If you Ask Jeeves, "what is the air-speed velocity of an unladen swallow?"
then he will ask you the proper question in response, and you can click on it to see where the hell this quote comes from.
The advantages of a good distribution network.
on
NYTimes Looks at Warez
·
· Score: 3, Insightful
It's obvious that what these guys were doing is illegal. Still I feel sorry for them, with their multi-year prison sentences, because they really weren't costing the software industry that much money in lost sales, and because they are scapegoats.
As many others have said, most people wouldn't have bought the very expensive applications anyhow. When someone makes a pirated copy of Photoshop to do web graphics, at worst, they are depriving The GIMP community of a new user, or depriving Jasc of $99 -- usually not depriving Adobe of $600. There is some financial impact on the industry, but the numbers are lower. Also, there are plenty of software copiers. Software "theft" won't be reduced one iota by locking these guys up.
The reason for that is, they were just functioning as a completely essential part of a healthy information economy -- the underground. Why is it essential? One reason is that, espescially near the turning points in society and revolution, information occasionally must transcend barriers created by law. If these underground data networks -- very small ones, if you believe the numbers in the NYT article -- are maintained, hidden, and keep working based on an economy of commercially available pilfered information, and if more citizens are trained in how to communicate covertly, and people are indoctrinated to know that storing or exchanging illegal information may not actually be wrong, then our surveillance-laden society has paid a fair price.
The loosely hierarchical distribution network used by warez kidz is analogous in form and function to those used in China and other repressive regimes by political dissidents. Capable of passing only information, peer-organized, and with a medium level of identity isolation -- bring down one and you bring down a few others, but not the whole group. Personally, I feel more secure knowing that there exist these sophisticated illegal networks, capable only of traffic in information, that would be rather difficult for any authority to completely shut down. Who knows when they may be needed...
-=Ivan (actually not very paranoid at all)
"Here are a few notes from the underground / load them at your pleasure / These are the dusty pictures that I found / while on my search for treasure" -- Information Society: Mirrorshades
Nah, you're reasonable. I didn't even look at your other posts. But maybe you could put your commercial links/ads in a sig or whatever. Anyhow, keep posting your opinions about being an independant software developer. I've got zero problems with a developer deciding to use whatever licensing terms they like. And keep shamelessly promoting too if you think it's worthwhile, just be subtle.
[ As for my license,... the part about honoring copyright, well, it still seems reasonable to me. ]
Honoring copyright is one thing. But I don't agree that all instances of unauthorized copying/display/etc of copyrighted works necessarily are an infringement of anyone's rights. That kind of 'simplification' is leading to the erosion of fair use rights. And according to your contract, if I disagree with that statement (it is part of the terms) I am not authorized to use the software.
OK I am rotflmao at that line so give me a second here.:-)...
I don't think anyone was complaining about advertising in general, as much as commenting on your use of the/. message boards to advertise your commercial software product. I agree that you have to hustle if you are independant. Keep that business card with you, you never know who you will meet. But advertising on these forums could be interpreted as spam.
Also, this is not exactly the Artistic License. Maybe your product is worth whatever you're charging, and maybe not. But you're no public service organization:
You agree not to copy or distribute this software or change the source code without prior written permission... Unauthorized copying, distribution, modification, public display, or public performance of copyrighted works is an infringement of the copyright holders' rights. If you disagree with any of these terms, you are not authorized to use this software... [emphasis added]
Well, I wouldn't donate towards your product development in order to support the community's effort, or to show my distaste for the RIAA's interpretation of copyright law! As another poster suggested, JukeX seems like a much more viable alternative from those perspectives.
Come on, you don't hear positive reports about the software because there's no news to be had in it! Here's the thing. Nobody is really keeping the block list a secret. You can find out how any particular site is listed using the web link at the top of the article. You just can't see the whole list at once, because the company feels the need to protect its intellectual property. This is so that the income they recieve for providing the control list service will allow them to pay wages of human editors to categorize their control list.
You talk about the "censors that keep the block lists secret" -- but do you think the employees who categorize websites make choices about whether the control list contents should be released? Their opinions about the validity of the company's intellectual property are not the subject of this conversation.
And yes, I do think they deserve to be cut a little slack now and then.
It is just too bad that schools use such products at all. Government funded filtering is censorship, no matter who carries it out. It's the same as pulling controversial books off of school library shelves -- it just shouldn't be done.
Any list containing a ton of URL's supposedly classified by category is going to contain some mistakes. A few erroneous submissions that aren't noticed until much later could drastically decrease the percieved accuracy of this kind of product. That doesn't matter too much to a corporate IT department who wants to be able to say that they are doing everything they can to reduce sexual harassment liability and increase productivity. Perhaps it shouldn't matter to them, if those are the priorities. (Personally I think those responsibilities should be handled by management, not IT)
But I certainly hope that this kind of censorship matters to parents, and educators, and librarians.
I used to work for Secure Computing Corporation. I was a programmer on SmartFilter. And when I found out they were actively trying to get S.97 (the mandatory library censorship bill) passed, I turned in my resignation.
However, at that time, they had just come under new management, after the prior CEO and his cronies decieved the employees, fleeced the company, then left. The new management listened to what I had to say. They approved of my changing category definitions (removing STD and safe sex info from the Sex category, for instance). They agreed that supporting S.97 and library censorship (though I don't think they'd use that word) would probably not impress their enterprise clients. Of course, it was too late -- S.97 was going to pass anyhow. It did.
I recommend that you activists out there focus your efforts on trying to get CIPA ruled unconstitutional, rather than focusing on (doing something to? flaming?) one of the many companies that provides a filtering product. You may think that Secure Computing's past support of S.97 is unforgivable, but at least it was done under different management: their current line is that the product is made for corporations, a fairly consistant message throughout their website.
