And no matter how the 'witty' people who post the 'Al Gore invented the internet' posts try to spin it, Al Gore never said anything even close to implying that he invented the internet.
Hmm, according to Snopes, you are both right and wrong.
I say that because Snopes classifies it as "False" but the explanation itself seems to be a spin. The quote itself on Snopes was:
"During my service in the United States Congress, I took the initiative in creating the Internet. I took the initiative in moving forward a whole range of initiatives that have proven to be important to our country's economic growth and environmental protection, improvements in our educational system."
Snopes then goes on to say that it is rediculous to believe that Al Gore believed he created the Internet. That's not in question. No-one believes Al Gore created the Internet and I doubt Al Gore himself believes it, but the fact of the matter is he said: "During my service in the United States Congress, I took the initiative in creating the Internet."
Perhaps, as Snopes concludes, it was simply a clumsy and self-serving phrasing that Gore used, but he did say it. I figure at worst it was self-serving and at best it was just stupid on Gore's part to say whatever it is he meant to say in that manner. But to say that others are spinning what Gore said is inaccurate. Many people jokingly mention it but, in the end, Gore DID say it in the above context--regardless of how much you wish he hadn't.
I usually keep a few grand in the bank and I move over $100,000 through that account a year...
Does anyone know of a way I can keep the bank from wasting my valuable time?
First solution: If you are moving $100k through your account per year try leaving a cash buffer in your account. I move less than $100k/year through mine and my account hasn't fallen below $2000 in years. You've gotta be cutting things pretty close if the difference between a $250 authorization and a $50 charge is the difference between positive/negative. (Not defending the bank, I think they're wong, but for such a large income I wouldn't think you'd need to cut it so close to zero before payday).
Nope, those shows relied solely on a strong story line, and never resorted to sex and violence to sell the show. Yep, that Cassiopea was all about adding to the story line;-)
I see your response is tounge-in-cheek, but there is a difference between having attractive woman on a show (nothing wrong) and running commercials showing a naked female Cylon (back view) apparently in the middle of having sex (blatant sexuality). Maybe the scene is even useful to the storyline, but the fact that they use it as a centerpiece of their advertising is telling.
I can't believe my original post got modded flamebait. Drugged out, porn-addicted moderator I guess.:)
Giving me something scary, or disturbing. If I'm in the mood for porn, I'm not going to waste my time on basic cable.
You got that right. The whole "sexy baby" thing is annoying, not just when they're the bad guys. Don't get me wrong, woman are beautiful. But like you said... if you want to see some hot chicks, you know where to go. Not EVERYTHING has to be about hot babes, and I'd like to be able to enjoy some fun science fiction without having to feel like I'm watching something that's really targetting some puberty-controlled teenage boy.
There was a Battlestar Galactica marathon on SciFi about two months ago and I got to watch a few episodes. Of course, most of the commercials were for the new Galactica that they were going to release in December. One of the commercials actually was talking/showing a photoshoot from Maxim magazine where apparently one of the actresses came from (that ought to tell you something... their cast consists of models, not actors apparently).
In other words, screw promoting the series based on its storyline (apparently no longer involving earth) or characters (which apparently have had sex changes and I get the impression that they even had to put a gay character in there, you know, to be PC and all)... rather just promote the blatant sexuality of the series. After all, with a sensual sex scene with a female Cylon, showing previous photoshoots of the actress^H^H^H^H^H model, and using the "Battlestar Galactica" goodwill, profit is guaranteed.
I don't plan on watching. I hope to buy the original series DVD box set this month, maybe even by Monday, and I'll watch that instead. Sure, this new series is completely different and should be judged on its own merits. It might even be fun. But I don't want to contaminate my memory of Battlestar Galactica by even associating this new material with the name.
I will be interested to see, though, how well the series does. They've basically alienated the real fan base that's been waiting for a new Battlestar Galactica for 25 years and I'm not convinced the new generation is really dying to see a remake. Who IS their market? Maybe it IS just the pubery-driven teenage boys.
What about the companies that are exclusively anti-spam? Maybe they don't even have to have a spamming component... The point is that whether or not a company does the "right" thing is the important thing, not whether it offers x product.
Despite the fact that we now often see certain anti-spam products being promoted in, of all things, spam, I'd like to believe that most anti-spam offerings are honestly interested in reducing spam for their customers.
We offer an anti-spam service which has been very, very successful and received very positive feedback from our satisfied customers. It's rewarding to receive that feedback. But a few weeks ago we received an email that I knew someone would send someday: They said it was unethical and suspicious of us to be selling an anti-spam service. Don't we benefit from spam by offering the anti-spam service for a fee? Maybe we are guilty of generating spam to drive people to our service?
