally) to Western Europe. So, there's not really a contrast to the "Western World", they have just joined it with quite some success. They are not that different, there is also economic hardship in the Baltic states, especially in Latvia and Lithuania, although things are improving.
OK, they can't depend on other countries for military aggressions...
But they can have a huge trade deficit and depend on lost of money coming to the US from other countries?
Of course, trade wars usually affect both sides, so everyone is careful for the time being, but when US politics gets even further away of international law and becomes even more threatening, this will sooner or later trigger co-ordinated economic actions by the European Union, Japan and other countries that will bring the US to its knees economically. Of course, some Americans might then consider conquering these countries, after all, the military is the only area in which the US is stronger than the rest of the world if they work together, but an impoverished US could only drop bombs, a successful occupation to reap economic benefits would not be possible. Probably, US politicians will come to reason before that happens.
Well, at least, the more difficult cleaning the data is the more expensive address marketing gets...
Something that should work quite well is using different first names. Once it happened by accident, and then I could see who gave the address to whom. Both the snail mail service and the marketeers probably assume that it's another member of the same family.
Of course, the average windows user does not have his or her own domain, but I think services like Emailias, Sneakemail, SpamEx or SpamMotel are in principle suited for the masses. It's true that they are not very known, and most people seem to take it for granted that one has a limited and relatively small number of e-mail addresses, the idea to create a new address for every new contact (with a few mouse clicks) seems strange to them. I think that can change when mainstream e-mail providers (Hotmail, AOL, GMX etc) offer disposable e-mail addresses.
So, if they are serious about preventing spamming, these large mainstream mail providers should acquire a few hundred domains and let people create additional addresses for their existing accounts in an easy way on the web interface. That would be relatively easy to do.
I don't think the basic definition of crime implies injury, 'attempted murder' has a long tradition. No one is injured. I think that is a crime that should be punished.
Two things are important:
- planning is different from attempting. In many countries, planning to overthrow the government is a crime, the rationale behind this is that this must be prosecuted in an early stage because after the coup the current justice system won't work any more - that's of course prone to abuse -, but generally planning a crime is not a crime by itself. The important difference is that a person who planned, but didn't attempt to do something could still have given up the plan - someone who buys a gun and draws maps of the victim's house can still decide not to commit the murder. I don't know enough about the case, but my impression from the article is that the sale of the devices was still in the stage of planning and therefore shouldn't be punished.
- One cannot be prosecuted in general for attempting to do something illegal, only for severe crimes like murder. I don't know exactly about the US, but in Switzerland, people can be sent to prison or fined for attempted morder or attempted incurring of grievous bodily harm, but certainly not for attempting to cause material damage. I think that is justified, someone who shot at people to kill them and missed them must be punished, but not someone who attempted to smash a window and missed it (or developed devices for decoding satellite TV signals, but was stopped before he could sell them).
I think LyX, which has already been mentioned in another post, is a good compromise. It is based on LaTeX, so you can use LaTeX for formulae, tree diagrams etc, and on the other hand, it comes quite close to WYSIWYG (if you count previewing the DIV files, it is WYSIWYG). People who are used to LaTeX perhaps prefer writing LaTeX source code, but I think for many LyX is a good way to have the best of both worlds.
How long was that name? The hotmail space is so overcrowded that not only dictionary attacks, but also brute force is used quite efficiently by spammers.
I haven't used Popfile, but I'm generally quite sceptical about Bayesian filtering. I tried out the Bayesian filtering system of Mozilla, it needed an awful lot of mails for training until it had somehow acceptable results, and even then, it had quite a lot of false positives. Maybe, I expect too much. I expect that an e-mail in a language the filtering system has not seen before will not be misrecognized as spam, but invariably, some of the first mails in a language not seen before were labeled as spam.
I think there is a more fundamental problem with Bayesian filtering. Of course, you can get very high percentages with them, but that doesn't mean so much. Quantitatively, most e-mails we usually receive follow certain patterns (mails from mailing lists on a certain subjects, in general mails on subjects we often deal with), and it's quite easy for Bayesian filters to learn their characteristics. However, I think, as a rule, unusual messages are more important than those that are very similar to thousands of others (rare and unexpected events have more informational value), and it's just these unusual mails that are more likely to end up as false positives. So, the nice percentages can be quite misleading.
Maybe, I'm too radical here, for people who only rely on receiving mails in few languages for which the Bayesian system soon has enough samples, their performance isn't that bad. But I strongly dislike the principle underlying Bayesian filters: "What conforms to the rest is good, what is unusual is suspicious."
An example to illustrate what I mean (it's not very realistic, of course): Person A daily receives e-mails from people of the opposite sex wanting a date. So, every new mail of that category are safe. However, A does not have a job and always waits in vain for e-mail with job offers. When finally one such mail arrives, it goes into the spam folder and is overlooked because it doesn't have the characteristics of the legitimate mails A usually receives (and has some distant similarity with work-at-home and MMF spam).
B daily receives e-mails with great job offerings and makes an unusual career. So, every new mail of that category is safe. However, B is lonely in private life and waits in vain for a date proposal. When finally one such mail arrives, it's put into the spam folder and overlooked because it doesn't have the characteristics of the legitimate mails B usually receives (and has some distant similarity with spam advertising dating sites and telephone numbers).