If you read his later post, you will note that his figure was only measuring accuracy *of blocked sites*, not total accesses. If blocked sites were actually 1 percent of requests (just a random guess) then 0.03 percent of requests are incorrectly blocked.
I liked your analysis, so I'll rerun the numbers. Given a 3 percent error on blocked sites, and 1 percent of requests are blocked, accessing 24 sites:
(1.00-0.0003)^24 = 99.3 percent chance of unimpeded access along with accurate blocking, 0.7 percent chance of being unfairly blocked. That's a lot better.
On the other hand, the success rate would be even closer to 100 percent if the subject quits their job and goes to work someplace that doesn't put obnoxious filtering software between them and the web.
[ (wager is the guys name, no _human_ whould miss that!) ]
OK, but the URL is actually "wwager" which would not be caught by any self respecting word search algorithm. Still, if you spend long enough testing URL's in an unscientific fashion, you will eventually find one URL to support your hypothesis that SmartFilter uses weak wordsearch algorithms to classify Gambling sites. Even if they did, do you *really* think they used an algorithm that classifies every site that contains the word "wager" as a gambling site? That's insane. If you think it's really that simple, do you think that the other half-million pages on Google with the word "wager" are all categorized as Gambling by SmartFilter?
On the other hand, perhaps that one URL was just a random mistake, like SourceForge being classed as an MP3 site. Someone clicked the wrong button and didn't notice. Or maybe even some "creative" ex-employee figured out how to batch-import a bunch of search engine results a few years ago, trying to save some time or whatever. (I would still consider that "operator error")
[ Maby you could have a friend post and admit they made the error the article gave as an example?:) ]
Would you, after such a shitstorm? Anyhow, I already suggested in another post that the company post a statement on/., it is up to the company if they are even reading this.
[ just because there are (ahumn...) _2^30_ combinations, does'nt mean you have to search 2^30 space to find the answer. ]
*turns beet red* Yes, I meant to say 2^30. Brain fart. Anyhow, that means that the final result would have 2^30 possible outcomes. If you divide it up into 30 operations, allegedly using some sort of word search algorithm for each category, you increase the probability of inaccuracy 30 times, effectively searching the entire 2^30 answerspace and leaving precious little chance that you would have any sort of accuracy in the end.
The CensorWare report talked about over 15 million pages (53 million URLs), accessed from school property, of which around 1 percent were blocked, and a few hundred of those inaccurately so. The raw numbers in that report were once used by Secure Computing spinmen to show that SmartFilter was astoundingly accurate.
However, that report also contains a lot of complete fabrications by the author regarding how SmartFilter categorizes sites! The writer didn't make even the slightest attempt to fact-check the allegations. Take their explanation for the "william wager" site -- they just made it all up based on their own guesses about the URL. They didn't ask Secure Computing for info, such as how certain sites were categorized, because they weren't even trying to be impartial.
I'm sorry to say that the CensorWare report on SmartFilter is unscientific crap. Really sorry, because I too believe that SmartFilter should never come within 100 yards of a school or government funded institution. But fabricating facts about how it works is NOT the right way to stop the government buying it!
[ "This document last updated on Thursday September 07 2000." ]
The web page containing the article to which you are referring was "last updated" then. The article was written in 1999, using data from 1997 IIRC.
-=Ivan
(disclosure: A long time ago, I used to write code for SmartFilter. Of course nothing in this message is confidential. All facts are suspect.)
[Actually, my stat is based on real numbers in our environment.]
Alright, I'll take you at your word. But given your methodology -- you are not measuring the accuracy of the filter, just the accuracy exhibited when it blocks a site (i.e. you are not comparing to the total number of URLs that were not blocked). That is, you're not measuring when it correctly does not block a site, only when it incorrectly blocks one. Which may be valid, depending on your purpose, but certainly serves to drastically inflate the percentage figure as well.
[I don't know enough about search technology to argue the point well, but what you're suggesting is that a group of humans could do it instead? I don't really think that makes much sense...]
I think it makes sense -- the problem is, define which of 30 categories site X falls into. There are 2^30 possible answers. A trained human can look at the site, and using his/her brain, make subjective decisions about the categories and quickly mark some checkboxes. Automated bots (AIs) can only provide loose suggestions at best, when the problem is this complex and subjective. Computers are not generally very good at solving problems of that sort.
On the other hand, you can use all sorts of different inputs to categorization for one specific category. For example, you could tell a bot to periodically index some public web page with a large list of certain types of website (for example, persiankitty) and then auto-classify links that appear on that page.
The filtering companies have to be careful about how they SAY they do filtering, from a marketing perspective. If they say that they only use bots, customers will be concerned about the intelligence of the bots. If they say that they only use real people, customers may think the product is not as efficient as it could be. If they tell you exactly what they do, they are giving away their intellectual property.
So keep in mind that whatever SurfControl, SmartFilter, etc is telling you about how they classify sites, it has probably been "filtered" through a marketroid.
[I'm suggesting it got catagorized automatically with no checks, or someone wasn't paying attention when they clicked the button.]
I would definitely lean towards the latter explanation. I doubt SmartFilter would try to use a text search heuristic for categories like MP3 *without human verification*.
-=Ivan
(disclosure: I used to write code for SmartFilter, but that was a long time ago. There's nothing confidential here. I don't represent anyone's opinions but my own. As usual, all facts are suspect.)
What article are you talking about? The one on Peacefire that is 5 years old, out of date and littered with factual inaccuracy? Or the ancient report on CensorWare.org that is simply chock full of propaganda? Read my post again, and turn on your mind first.
Now that your mind is hopefully on -- how well do you think an automated system (an AI) is going to classify millions of websites into 30 categories? That's 30^2 combinations. I would be surprised if such a system EVER got one page right. Therefore, they MUST use humans. And they do. Your implication that I did not "read the article" is just as incorrect as your implication that the links are not categorized by humans. Turns out I actually know a couple of the humans that do this categorizing, so I'm pretty sure I didn't just dream it up.