Well, ignoring the minimal gains we've earned from the service (far outweighed by bandwidth and development costs), I feel TRUE anti-spam services are no more unethical than doctors. Are doctors unethical because they pay their bills curing others? They are unethical if they make people sick so they can charge to cure them, but if they truly are just curing people then that seems like a valuable service and not at all unethical (as long as they aren't ripping people off).
Fact is, there's so much spam that no company needs to create more of it to justify anti-spam remedies. That said, I don't think it is appropriate for any company to provide anti-spam and mass-mailing services or products at the same time. That just sounds wrong and definitely looks like a conflict of interest even if it might not be.
The real question is: If a company truly had the ability to end spam but it would end their cashflow, would they do it? I can't speak for other companies, but I know we would. We developed the service because WE were sick of spam. If we could make spam disappear tomorrow, we would do so and just find something else to do business-wise. We were solvent before spam appeared and we will be solvent after spam goes away (and I'm convinced it will, someday. Whether it is attacked with new mail protocols, filters, legislation, or all of the above, I don't think we'll be dealing with spam 5-10 years from now).
You notice the.biz thing because there are a lot fewer of them.
It's the ratio.
In my Bayesian corpus, the.COM extension in an HTML tag is a 90.43% spam probability (because most of my non-spam doesn't have HTML tags) and a 22.0% spam probability in free text.
Meanwhile, BIZ is a 99.92% spam probability when found in an HTML tag and a 90.5% spam probability in free text.
So, yes,.BIZ is a good spam token and I, too, have thought about filtering everything.BIZ. The main reason I don't is because my Bayesian filter catches 99.9% of it all anyway so there's no reason to bother increasing my false positives by filtering BIZ.
If they used online mailing, the phone line would be tied up for long periods doing nothing.
Well considering the story itself says that "BigPond will investigate cable and ADSL Internet customers sending more than 20 e-mails in a 10-minute period", I fail to see where staying online for long periods will tie up the phone line.
Unless ADSL and cable service in Australia ties up the phone?:)
I live in Mexico too. If you've been here for 5 years it'd be about time you learned spanish, so you wouldn't have to be "language isolated".
Hablo espanol perfectamente bien, pero aun asi extrano el poder hablar con gente diariamente en mi idioma nativo.
Also, there are lots of us 'technical' people here, and most of us speak english, some of us speak pretty good english, so you wouldn't have to feel 'isolated'.
I didn't say there aren't any technical people in Mexico. I just said that since I work alone in my home-office I don't get to talk to them very often. And when I do have lunch with colleagues from the Mexican company I used to work at we always speak in Spanish since I speak Spanish much better than they speak English.
So, step outside for a moment, and look around for the guys reading physics or programming books.
Heheh, I've never seen that.:) Then again, I don't lurk around universities either.
But I think you are misunderstanding what I was driving at. I miss being able to "shoot the breeze" in my own native language with technical people. Yes, I can do it in Spanish and sometimes do. And it's not that my Mexican friends don't know their stuff... But it's just not quite the same when it's not in your language.
By the way, you're not the american-looking guy who goes to Vips at San Jeronimo with a programming book, are you?
No, print is dead.:) But seriously, it's not me. What city are you in? I'm in Monterrey but I try to avoid Vips whenever possible. The only thing I can half-tolerate there is Consome de Pollo.:)
Tell me about it. I've been doing independent consulting from the home office for coming up on 5 years. Not only do I miss the office interaction (I kid you not), I'm technically isolated and "language" isolated in that I live in Mexico. I IM my tech colleagues from when I lived and worked in the US on a daily basis, but it does get a bit tiring 1) Working at home. 2) Not seeing or talking face-to-face with anyone technical. 3) Not seeing or talking face-to-face with anyone in English.
And people shouldn't discount the office interaction. Yes, we all work our 8 hours (usually more if you are self-employed), but in the office you have those minutes at the coffee machine, people stopping by your office or cube, or stepping outside to "have a smoke" (even if you don't smoke), or maybe going to lunch with some coworkers. They are things that seem incredibly trivial and often even annoying when you're in an office but you actually start to miss after working alone for awhile.
But I agree with you. Ironically it's a situation where you actually start to look forward to the on-site work--even if it's just meetings.:)
If the purpose of a search engine is to help us find the products/content we're looking for then why are they trying to filter out worthwhile search results?
Obviously you didn't RTFA. The problem is when the first 10 results in Google all point to the same company operating with different websites. This has happened to me on occasion, and it's downright annoying. I'm glad Google is taking action to fix this.
About 50% of the time when I'm searching, I AM looking for vendors of a product in order to do price comparisons... Maybe I'm missing something....
Yes... that the first 10 entries where you want to "compare" prices are really the same company so you're going to say "Wow, I guess that's what it costs!" rather than seeing results from other providers that might have a better price.