Both A and B have very low rates of false positives, so the Bayesian filters are working well...
I prefer systems that check for typical spam characteristics and mail source (Spamassassin, Blacklists), in my experience, they aren't less efficient than these hyped Bayesian filters, but in contrast to them, they do not promote conformity against diversity. Maybe, Bayesian criteria are useful as an add-on, but I would never want to base the decision whether something is spam solely on them. What should I do when Bayesian filtering becomes popular, thresholds are set lower and lower because of the rising spam problem and I want to write someone a message? Should I try to guess which wordings conform better to the usual correspondence of the person I write to - otherwise there's the risk that the message won't be seen?
That would already be total capitulation, spam would have defeated e-mail.
Sign up for a throwaway Hotmail address. Never give the address out to anyone. Never use it for registration. Just let it sit there for a month or so. Then log into it and see the mountains of spam it contains.
;-), but it's hard to believe that they would sell addresses to spammers. It's not in their interest because spammers can collect addresses quite cheaply, so the amount of money Microsoft would receive would be rather small, and when people receive too much spam, they may abandon the account and not see Microsoft's ads any more.
That's everyone's experience, I think. I used to create throwaway accounts at Hotmail, and spam came in right from beginning.
Since you never gave this address to anyone, the only possible way the spammers got the address is because Microsoft sold it to them.
Don't rush to such conclusions. May I guess who the addresses in question looked like? Probably like the ones I used to register - a letter for the first name combined with a given name or a word that can be found in any dictionary. Who knows, maybe someone else had the address previously, gave it away everywhere and let it expire - in any case, it's easy for spammers to guess such addresses. I think no one should use e-mail addresses with domains like hotmail.com or aol.com for anything else than throwaway accounts. With these domains, it is so likely that a relatively short, pronounceable address you make up or generate with a program corresponds to an actual e-mail address that dictionary attacks are very efficient and used very often. I experienced the difference with GMX - a mail provider that is probably at least as popular in central Europe as Hotmail -, I received much more spam to addresses at gmx.de than to addresses at gmx.ch (Switzerland is much smaller than Germany, the likelihood of a random address @gmx.ch existing is therefore smaller and dictionary attacks less likely).
I've just registered a Hotmail address with 30 random characters and numbers. I am quite certain that it won't receive any spam. If it does, I might be inclined to believe that Microsoft sells addresses (I'm certainly not the first person to conduct this experiment). Of course, Microsoft does a lot of evil
I don't think there's a danger that e-mail as a whole is dominated by Microsoft. I don't know the exact figures about mail servers, but those who decided to receive only mails sent via Microsoft servers would be in a disadvantaged position, even if it's a whole country.
On the other hand, I wouldn't be so optimistic about people flocking to the cheap alternative. In my view, many GNU/Linux distributions and OpenOffice.org can compete directly with MS systems for the everyday user as to useability and ease of use even now, but a very small proportion of the population uses them. Many have probably heard about Linux, but they have hardly seen it in action and would be very surprised how easy it is for an average Windows user to work with Gnome or KDE. What I find more striking is that most people I know, even people working in IT, haven't ever heard about OpenOffice.org. Somehow, they take it for granted that for presentations one needs MS Powerpoint, for writing formatted texts MS Word etc (some know about WordPerfect and other non-MS proprietary programs, but I haven't encoutered many users of OpenOffice.org), and they seem to accept that they either have to pay lots of money for that or to infringe on Microsoft's copyrights. I don't know how this is going to change - a huge advertising campaign might be needed to make average consumers more aware about free software, but that's not the way OSS spreads - who would finance it? In many cases when someone without special interest in IT uses OSS, it is probably due to a personal recommendations - I guess it's hard to predict whether that way a critical mass can be reached.
GNU/Linux distributors like Redhat, SuSE, Mandrake etc. do play an important role, Linux does have some presence in software shops, but, of course, they are very far from being able to run advertizing campaigns for everyday computer users like Microsoft or Apple.
Yes, but it's not only the right to be heard the US constitution doesn't guarantee, it doesn't guarantee freedom of speech, either, and it doesn't even try to protect it - as long as it isn't the state that prevents freedom of speech.
If you don't want to publish a reply, just don't, then the person who wants to have a reply published can go to court, and they decide whether the reply has to be published.
Yes, of course you'd have to take them to court, but just to demand that they publish the reply, not for some American-style multi-million dollar recompensation claim. Maybe that wouldn't work in the US because US lawyers and courts would make extremely expensive and complicated suits when someone demands that a reply is published, but it does work with European newspapers, and I assume it will also work for the Internet.
Is freedom of speech really that strong in the US? One aspect of it, namely prohibiting that the state limits freedom of speech is stricter in the US than in other countries.
But that is only one of several aspects of freedom of speech.
Freedom of speech can, for example, be restricted by the concentration of mainstream media and their business interest or the threat of multi-million damage claim. In these respects, the protection of freedom of speech is more developed in Western Europe. The right to get a rebuttal published diminishes the limitation of freedom of speech that arises from the power of large mainstream media, and the more moderate amounts of compensation payments and costs of litigation diminishes self-censorship.