The funniest thing Secure Computing did regarding the original CensorWare report was to officially respond stating something like "CensorWare's statistics proved SmartFilter was more than 99.9 percent accurate." It was 99.99 percent hilarious, because both sides were using some seriously twisted "new math" to calculate accuracy in terms of a single number that could be politicized and used to support their own viewpoints.
Now, I'm sorry to laugh at the CensorWare folks, because in a way I support much of their cause. This product shouldn't be used in schools, at all, and shouldn't be financially supported by our government! But I have a 6th sense for propaganda, and I don't like it when people play with numbers and play with the English language, while claiming that they are sticking to the "facts."
"Nuf said."
-=Ivan
(disclosure: I wrote code for SmartFilter a long time ago. There's nothing confidential in this message. Opinions expressed herein are mine alone. All facts are suspect.)
First of all, let me point out that your 3-5 percent error rate is probably a made up number (97.4 percent of all statistics are made up on the spot and yours is no exception!).:-)
Secondly, while all the list-based filter software companies probably are using search engines to find new sites (if they're smart), some of them like SmartFilter actually have a human categorizing the web pages. SF has something like 30 different categories. I guarantee you that, given a modest budget, you cannot write an AI that will class millions of web sites into 30 content categories (2^30 combinations) based on fuzzy human definitions. As far as I know, they hire a team of people to review the sites. Maybe some things are really obvious, like the occurrance of hundreds of 4 letter sex words in the meta tags might automatically flag it as a porn site. But many sites would have to be reviewed by hand.
You are correct to point out that system administrators can choose the blocking policy (even blocking nothing, and just logging category results). They can exempt sites, etc. But you forgot one really obvious thing. This mistake by SmartFilter was probably just that -- a mistake, an accident, human fallibility, operator error. Someone clicked the wrong button, oops, to be fixed in the next control list. Would be interesting to hear from some of the SmartFilter folks on this. Anyone out there? Is there an official company response to this issue?
-=Ivan
(disclosure: I wrote code for SmartFilter a long time ago. There's nothing confidential in this post. Facts herein are subject to my own human fallibility.)
OK, so Squid is not hard to configure. With commercial products like SmartFilter, you can block millions of porn sites by checking a box. Their customers are paying for the service of having a team of people and machines examining new sites as they pop up across the world.
With the open-source solutions, you can block whatever sites you want to type in, or you can rely on someone else's (probably unpaid and irregular) labor to track new porn sites as they pop up. When your source dries up, you might find another.
Now, I fully support open source software and use it whereever I can. But if you want to really actively block a high percentage of porn URL's, without using heuristic [over]blocking, you're going to have to get access to a really really big and frequently-updated list, which is really the service being sold by SmartFilter.
One more thing, none of the open source solutions for blocking that I've seen can even approach the raw URL's per second that can be handled by SmartFilter. They stopped supporting the standalone UNIX proxy version a LONG time ago, now you just plug it into your existing firewall or proxy (and reserve a few MB's of RAM for its huge hash tables). In my opinion, the reason open-source web filters are not as fast is probably just because developers don't spend a lot of time writing extremely efficient open-source list-based proxy server access control mechanisms. There's no money in it, and they probably have better things to do...
-=Ivan
(disclosure: I used to write code for SmartFilter a long time ago. There's no confidential info in this post. Opinions in this post don't represent anyone's views except my own. However, John McNulty, the CEO of Secure Computing, once agreed with me -- after I provided a convincing argument -- that SmartFilter is indeed designed for corporations, not education.)
You have seen that it is very inconvenient for some people when there is a mistake in these control lists. If you submit a site with incorrect category information, it seems like you are *trying* to make the product more inaccurate. Such tactics will NEVER stop SmartFilter from selling their product, but you might make them SLOWER to respond to requests to fix actual mistakes.
Thus, individuals who are behind such firewalls will have an even worse time of it. And eventually, SmartFilter may just take down the web form if it causes too much trouble. That would help the situation, right? No. No, it wouldn't.
It is not a control list editor's or programmer's fault that corporations go out looking to buy filtering software. It is not SF's fault that your stupid, fascist employer decides to buy this software and subject you to it. It's not their fault that you can't figure out how to circumvent the filter. (of course, that would be impossible)
> It's about impossible to block the > web on a page by page basis
Actually, that's not true. SmartFilter (and most other proxy-based blockers that require a list of sites to operate) can selectively control access to domains, servers or their IP addresses, subdirectories, pages or individual files. The policy about what categories are blocked will be determined by the IT administrator installing the filter.
The "editors" that create the control lists used by these products (actual human editors in the case of SmartFilter) require this kind of flexibility in order to categorize the web as accurately as possible.
Geez, only one person needs to submit the link to ensure that it will be reviewed again. 1000 people doing so isn't going to do much except flood the submission system with duplicates (which will probably be dequeued before they are seen by the reviewers, anyhow).
Listing SourceForge.net in the "MP3" category was almost certainly an accident. Secure Computing/Smartfilter has been very quick to resolve such issues in the past, typically providing automatic updates within a week or less.
Finally, if you want 100 percent accurate filtering software, you might as well give up right now. The nice thing about SmartFilter, if there is anything nice about any of these products, is that the links are reviewed and categorized by humans -- who are good, and trained, but not completely infallible. While processing thousands of sites, someone might hit the wrong button now and then. It's not a conspiracy, folks.
System admins who are frustrated by requests to un-block the site should simply add it to their local exemption file, at least until they recieve the next update to the control list.
-=Ivan
(disclosure: I used to work there a long time ago. There's no confidential information in this post. This message doesn't represent their official views or policies or anything. All facts stated in this message are potentially subtly incorrect.)