This isn't about ignoring valid results... this is about ignoring results that have been intentionally and artificially inflated in pursuit of the almighty dollar rather than in the pursuit of providing valuable content to the user.
By the way, foreigners "taking our jobs away" is a stupid argument. Human culture is built on a migrant work force. That's how a lot of cultural things such as a the wheel spread around the world. Imigration is healthy for a country. It's the ones that come to claim welfare that you don't want.
I agree with you completely. I have no problems with foreigners working in the U.S. I wasn't trying to complain, I was just pointing out that the U.S. is actually quite liberal when it comes to letting foreigners work in the country. More liberal than either Mexico or Colombia that both have much less to protect.
For call centers, perhaps, but I wouldn't bank on having the IT jobs return from cheaper lands. If the IT geek doesn't have to deal with the end user then the language barrier is virtually nonexistent, at least as far as the masses are concerned. Does the primary language of the person who programs your dialog boxes really matter?
Most software development does require interaction with "end user"--although the end user may be another department in the company.
And, yes, the primary language of the person who programs your dialog boxes does matter since most companies will want the source code that was outsourced and that source code better be readable. Source code can be hard enough to follow when it is written by someone in the same language. I have experience working with source code in English and Spanish. It's not pretty, and I *speak* Spanish.
To answer the article: Yes, I firmly believe that this is the first of what will be many outsourcing projects coming back to the US. I've been saying it for years... Outsourcing to the lowest bidder on the other side of the globe is a "fad" in the post-bubble world of trying to save every cent. I had predicted it'd be about 2004 or 2005 when we start seeing most of the outsourced projects coming back as one company after another found out that outsourcing to 3rd-world countries is not all that it was promised to be.
I'm not an expert here, but I don't think that's the case. As far as I'm aware, to get a job in America on a temp visa, a company must have no valid native applicants. There are many other countries that are far more open.
I'm sure there are some countries that are more open, but Mexico is identical: A company must have no other valid native applicants before they can higher a foreigner. Same with Colombia. There are are 200+ countries in the world and I don't claim to know all of them, but 3 out of 3 (U.S., Mexico, and Colombia) are all the same. Like I said, I just got a Mexican work visa because I was lucky enough to be a on a first-name basis with the Mexican consul in Denver.
I also have a friend that lives and works in the UK. My understanding is that was only possible because the company he worked for wanted him to work in their UK division on a temporary basis. My understanding is that his American ex-wife (they divorced while in UK) is having trouble getting permission to continue to live and work there.
In the U.S., however, the "no valid native applicants" is not very well enforced. That's why you have hundreds of thousands of foreigners running around working in the U.S. on B2 visas. Do you really think there are no Americans that can do the job of these B2 people from India and Pakistan? Of course there are but the foreigners are in the U.S. taking our jobs anyway.
So I still believe the U.S. is certainly no worse than most countries and arguably more lax than some since the U.S. has the same requirements as Mexico and Colombia but really don't enforce it that much.
Client side? I'll take server-side any day. Why would I want to download 250+ spams per day when the server could just as easily filter them for me?
If you have your mail on a POP server (ISP, hosting provider, etc.) try PrismEmail. It filters between your server and you so there is effectively no time or load on your computer, plus it works with virtually any mail client with nothing to install on the server or on the client.
First, I have to trust a company I've never heard of. I have to trust them enough that I give them my passwords and such, so they can log in to my server. Eventually, I'd have to trust them with a credit card number, too.
I'm sure you trust one or more companies with that same information now for your Internet access and/or mail now.
At my current levels, lets say it averages to 500 emails a day. Multiply * 365 = 182,500. Based on the PrismEmail pricing page, that means I'll need to be in the "Power User" range, even though the vast majority is spam.
Pricing is based on bandwidth, not mail or spam count. Those are provided as conservative examples. Luckily most spams are still less than 20k and you can receive quite a ton of it without going into plans that cost more than $20. I currently receive 250+ spams per day and receive quite a bit of good email, even attaches. I'm not even close to hitting the bandwidth limit of the $20/year plan.
I would have to poll their system. Their system then gets the mail from my POP3 server, then filters it, then sends me the good parts and does whatever it does with the rest. Is that really faster than letting MailWasher help as it does now? I'm not sure it does.
Depends. Is your POP3 currently local to your mail client? If so, it might not be faster. But if your mail is somewhere else then this is done in real-time anyway. It's not like it takes any longer to get the email to your client. The only difference is the mail that does get to your client is only good email.
That's one of the funnyiest things I've heard in a while. Is that what most Americans believe? No wonder less than 1 in 10 have a passport!!