Re:I had a pet virus once.
on
Pet Bugs?
·
· Score: 1
Yes. That's 100 percent correct.
ALL boot sector viruses are required by ANSI (surely you've heard them, they are the infamous cabal of hackers that invented the "ANSI bomb") to infect floppies within 5 seconds of accessing the disk. Since virus authors are required by Canadian law to adhere to these standards, it seems VERY doubtful that someone would write a boot sector virus that did not conform to the specification.
...
Some of my favorite and bugs along this line: * The bug in 1813/Jerusalem-B that caused it to re-infect the same files over and over again. * The bug in MacMag (the advertising virus!) that caused it to crash the MacII after infection. * The bug in Cascade that caused it to infect systems with IBM BIOS's even though it tried to avoid them.
I see your point. However, it did serve to establish the institute's record and qualifications. You kind of have to be a team of computer security experts, like these ADTI folks, in order to formulate a reasonable opinion on such a complicated and specialized subject.
In fact, there are groups of specialists that deal with issues such as "security by obscurity" and "peer analysis," and can make decisions like whether they are good... or evil! It's clear that ADTI's researchers, perhaps even their IT contractors and "power users," are among these proud few.
Indeed, it was this very institution's namesake, Alexis de Tocqueville, that once wrote, "We are superior to the beasts in this, that we use our souls to find out those material benefits to which they are only led by instinct."
You're way off there. US public libraries are not kicking back a dime (or even a fraction of a penny) to the author on a per-checkout basis. Why would they, when funds are limited and there is no law compelling them to do so?
Authors only get royalties from a library for copies made, and even then only in "non-fair-use" circumstances. If you think I'm wrong, show me a shred of evidence that public libraries are paying authors royalties each time a book is checked out. It's simply untrue. Even in Canada.
As for buying from a local used bookstore, I would check there before going to Amazon if possible, simply because I believe in supporting local commerce 1st. But what I'm interpreting from your statements is that you are anti-Amazon for other reasons -- patent abuse, etc. -- and are perhaps jumping on the (rather small) bandwagon that's condemning Amazon for doing something that is actually a really good and useful idea.
Your argument that "used is theft" flies in the face of the first sale doctrine, which is perhaps all that keeps libraries and used bookstores going today. Just as used record stores need not pay the publishing company for a second license to the record, these other great institutions are free to buy and sell works for which the license fee is already paid.
About your fictitious (or perhaps just very tiny) boycott of Amazon, who the hell are the "Writers of America"? OK, I am aware that the American Publishers Guild and the Author's Guild have issued a joint whine about Amazon putting too much emphasis on good customer service by helping people sell their old books... Which reminds me -- I am more likely to buy an unknown book if I know that I can resell it for a good price later! Those same organizations also complained loudly about Amazon's policy of publishing customer reviews.
But back to the point. In the US, there is no "license" to read a book or lend one. Nor is any license needed to buy or sell a copy of a book. You only need a license for the RIGHT to make a COPY. That's what copyright (right to copy) is about, or at least, what it's supposed to be about. The DMCA weakens this, although I suppose you're a big fan of that too.
As for library kick-backs.. Good choice of words. Maybe that happens in Canada, where record companies bribed weak-minded government officials to give the RIAA free money for sales of *BLANK* audio tapes. But I still think you're way off there: as far as I know, even Canadian libraries only pay royalties on items that they actually COPY, with numerous "fair dealing" exceptions (similar to "fair use" in the USA). I'm certain that my local library doesn't waste my tax money by paying copy royalties that have already been paid.
What country are you from?!? It certainly can be illegal and an abuse of power to charge someone with a crime that they actually are committing.
Ever heard of the Equal Protection Act? A conviction cannot be based upon state activity that denies an accused equal protection of the law. Wright v. Georgia, 373 U.S. 284 (1963).
Oh yeah! I leeched a bunch of Purple Motion's songs -- get this -- available as MOD files! crazy old school trackers. check it out. man, it's been so long since I watched 2nd Reality.. *rock*
... that is, unless a large electrostatic discharge is released across your skin, through the Palm and Cradle, and into one of the non-grounded pins connected to your UART. Which is exactly what this guy is talking about. This is not impossible, you know. Sensitive electronic equipment gets killed by static shocks all the time.
However, there may be something about the design of the particular UART that makes it more susceptible to this, or there might be something about the design of the Palm or Cradle that makes it more likely. But don't dismiss a problem report simply because you don't understand it.
(I don't usually respond to trolls. But this is an insidious meme that must be fought.)
The reason for your lack of power is that you are apathetic. You say that other people do not have the power to change the world, but that is not true. You can not deny that change happens. And it must always start with an individual.
But you attempt to discourage others from working for change, so that they will become impotent like you. You are using what little power you have to try to make this negative change occur. That's not a very nice thing to do.
OK.. but why not vote too, in addition to any other tactics you take to effect change? It does not cost you anything except a few minutes of your time. And even if your chosen candidates or laws do not win, you have grown in your knowledge about the elected officials and laws of your home.
If voting is like gambling, then living in a democracy and not voting is very much like hanging out in a casino, getting a free $5 credit and throwing it away, even though you stay in the casino.
Please, don't vote if you don't want to or don't care. But there's no need to discourage others from learning and participating in the process. By doing so, you are promoting apathy, not change. If people participate, they may learn that there are things wrong with the system, and they may try to change it or at least believe that it needs to be changed. If they are discouraged and do not participate, how can they avoid apathy? Will they care about changing the system if they feel that they are not a part of it anyhow?
Do you really care about changing the system or are you just promoting a rather anti-social version of anarchy?
Slashdot Mirror
If you Ask Jeeves,
"what is the air-speed velocity of an unladen swallow?"
then he will ask you the proper question in response, and you can click on it to see where the hell this quote comes from.