I'm an American living in Mexico. I was able to get a Mexican work visa pretty much because I was providing Internet access to the Mexican consulate in Denver at the time and was on a first-name basis with him. I mentioned that I'd like to move to Mexico and work there. He made the paperwork happen.
It's not usually that easy. Mexico isn't anxious to give up its jobs to foreigners. I know Colombia isn't either.
For all the crap that gets slung at the U.S., we are still one of the most open when it comes to accept foreigners into our country and letting them live and work here.
The Federal Government should stay out of it. With or without a tax, a new law would cost money. You'd need enforcement to ensure that the tax is paid (or, in your example, the Bayesian filter is installed), etc . ..
Oh, I agree. But if the Feds are going to get involved I'd rather see a technical solution fix the problem rather than a new tax.
In theory, a bayesian filter can run on the server. I'm told that there are some that do that. But then you lose the ability to interact with the program, telling it "This was spam you let through" and "this was legitimate mail which you didn't deliver". The ones currently available let you do those things via a website. Sure, that's what I want to do, go to a website to figure out where my legitimate mail is at. Sorry, I'm just not willing to waste that much time.
At the risk of being flamed for promoting a useful service, try the site in my sig. There is a "Report this as spam" link in every message--click it and you just reported it as spam. Or you can receive a nightly summary of all the spams that were trapped in the last 24 hours and all the mail that was let through... then just quickly click on each one that was miscategorized. No need to go to the website to check for miscategorized mail--just re-categorize it once per day from the nightly mailing.
Bayesian filtering should be done on the server, definitely. It's the only logical place to do it.
What I *would* like to see is a "spam reporting RFC" that defines a protocol whereby email programs can submit "This is a spam" or "This was a false positive" to the server without using HTTP. That'd be very, very useful in opening up a new level of integration between the email client and email server.
That would assume that bayesian filters continue to be effective. They aren't. The filters match words, so the spammers stopped using words, which is why your email titles now read "G@t G.E.n.E.r.I.C. V1agra", and contain not a single full word in the entire email, nor the same randomised spelling on any two emails.
Yes, it does work. Let's take your example "G@t G.E.n.E.r.I.C. V1agra" and look at my current corpus:
If you sent me the above spam, it would be caught by my Bayesian filter. If you spelled every word one letter at a time it would be caught by my Bayesian filter. If you used a domain name it would be even easier to catch. If you use an IMG to embed an image rather than spelling out your message you'd have a 96.7% spam score right off the bat. If you embed a JPG iamge you have a 98.5% spam score while a GIF gives you a 95.1% spam score. And don't forget the headers themselves--as Paul Graham mentioned, it's easy to think the headers aren't important but there is a ton of useful information there.
Many people seem to think Bayesian filters aren't the solution or that spammers will get around them. Mark my words, they won't. They can't.
Miniscule... for now. One should never allow the government to start taxing something new, for the tax will inevitably grow. And considering how many emails are sent every day do you think the government would be shy in wanting to increase that?
If the Federal Government is going to pass any law that addresses spam, it should be a law that says that every user of email is required to use a Bayesian filter. That would lead to a virtually zero response rate within months which would lead to spam going away. And, wow, all without new taxes!
As a follow-up to my previous message, I just did some looking through my historical spam corpus (I keep it to test improvements to my Bayesian filter).
1. One message had the entire Bill of Rights (1st through 10th Amendments) scattered throughout the spam, in white font. The message still got a spam score of 99.061171%.
2. Another message just blatantly included the following words (also in white text) to try to lower the spam score. The terms they included are listed along with their spam probability in my corpus. NA=Not Applicable because the term has not been used sufficiently to call it spam or good, so it receives a 40% score. OS=Only Spam has used this term, so automatically 99.9% score.
In almost all cases of NA (which means the term did not effect the Bayesian score for the message), the only usages were in spam--which means after just a few more messages these are going to all convert to 99.9% terms. In which case the use of these terms in a future spam will bury it.
The only term that was really low (forum at 0.8%) is because I run a forum.
There were two terms that are *ONLY* used in spam (99.9% score), and 1 term that had an 80% score. So by inserting "innocent" text, this spam actually gave me two more terms that were very much spammy. In this case, the two 99.9% terms effectively canceled out their lucky hit on forum (0.8%) and less lucky hit on incomplete (8.5%).
Even though they hit a single good term (forum), it is really irrelevant. Bayesian doesn't look at ALL terms, it looks at the most INTERESTING terms. That means the most spammy and the least spammy are considered--nothing in between--I use Paul Graham's implementation (15 most interesting terms). As it turns out, almost all of the 15 most interesting terms in this spam had 99.9% ratings, and all of them were above 90%. So, at best, 2 or 3 of the random terms were considered (forum, incomplete, determine). But even so, they were no match for the overwhelmingly spammy words (and HTML tags) used in the spam.