It's obvious that what these guys were doing is illegal. Still I feel sorry for them, with their multi-year prison sentences, because they really weren't costing the software industry that much money in lost sales, and because they are scapegoats.
As many others have said, most people wouldn't have bought the very expensive applications anyhow. When someone makes a pirated copy of Photoshop to do web graphics, at worst, they are depriving The GIMP community of a new user, or depriving Jasc of $99 -- usually not depriving Adobe of $600. There is some financial impact on the industry, but the numbers are lower. Also, there are plenty of software copiers. Software "theft" won't be reduced one iota by locking these guys up.
The reason for that is, they were just functioning as a completely essential part of a healthy information economy -- the underground. Why is it essential? One reason is that, espescially near the turning points in society and revolution, information occasionally must transcend barriers created by law. If these underground data networks -- very small ones, if you believe the numbers in the NYT article -- are maintained, hidden, and keep working based on an economy of commercially available pilfered information, and if more citizens are trained in how to communicate covertly, and people are indoctrinated to know that storing or exchanging illegal information may not actually be wrong, then our surveillance-laden society has paid a fair price.
The loosely hierarchical distribution network used by warez kidz is analogous in form and function to those used in China and other repressive regimes by political dissidents. Capable of passing only information, peer-organized, and with a medium level of identity isolation -- bring down one and you bring down a few others, but not the whole group. Personally, I feel more secure knowing that there exist these sophisticated illegal networks, capable only of traffic in information, that would be rather difficult for any authority to completely shut down. Who knows when they may be needed...
-=Ivan (actually not very paranoid at all)
"Here are a few notes from the underground / load them at your pleasure / These are the dusty pictures that I found / while on my search for treasure" -- Information Society: Mirrorshades
Nah, you're reasonable. I didn't even look at your other posts. But maybe you could put your commercial links/ads in a sig or whatever. Anyhow, keep posting your opinions about being an independant software developer. I've got zero problems with a developer deciding to use whatever licensing terms they like. And keep shamelessly promoting too if you think it's worthwhile, just be subtle.
... the part about honoring copyright, well, it still seems reasonable to me. ]
[ As for my license,
Honoring copyright is one thing. But I don't agree that all instances of unauthorized copying/display/etc of copyrighted works necessarily are an infringement of anyone's rights. That kind of 'simplification' is leading to the erosion of fair use rights. And according to your contract, if I disagree with that statement (it is part of the terms) I am not authorized to use the software.
-=Ivan
OK I am rotflmao at that line so give me a second here.
I don't think anyone was complaining about advertising in general, as much as commenting on your use of the
Also, this is not exactly the Artistic License. Maybe your product is worth whatever you're charging, and maybe not. But you're no public service organization: Well, I wouldn't donate towards your product development in order to support the community's effort, or to show my distaste for the RIAA's interpretation of copyright law! As another poster suggested, JukeX seems like a much more viable alternative from those perspectives.
-=Ivan
Come on, you don't hear positive reports about the software because there's no news to be had in it! Here's the thing. Nobody is really keeping the block list a secret. You can find out how any particular site is listed using the web link at the top of the article. You just can't see the whole list at once, because the company feels the need to protect its intellectual property. This is so that the income they recieve for providing the control list service will allow them to pay wages of human editors to categorize their control list.
You talk about the "censors that keep the block lists secret" -- but do you think the employees who categorize websites make choices about whether the control list contents should be released? Their opinions about the validity of the company's intellectual property are not the subject of this conversation.
And yes, I do think they deserve to be cut a little slack now and then.
-=Ivan
It is just too bad that schools use such products at all. Government funded filtering is censorship, no matter who carries it out. It's the same as pulling controversial books off of school library shelves -- it just shouldn't be done.
Any list containing a ton of URL's supposedly classified by category is going to contain some mistakes. A few erroneous submissions that aren't noticed until much later could drastically decrease the percieved accuracy of this kind of product. That doesn't matter too much to a corporate IT department who wants to be able to say that they are doing everything they can to reduce sexual harassment liability and increase productivity. Perhaps it shouldn't matter to them, if those are the priorities. (Personally I think those responsibilities should be handled by management, not IT)
But I certainly hope that this kind of censorship matters to parents, and educators, and librarians.
I used to work for Secure Computing Corporation. I was a programmer on SmartFilter. And when I found out they were actively trying to get S.97 (the mandatory library censorship bill) passed, I turned in my resignation.
However, at that time, they had just come under new management, after the prior CEO and his cronies decieved the employees, fleeced the company, then left. The new management listened to what I had to say. They approved of my changing category definitions (removing STD and safe sex info from the Sex category, for instance). They agreed that supporting S.97 and library censorship (though I don't think they'd use that word) would probably not impress their enterprise clients. Of course, it was too late -- S.97 was going to pass anyhow. It did.
I recommend that you activists out there focus your efforts on trying to get CIPA ruled unconstitutional, rather than focusing on (doing something to? flaming?) one of the many companies that provides a filtering product. You may think that Secure Computing's past support of S.97 is unforgivable, but at least it was done under different management: their current line is that the product is made for corporations, a fairly consistant message throughout their website.
-=Ivan
If you read his later post, you will note that his figure was only measuring accuracy *of blocked sites*, not total accesses. If blocked sites were actually 1 percent of requests (just a random guess) then 0.03 percent of requests are incorrectly blocked.
I liked your analysis, so I'll rerun the numbers. Given a 3 percent error on blocked sites, and 1 percent of requests are blocked, accessing 24 sites:
(1.00-0.0003)^24 = 99.3 percent chance of unimpeded access along with accurate blocking, 0.7 percent chance of being unfairly blocked. That's a lot better.
On the other hand, the success rate would be even closer to 100 percent if the subject quits their job and goes to work someplace that doesn't put obnoxious filtering software between them and the web.