Result: The message was caught with a spam score of over 90%.
Statistics are fun, and it'll be interesting to see how long it takes spammers to realize that they can't get around Bayesian. Their attempts to get around it, as shown in the above random example, at best are a wash (no effect on its status as spam) or may even INCREASE the spam score since they may just as easily hit spammy words as innocent words.
If spammers can get filters to chuck a small percentage of legitimate e-mail, users will turn off the spam filters.
That's the only thing they could possibly be thinking that is somewhat reasonable... but even then I don't think they can make a difference. The terms that Bayesian ends up using to decide something is legitimate mail are going to be the name of your mother, your best friends, the bar where you get together with your friends, etc. It is very unlikely spammers will be able to seed messages that happen to have those words in it--if they did, they could just try sending you spam and use those same words to decrease the spamminess of their message.
The terms they inevitably will end up including in either a seed message or in the spam itself (with the intent of lowering the spamminess) are the terms that are not going to be particularly biased one way or the other, aren't going to be used in determining spaminess, so aren't very useful when it comes to seeding.
Next time I see a spam in my spam folder that has an attempt to evade Bayesian I will run through their "seed" words and see what kind of spam percentages those words have in my corpus. It ought to be interesting.
Slashdot Mirror
Hmm, according to Snopes, you are both right and wrong.
I say that because Snopes classifies it as "False" but the explanation itself seems to be a spin. The quote itself on Snopes was:
Snopes then goes on to say that it is rediculous to believe that Al Gore believed he created the Internet. That's not in question. No-one believes Al Gore created the Internet and I doubt Al Gore himself believes it, but the fact of the matter is he said: "During my service in the United States Congress, I took the initiative in creating the Internet."
Perhaps, as Snopes concludes, it was simply a clumsy and self-serving phrasing that Gore used, but he did say it. I figure at worst it was self-serving and at best it was just stupid on Gore's part to say whatever it is he meant to say in that manner. But to say that others are spinning what Gore said is inaccurate. Many people jokingly mention it but, in the end, Gore DID say it in the above context--regardless of how much you wish he hadn't.
First solution: If you are moving $100k through your account per year try leaving a cash buffer in your account. I move less than $100k/year through mine and my account hasn't fallen below $2000 in years. You've gotta be cutting things pretty close if the difference between a $250 authorization and a $50 charge is the difference between positive/negative. (Not defending the bank, I think they're wong, but for such a large income I wouldn't think you'd need to cut it so close to zero before payday).
Second suggestion: Get a new bank. That easy.
I see your response is tounge-in-cheek, but there is a difference between having attractive woman on a show (nothing wrong) and running commercials showing a naked female Cylon (back view) apparently in the middle of having sex (blatant sexuality). Maybe the scene is even useful to the storyline, but the fact that they use it as a centerpiece of their advertising is telling.
I can't believe my original post got modded flamebait. Drugged out, porn-addicted moderator I guess. :)
You got that right. The whole "sexy baby" thing is annoying, not just when they're the bad guys. Don't get me wrong, woman are beautiful. But like you said... if you want to see some hot chicks, you know where to go. Not EVERYTHING has to be about hot babes, and I'd like to be able to enjoy some fun science fiction without having to feel like I'm watching something that's really targetting some puberty-controlled teenage boy.
There was a Battlestar Galactica marathon on SciFi about two months ago and I got to watch a few episodes. Of course, most of the commercials were for the new Galactica that they were going to release in December. One of the commercials actually was talking/showing a photoshoot from Maxim magazine where apparently one of the actresses came from (that ought to tell you something... their cast consists of models, not actors apparently).
In other words, screw promoting the series based on its storyline (apparently no longer involving earth) or characters (which apparently have had sex changes and I get the impression that they even had to put a gay character in there, you know, to be PC and all)... rather just promote the blatant sexuality of the series. After all, with a sensual sex scene with a female Cylon, showing previous photoshoots of the actress^H^H^H^H^H model, and using the "Battlestar Galactica" goodwill, profit is guaranteed.
I don't plan on watching. I hope to buy the original series DVD box set this month, maybe even by Monday, and I'll watch that instead. Sure, this new series is completely different and should be judged on its own merits. It might even be fun. But I don't want to contaminate my memory of Battlestar Galactica by even associating this new material with the name.
I will be interested to see, though, how well the series does. They've basically alienated the real fan base that's been waiting for a new Battlestar Galactica for 25 years and I'm not convinced the new generation is really dying to see a remake. Who IS their market? Maybe it IS just the pubery-driven teenage boys.
Despite the fact that we now often see certain anti-spam products being promoted in, of all things, spam, I'd like to believe that most anti-spam offerings are honestly interested in reducing spam for their customers.
We offer an anti-spam service which has been very, very successful and received very positive feedback from our satisfied customers. It's rewarding to receive that feedback. But a few weeks ago we received an email that I knew someone would send someday: They said it was unethical and suspicious of us to be selling an anti-spam service. Don't we benefit from spam by offering the anti-spam service for a fee? Maybe we are guilty of generating spam to drive people to our service?
Well, ignoring the minimal gains we've earned from the service (far outweighed by bandwidth and development costs), I feel TRUE anti-spam services are no more unethical than doctors. Are doctors unethical because they pay their bills curing others? They are unethical if they make people sick so they can charge to cure them, but if they truly are just curing people then that seems like a valuable service and not at all unethical (as long as they aren't ripping people off).
Fact is, there's so much spam that no company needs to create more of it to justify anti-spam remedies. That said, I don't think it is appropriate for any company to provide anti-spam and mass-mailing services or products at the same time. That just sounds wrong and definitely looks like a conflict of interest even if it might not be.
The real question is: If a company truly had the ability to end spam but it would end their cashflow, would they do it? I can't speak for other companies, but I know we would. We developed the service because WE were sick of spam. If we could make spam disappear tomorrow, we would do so and just find something else to do business-wise. We were solvent before spam appeared and we will be solvent after spam goes away (and I'm convinced it will, someday. Whether it is attacked with new mail protocols, filters, legislation, or all of the above, I don't think we'll be dealing with spam 5-10 years from now).
It's the ratio.
In my Bayesian corpus, the .COM extension in an HTML tag is a 90.43% spam probability (because most of my non-spam doesn't have HTML tags) and a 22.0% spam probability in free text.
Meanwhile, BIZ is a 99.92% spam probability when found in an HTML tag and a 90.5% spam probability in free text.
So, yes, .BIZ is a good spam token and I, too, have thought about filtering everything .BIZ. The main reason I don't is because my Bayesian filter catches 99.9% of it all anyway so there's no reason to bother increasing my false positives by filtering BIZ.
Yeah, but since Australia isn't in Europe it really doesn't matter since it fails the "relevant law" test. :)
Well considering the story itself says that "BigPond will investigate cable and ADSL Internet customers sending more than 20 e-mails in a 10-minute period", I fail to see where staying online for long periods will tie up the phone line.
Unless ADSL and cable service in Australia ties up the phone? :)
Hablo espanol perfectamente bien, pero aun asi extrano el poder hablar con gente diariamente en mi idioma nativo.
Also, there are lots of us 'technical' people here, and most of us speak english, some of us speak pretty good english, so you wouldn't have to feel 'isolated'.
I didn't say there aren't any technical people in Mexico. I just said that since I work alone in my home-office I don't get to talk to them very often. And when I do have lunch with colleagues from the Mexican company I used to work at we always speak in Spanish since I speak Spanish much better than they speak English.
So, step outside for a moment, and look around for the guys reading physics or programming books.
Heheh, I've never seen that. :) Then again, I don't lurk around universities either.
But I think you are misunderstanding what I was driving at. I miss being able to "shoot the breeze" in my own native language with technical people. Yes, I can do it in Spanish and sometimes do. And it's not that my Mexican friends don't know their stuff... But it's just not quite the same when it's not in your language.
By the way, you're not the american-looking guy who goes to Vips at San Jeronimo with a programming book, are you?
No, print is dead. :) But seriously, it's not me. What city are you in? I'm in Monterrey but I try to avoid Vips whenever possible. The only thing I can half-tolerate there is Consome de Pollo. :)
And people shouldn't discount the office interaction. Yes, we all work our 8 hours (usually more if you are self-employed), but in the office you have those minutes at the coffee machine, people stopping by your office or cube, or stepping outside to "have a smoke" (even if you don't smoke), or maybe going to lunch with some coworkers. They are things that seem incredibly trivial and often even annoying when you're in an office but you actually start to miss after working alone for awhile.
But I agree with you. Ironically it's a situation where you actually start to look forward to the on-site work--even if it's just meetings. :)
Obviously you didn't RTFA. The problem is when the first 10 results in Google all point to the same company operating with different websites. This has happened to me on occasion, and it's downright annoying. I'm glad Google is taking action to fix this.
About 50% of the time when I'm searching, I AM looking for vendors of a product in order to do price comparisons... Maybe I'm missing something....
Yes... that the first 10 entries where you want to "compare" prices are really the same company so you're going to say "Wow, I guess that's what it costs!" rather than seeing results from other providers that might have a better price.
This isn't about ignoring valid results... this is about ignoring results that have been intentionally and artificially inflated in pursuit of the almighty dollar rather than in the pursuit of providing valuable content to the user.