-=Ivan
[ (wager is the guys name, no _human_ whould miss that!) ]
:) ]
/., it is up to the company if they are even reading this.
OK, but the URL is actually "wwager" which would not be caught by any self respecting word search algorithm. Still, if you spend long enough testing URL's in an unscientific fashion, you will eventually find one URL to support your hypothesis that SmartFilter uses weak wordsearch algorithms to classify Gambling sites. Even if they did, do you *really* think they used an algorithm that classifies every site that contains the word "wager" as a gambling site? That's insane. If you think it's really that simple, do you think that the other half-million pages on Google with the word "wager" are all categorized as Gambling by SmartFilter?
On the other hand, perhaps that one URL was just a random mistake, like SourceForge being classed as an MP3 site. Someone clicked the wrong button and didn't notice. Or maybe even some "creative" ex-employee figured out how to batch-import a bunch of search engine results a few years ago, trying to save some time or whatever. (I would still consider that "operator error")
[ Maby you could have a friend post and admit they made the error the article gave as an example?
Would you, after such a shitstorm? Anyhow, I already suggested in another post that the company post a statement on
[ just because there are (ahumn...) _2^30_ combinations, does'nt mean you have to search 2^30 space to find the answer. ]
*turns beet red* Yes, I meant to say 2^30. Brain fart. Anyhow, that means that the final result would have 2^30 possible outcomes. If you divide it up into 30 operations, allegedly using some sort of word search algorithm for each category, you increase the probability of inaccuracy 30 times, effectively searching the entire 2^30 answerspace and leaving precious little chance that you would have any sort of accuracy in the end.
The CensorWare report talked about over 15 million pages (53 million URLs), accessed from school property, of which around 1 percent were blocked, and a few hundred of those inaccurately so. The raw numbers in that report were once used by Secure Computing spinmen to show that SmartFilter was astoundingly accurate.
However, that report also contains a lot of complete fabrications by the author regarding how SmartFilter categorizes sites! The writer didn't make even the slightest attempt to fact-check the allegations. Take their explanation for the "william wager" site -- they just made it all up based on their own guesses about the URL. They didn't ask Secure Computing for info, such as how certain sites were categorized, because they weren't even trying to be impartial.
I'm sorry to say that the CensorWare report on SmartFilter is unscientific crap. Really sorry, because I too believe that SmartFilter should never come within 100 yards of a school or government funded institution. But fabricating facts about how it works is NOT the right way to stop the government buying it!
[ "This document last updated on Thursday September 07 2000." ]
The web page containing the article to which you are referring was "last updated" then. The article was written in 1999, using data from 1997 IIRC.
-=Ivan
(disclosure: A long time ago, I used to write code for SmartFilter. Of course nothing in this message is confidential. All facts are suspect.)
[Actually, my stat is based on real numbers in our environment.]
Alright, I'll take you at your word. But given your methodology -- you are not measuring the accuracy of the filter, just the accuracy exhibited when it blocks a site (i.e. you are not comparing to the total number of URLs that were not blocked). That is, you're not measuring when it correctly does not block a site, only when it incorrectly blocks one. Which may be valid, depending on your purpose, but certainly serves to drastically inflate the percentage figure as well.
[I don't know enough about search technology to argue the point well, but what you're suggesting is that a group of humans could do it instead? I don't really think that makes much sense...]
I think it makes sense -- the problem is, define which of 30 categories site X falls into. There are 2^30 possible answers. A trained human can look at the site, and using his/her brain, make subjective decisions about the categories and quickly mark some checkboxes. Automated bots (AIs) can only provide loose suggestions at best, when the problem is this complex and subjective. Computers are not generally very good at solving problems of that sort.
On the other hand, you can use all sorts of different inputs to categorization for one specific category. For example, you could tell a bot to periodically index some public web page with a large list of certain types of website (for example, persiankitty) and then auto-classify links that appear on that page.
The filtering companies have to be careful about how they SAY they do filtering, from a marketing perspective. If they say that they only use bots, customers will be concerned about the intelligence of the bots. If they say that they only use real people, customers may think the product is not as efficient as it could be. If they tell you exactly what they do, they are giving away their intellectual property.
So keep in mind that whatever SurfControl, SmartFilter, etc is telling you about how they classify sites, it has probably been "filtered" through a marketroid.
[I'm suggesting it got catagorized automatically with no checks, or someone wasn't paying attention when they clicked the button.]
I would definitely lean towards the latter explanation. I doubt SmartFilter would try to use a text search heuristic for categories like MP3 *without human verification*.
-=Ivan
(disclosure: I used to write code for SmartFilter, but that was a long time ago. There's nothing confidential here. I don't represent anyone's opinions but my own. As usual, all facts are suspect.)
What article are you talking about? The one on Peacefire that is 5 years old, out of date and littered with factual inaccuracy? Or the ancient report on CensorWare.org that is simply chock full of propaganda? Read my post again, and turn on your mind first.
Now that your mind is hopefully on -- how well do you think an automated system (an AI) is going to classify millions of websites into 30 categories? That's 30^2 combinations. I would be surprised if such a system EVER got one page right. Therefore, they MUST use humans. And they do. Your implication that I did not "read the article" is just as incorrect as your implication that the links are not categorized by humans. Turns out I actually know a couple of the humans that do this categorizing, so I'm pretty sure I didn't just dream it up.
The funniest thing Secure Computing did regarding the original CensorWare report was to officially respond stating something like "CensorWare's statistics proved SmartFilter was more than 99.9 percent accurate." It was 99.99 percent hilarious, because both sides were using some seriously twisted "new math" to calculate accuracy in terms of a single number that could be politicized and used to support their own viewpoints.
Now, I'm sorry to laugh at the CensorWare folks, because in a way I support much of their cause. This product shouldn't be used in schools, at all, and shouldn't be financially supported by our government! But I have a 6th sense for propaganda, and I don't like it when people play with numbers and play with the English language, while claiming that they are sticking to the "facts."
"Nuf said."
-=Ivan
(disclosure: I wrote code for SmartFilter a long time ago. There's nothing confidential in this message. Opinions expressed herein are mine alone. All facts are suspect.)
First of all, let me point out that your 3-5 percent error rate is probably a made up number (97.4 percent of all statistics are made up on the spot and yours is no exception!). :-)
Secondly, while all the list-based filter software companies probably are using search engines to find new sites (if they're smart), some of them like SmartFilter actually have a human categorizing the web pages. SF has something like 30 different categories. I guarantee you that, given a modest budget, you cannot write an AI that will class millions of web sites into 30 content categories (2^30 combinations) based on fuzzy human definitions. As far as I know, they hire a team of people to review the sites. Maybe some things are really obvious, like the occurrance of hundreds of 4 letter sex words in the meta tags might automatically flag it as a porn site. But many sites would have to be reviewed by hand.
You are correct to point out that system administrators can choose the blocking policy (even blocking nothing, and just logging category results). They can exempt sites, etc. But you forgot one really obvious thing. This mistake by SmartFilter was probably just that -- a mistake, an accident, human fallibility, operator error. Someone clicked the wrong button, oops, to be fixed in the next control list. Would be interesting to hear from some of the SmartFilter folks on this. Anyone out there? Is there an official company response to this issue?
-=Ivan
(disclosure: I wrote code for SmartFilter a long time ago. There's nothing confidential in this post. Facts herein are subject to my own human fallibility.)
OK, so Squid is not hard to configure. With commercial products like SmartFilter, you can block millions of porn sites by checking a box. Their customers are paying for the service of having a team of people and machines examining new sites as they pop up across the world.
With the open-source solutions, you can block whatever sites you want to type in, or you can rely on someone else's (probably unpaid and irregular) labor to track new porn sites as they pop up. When your source dries up, you might find another.
Now, I fully support open source software and use it whereever I can. But if you want to really actively block a high percentage of porn URL's, without using heuristic [over]blocking, you're going to have to get access to a really really big and frequently-updated list, which is really the service being sold by SmartFilter.
One more thing, none of the open source solutions for blocking that I've seen can even approach the raw URL's per second that can be handled by SmartFilter. They stopped supporting the standalone UNIX proxy version a LONG time ago, now you just plug it into your existing firewall or proxy (and reserve a few MB's of RAM for its huge hash tables). In my opinion, the reason open-source web filters are not as fast is probably just because developers don't spend a lot of time writing extremely efficient open-source list-based proxy server access control mechanisms. There's no money in it, and they probably have better things to do...
-=Ivan
(disclosure: I used to write code for SmartFilter a long time ago. There's no confidential info in this post. Opinions in this post don't represent anyone's views except my own. However, John McNulty, the CEO of Secure Computing, once agreed with me -- after I provided a convincing argument -- that SmartFilter is indeed designed for corporations, not education.)
You have seen that it is very inconvenient for some people when there is a mistake in these control lists. If you submit a site with incorrect category information, it seems like you are *trying* to make the product more inaccurate. Such tactics will NEVER stop SmartFilter from selling their product, but you might make them SLOWER to respond to requests to fix actual mistakes.
Thus, individuals who are behind such firewalls will have an even worse time of it. And eventually, SmartFilter may just take down the web form if it causes too much trouble. That would help the situation, right? No. No, it wouldn't.
It is not a control list editor's or programmer's fault that corporations go out looking to buy filtering software. It is not SF's fault that your stupid, fascist employer decides to buy this software and subject you to it. It's not their fault that you can't figure out how to circumvent the filter. (of course, that would be impossible)
-=Ivan
> It's about impossible to block the
> web on a page by page basis
Actually, that's not true. SmartFilter (and most other proxy-based blockers that require a list of sites to operate) can selectively control access to domains, servers or their IP addresses, subdirectories, pages or individual files. The policy about what categories are blocked will be determined by the IT administrator installing the filter.
The "editors" that create the control lists used by these products (actual human editors in the case of SmartFilter) require this kind of flexibility in order to categorize the web as accurately as possible.
-=Ivan
Geez, only one person needs to submit the link to ensure that it will be reviewed again. 1000 people doing so isn't going to do much except flood the submission system with duplicates (which will probably be dequeued before they are seen by the reviewers, anyhow).
Listing SourceForge.net in the "MP3" category was almost certainly an accident. Secure Computing/Smartfilter has been very quick to resolve such issues in the past, typically providing automatic updates within a week or less.
Finally, if you want 100 percent accurate filtering software, you might as well give up right now. The nice thing about SmartFilter, if there is anything nice about any of these products, is that the links are reviewed and categorized by humans -- who are good, and trained, but not completely infallible. While processing thousands of sites, someone might hit the wrong button now and then. It's not a conspiracy, folks.
System admins who are frustrated by requests to un-block the site should simply add it to their local exemption file, at least until they recieve the next update to the control list.
-=Ivan
(disclosure: I used to work there a long time ago. There's no confidential information in this post. This message doesn't represent their official views or policies or anything. All facts stated in this message are potentially subtly incorrect.)
Blade Communications Inc.
541 N. Superior St.
Toledo, OH 43660
Phone: 419-724-6000
Fax: 419-724-6080
Online: Web Site
Mr. Paul W. Shryock
Buckeye Cablevision, Inc.
5566 Southwyck Blvd.
Toledo, Ohio 43607
Fax: 419-724-7074
" You're welcome to try Buckeye Express for yourself in our Customer Service lobby at 5566 Southwyck Blvd [Toledo OH]. "
Buckeye info: 419-724-9800
Tech Support: 419-724-3278
-=Ivan
Yes. That's 100 percent correct.