I agree with you completely. I have no problems with foreigners working in the U.S. I wasn't trying to complain, I was just pointing out that the U.S. is actually quite liberal when it comes to letting foreigners work in the country. More liberal than either Mexico or Colombia that both have much less to protect.
Most software development does require interaction with "end user"--although the end user may be another department in the company.
And, yes, the primary language of the person who programs your dialog boxes does matter since most companies will want the source code that was outsourced and that source code better be readable. Source code can be hard enough to follow when it is written by someone in the same language. I have experience working with source code in English and Spanish. It's not pretty, and I *speak* Spanish.
To answer the article: Yes, I firmly believe that this is the first of what will be many outsourcing projects coming back to the US. I've been saying it for years... Outsourcing to the lowest bidder on the other side of the globe is a "fad" in the post-bubble world of trying to save every cent. I had predicted it'd be about 2004 or 2005 when we start seeing most of the outsourced projects coming back as one company after another found out that outsourcing to 3rd-world countries is not all that it was promised to be.
I'm sure there are some countries that are more open, but Mexico is identical: A company must have no other valid native applicants before they can higher a foreigner. Same with Colombia. There are are 200+ countries in the world and I don't claim to know all of them, but 3 out of 3 (U.S., Mexico, and Colombia) are all the same. Like I said, I just got a Mexican work visa because I was lucky enough to be a on a first-name basis with the Mexican consul in Denver.
I also have a friend that lives and works in the UK. My understanding is that was only possible because the company he worked for wanted him to work in their UK division on a temporary basis. My understanding is that his American ex-wife (they divorced while in UK) is having trouble getting permission to continue to live and work there.
In the U.S., however, the "no valid native applicants" is not very well enforced. That's why you have hundreds of thousands of foreigners running around working in the U.S. on B2 visas. Do you really think there are no Americans that can do the job of these B2 people from India and Pakistan? Of course there are but the foreigners are in the U.S. taking our jobs anyway.
So I still believe the U.S. is certainly no worse than most countries and arguably more lax than some since the U.S. has the same requirements as Mexico and Colombia but really don't enforce it that much.
If you have your mail on a POP server (ISP, hosting provider, etc.) try PrismEmail. It filters between your server and you so there is effectively no time or load on your computer, plus it works with virtually any mail client with nothing to install on the server or on the client.
I'm at 99.9% accuracy so far this month.
I'm sure you trust one or more companies with that same information now for your Internet access and/or mail now.
At my current levels, lets say it averages to 500 emails a day. Multiply * 365 = 182,500. Based on the PrismEmail pricing page, that means I'll need to be in the "Power User" range, even though the vast majority is spam.
Pricing is based on bandwidth, not mail or spam count. Those are provided as conservative examples. Luckily most spams are still less than 20k and you can receive quite a ton of it without going into plans that cost more than $20. I currently receive 250+ spams per day and receive quite a bit of good email, even attaches. I'm not even close to hitting the bandwidth limit of the $20/year plan.
I would have to poll their system. Their system then gets the mail from my POP3 server, then filters it, then sends me the good parts and does whatever it does with the rest. Is that really faster than letting MailWasher help as it does now? I'm not sure it does.
Depends. Is your POP3 currently local to your mail client? If so, it might not be faster. But if your mail is somewhere else then this is done in real-time anyway. It's not like it takes any longer to get the email to your client. The only difference is the mail that does get to your client is only good email.
I'm an American living in Mexico. I was able to get a Mexican work visa pretty much because I was providing Internet access to the Mexican consulate in Denver at the time and was on a first-name basis with him. I mentioned that I'd like to move to Mexico and work there. He made the paperwork happen.
It's not usually that easy. Mexico isn't anxious to give up its jobs to foreigners. I know Colombia isn't either.
For all the crap that gets slung at the U.S., we are still one of the most open when it comes to accept foreigners into our country and letting them live and work here.
Oh, I agree. But if the Feds are going to get involved I'd rather see a technical solution fix the problem rather than a new tax.
At the risk of being flamed for promoting a useful service, try the site in my sig. There is a "Report this as spam" link in every message--click it and you just reported it as spam. Or you can receive a nightly summary of all the spams that were trapped in the last 24 hours and all the mail that was let through... then just quickly click on each one that was miscategorized. No need to go to the website to check for miscategorized mail--just re-categorize it once per day from the nightly mailing.
Bayesian filtering should be done on the server, definitely. It's the only logical place to do it.
What I *would* like to see is a "spam reporting RFC" that defines a protocol whereby email programs can submit "This is a spam" or "This was a false positive" to the server without using HTTP. That'd be very, very useful in opening up a new level of integration between the email client and email server.