ALL boot sector viruses are required by ANSI (surely you've heard them, they are the infamous cabal of hackers that invented the "ANSI bomb") to infect floppies within 5 seconds of accessing the disk. Since virus authors are required by Canadian law to adhere to these standards, it seems VERY doubtful that someone would write a boot sector virus that did not conform to the specification.
...
Some of my favorite and bugs along this line:
* The bug in 1813/Jerusalem-B that caused it to re-infect the same files over and over again.
* The bug in MacMag (the advertising virus!) that caused it to crash the MacII after infection.
* The bug in Cascade that caused it to infect systems with IBM BIOS's even though it tried to avoid them.
-=Ivan
> Nice to see no politics being spouted here.
I see your point. However, it did serve to establish the institute's record and qualifications. You kind of have to be a team of computer security experts, like these ADTI folks, in order to formulate a reasonable opinion on such a complicated and specialized subject.
In fact, there are groups of specialists that deal with issues such as "security by obscurity" and "peer analysis," and can make decisions like whether they are good... or evil! It's clear that ADTI's researchers, perhaps even their IT contractors and "power users," are among these proud few.
Indeed, it was this very institution's namesake, Alexis de Tocqueville, that once wrote, "We are superior to the beasts in this, that we use our souls to find out those material benefits to which they are only led by instinct."
-=Ivan
You're way off there. US public libraries are not kicking back a dime (or even a fraction of a penny) to the author on a per-checkout basis. Why would they, when funds are limited and there is no law compelling them to do so?
Authors only get royalties from a library for copies made, and even then only in "non-fair-use" circumstances. If you think I'm wrong, show me a shred of evidence that public libraries are paying authors royalties each time a book is checked out. It's simply untrue. Even in Canada.
As for buying from a local used bookstore, I would check there before going to Amazon if possible, simply because I believe in supporting local commerce 1st. But what I'm interpreting from your statements is that you are anti-Amazon for other reasons -- patent abuse, etc. -- and are perhaps jumping on the (rather small) bandwagon that's condemning Amazon for doing something that is actually a really good and useful idea.
-=Ivan
Your argument that "used is theft" flies in the face of the first sale doctrine, which is perhaps all that keeps libraries and used bookstores going today. Just as used record stores need not pay the publishing company for a second license to the record, these other great institutions are free to buy and sell works for which the license fee is already paid.
About your fictitious (or perhaps just very tiny) boycott of Amazon, who the hell are the "Writers of America"? OK, I am aware that the American Publishers Guild and the Author's Guild have issued a joint whine about Amazon putting too much emphasis on good customer service by helping people sell their old books... Which reminds me -- I am more likely to buy an unknown book if I know that I can resell it for a good price later! Those same organizations also complained loudly about Amazon's policy of publishing customer reviews.
But back to the point. In the US, there is no "license" to read a book or lend one. Nor is any license needed to buy or sell a copy of a book. You only need a license for the RIGHT to make a COPY. That's what copyright (right to copy) is about, or at least, what it's supposed to be about. The DMCA weakens this, although I suppose you're a big fan of that too.
As for library kick-backs.. Good choice of words. Maybe that happens in Canada, where record companies bribed weak-minded government officials to give the RIAA free money for sales of *BLANK* audio tapes. But I still think you're way off there: as far as I know, even Canadian libraries only pay royalties on items that they actually COPY, with numerous "fair dealing" exceptions (similar to "fair use" in the USA). I'm certain that my local library doesn't waste my tax money by paying copy royalties that have already been paid.
-=Ivan
What country are you from?!? It certainly can be illegal and an abuse of power to charge someone with a crime that they actually are committing.
Ever heard of the Equal Protection Act? A conviction cannot be based upon state activity that denies an accused equal protection of the law. Wright v. Georgia, 373 U.S. 284 (1963).
Oh yeah! I leeched a bunch of Purple Motion's songs -- get this -- available as MOD files! crazy old school trackers. check it out. man, it's been so long since I watched 2nd Reality.. *rock*
http://www.modarchive.com/artists/pm/
http://www.modarchive.com/artists/skaven/
... that is, unless a large electrostatic discharge is released across your skin, through the Palm and Cradle, and into one of the non-grounded pins connected to your UART. Which is exactly what this guy is talking about. This is not impossible, you know. Sensitive electronic equipment gets killed by static shocks all the time.
However, there may be something about the design of the particular UART that makes it more susceptible to this, or there might be something about the design of the Palm or Cradle that makes it more likely. But don't dismiss a problem report simply because you don't understand it.
(I don't usually respond to trolls. But this is an insidious meme that must be fought.)
The reason for your lack of power is that you are apathetic. You say that other people do not have the power to change the world, but that is not true. You can not deny that change happens. And it must always start with an individual.
But you attempt to discourage others from working for change, so that they will become impotent like you. You are using what little power you have to try to make this negative change occur. That's not a very nice thing to do.
-=Ivan
OK.. but why not vote too, in addition to any other tactics you take to effect change? It does not cost you anything except a few minutes of your time. And even if your chosen candidates or laws do not win, you have grown in your knowledge about the elected officials and laws of your home.
If voting is like gambling, then living in a democracy and not voting is very much like hanging out in a casino, getting a free $5 credit and throwing it away, even though you stay in the casino.
Please, don't vote if you don't want to or don't care. But there's no need to discourage others from learning and participating in the process. By doing so, you are promoting apathy, not change. If people participate, they may learn that there are things wrong with the system, and they may try to change it or at least believe that it needs to be changed. If they are discouraged and do not participate, how can they avoid apathy? Will they care about changing the system if they feel that they are not a part of it anyhow?
Do you really care about changing the system or are you just promoting a rather anti-social version of anarchy?