Yes, it does work. Let's take your example "G@t G.E.n.E.r.I.C. V1agra" and look at my current corpus:
G: 82.7% spam
E: 88.4% spam
N: 83.1% spam
R: 81.3% spam
I: 66.0% spam
C: 76.1% spam
V1AGRA: 99.9% spam
If you sent me the above spam, it would be caught by my Bayesian filter. If you spelled every word one letter at a time it would be caught by my Bayesian filter. If you used a domain name it would be even easier to catch. If you use an IMG to embed an image rather than spelling out your message you'd have a 96.7% spam score right off the bat. If you embed a JPG iamge you have a 98.5% spam score while a GIF gives you a 95.1% spam score. And don't forget the headers themselves--as Paul Graham mentioned, it's easy to think the headers aren't important but there is a ton of useful information there.
Many people seem to think Bayesian filters aren't the solution or that spammers will get around them. Mark my words, they won't. They can't.
If the Federal Government is going to pass any law that addresses spam, it should be a law that says that every user of email is required to use a Bayesian filter. That would lead to a virtually zero response rate within months which would lead to spam going away. And, wow, all without new taxes!
1. One message had the entire Bill of Rights (1st through 10th Amendments) scattered throughout the spam, in white font. The message still got a spam score of 99.061171%.
2. Another message just blatantly included the following words (also in white text) to try to lower the spam score. The terms they included are listed along with their spam probability in my corpus. NA=Not Applicable because the term has not been used sufficiently to call it spam or good, so it receives a 40% score. OS=Only Spam has used this term, so automatically 99.9% score.
- rainstorm (NA), lufthansa (NA), officio (NA), lullaby (NA), aspect (22.7%),
democracy (OS/99.9%), hotelman (NA), rhodes (NA), roost (NA), embraceable (NA), chattanooga (NA),
austenite (NA), assess (NA), quail (NA), corvette (NA), curia (NA), degenerate (NA),
takeover (OS/99.9%),
brisk (NA), gully (NA), determine (8.9%), condescension (NA), count (12.8%), chevalier (NA),
contributory (NA), importune (NA), complaisant (NA), godhead (NA), taxpayer (NA),
khmer (NA), clothesmen (NA), forum (0.8%), dispel (NA), afterlife (NA), swart (NA), revenue (43.3%),
crucify (NA), abject (NA), imposture (NA), honduras (NA), newsletter (80.0%), hangmen (NA),
digram (NA), inhere (NA), lawmen (NA), expenditure (NA), lord (38.4%), incomplete (8.5%), bedside (NA), armistice (BA), babbitt (NA), acrimony (NA), patsy (NA), adverbial (NA).
A few observations of the above list:- In almost all cases of NA (which means the term did not effect the Bayesian score for the message), the only usages were in spam--which means after just a few more messages these are going to all convert to 99.9% terms. In which case the use of these terms in a future spam will bury it.
- The only term that was really low (forum at 0.8%) is because I run a forum.
- There were two terms that are *ONLY* used in spam (99.9% score), and 1 term that had an 80% score. So by inserting "innocent" text, this spam actually gave me two more terms that were very much spammy. In this case, the two 99.9% terms effectively canceled out their lucky hit on forum (0.8%) and less lucky hit on incomplete (8.5%).
- Even though they hit a single good term (forum), it is really irrelevant. Bayesian doesn't look at ALL terms, it looks at the most INTERESTING terms. That means the most spammy and the least spammy are considered--nothing in between--I use Paul Graham's implementation (15 most interesting terms). As it turns out, almost all of the 15 most interesting terms in this spam had 99.9% ratings, and all of them were above 90%. So, at best, 2 or 3 of the random terms were considered (forum, incomplete, determine). But even so, they were no match for the overwhelmingly spammy words (and HTML tags) used in the spam.
- Result: The message was caught with a spam score of over 90%.
Statistics are fun, and it'll be interesting to see how long it takes spammers to realize that they can't get around Bayesian. Their attempts to get around it, as shown in the above random example, at best are a wash (no effect on its status as spam) or may even INCREASE the spam score since they may just as easily hit spammy words as innocent words.That's the only thing they could possibly be thinking that is somewhat reasonable... but even then I don't think they can make a difference. The terms that Bayesian ends up using to decide something is legitimate mail are going to be the name of your mother, your best friends, the bar where you get together with your friends, etc. It is very unlikely spammers will be able to seed messages that happen to have those words in it--if they did, they could just try sending you spam and use those same words to decrease the spamminess of their message.
The terms they inevitably will end up including in either a seed message or in the spam itself (with the intent of lowering the spamminess) are the terms that are not going to be particularly biased one way or the other, aren't going to be used in determining spaminess, so aren't very useful when it comes to seeding.
Next time I see a spam in my spam folder that has an attempt to evade Bayesian I will run through their "seed" words and see what kind of spam percentages those words have in my corpus. It ought to be interesting.