I'm actually going back to Windows XP / Linux dual-boot, from three years on a Linux-only box. The reason is simple: I want to pick up an MCSD and MCDBA, and it's easier for me to do that with XP and some form of Visual Studio than without. I will probably keep XP there so I can test my software on both Windows/.NET and Linux/Mono, and so that I can play with Palm development. My main motivation for all this is that I want to beef up my career.
Of course, for everything else, including games, email, publishing, music, UNIX development and web surfing, there's Linux.
(VMWare would be nice but there's that $300 price tag.)
The number of offers you get depends on a lot of things: what your experience is, what your skills are, where you live. A noob is less likely to find a job than a 10-year veteran, a Ruby hacker is in less demand than an Oracle Certified DBA, and your job opportunities are much better if you're willing to move anywhere than if you want to maintain a support network.
My employer may have to lay off a lot of employees in the next few months. Personally, if I get laid off again, I'm acting on the assumption that programming work (in the Portland area) will be VERY hard for me to find. I'm teaching myself Visual Basic and C#, getting certifications (some people actually look for them.. I can't afford to get overlooked by these people), and looking for opportunities to work menial type jobs so I can get insurance until the "real job" comes along. Or until I start my own business. Or until I win the lottery.
1) I apologize for saying you were wrong.
2) I will not pay you and I expect no money from you, even if you get 0%. However, I doubt you will get larger than a 2% overblock rate for porn. No, I'm not going to wager on this. The only thing you have to gain by doing this is the right to publish your methology and results, and make me look like an idiot.
Not really, I didn't have time, but that's not a good excuse.
Last night I did some digging around and found that the software in question *was* using RS technology. However, that still doesn't answer the question, since the sample was ***by PeaceFire's own admission*** not a scientific sample, and because the technology has evolved since then. Thus, the 34% number is grossly misleading.
A better review of the technology would be to take a randomly selected sample of non-porn websites and compare it against a randomly selected sample of porn web sites, and see how AOL's parental controls perform. That's the best test of the technology that I could imagine.
He lost his primary precisely because he opposed key anti-privacy provisions of the Homeland Security Act. After that, the GOP targeted him for un-election, and the Rush Limbaugh listening suburbanite dronies of Cobb County out-voted the hard-core, pro-privacy conservatives who used to be Barr's core constituency. The fact that the Georgia Democrats redistricted Barr into Newt's old district didn't help much, either.
Don't get me wrong, I had a lot of bones to pick with Bob Barr, particularly when it came to the religious freedom of our Armed Forces. But I'm sad to see him go, because now we need privacy advocates more than ever before.
Bah.. sorry for the double posting, but here's some new information to make up for it. This information is more current than what PeaceFire has. It basically undermines your argument, but on the other hand, it doesn't contain anything to support my claims. Oh well.
Nice troll. The 34% figure that you presented was collected on 10/23/00, which, to the best of my knowledge, pre-dates their use of RS's technology.
If you really want to test their technology, your best bet would be to do a survey of the web with AOL's parental filters in place. They're not perfect but I've seen the numbers and I doubt that you'll get as high as 34% for false negatives.
Nice troll. Problem is, the 34% error rate report that you presented from Peacefire is dated 10/23/00. To the best of my knowledge this is before RuleSpace became one of their customers.
Sheesh.
Anyway, if you can dig up hard numbers on AOL, feel free to post them.. they've been using RS's technology since mid-May of 2001. As far as I know that's where RS presents its best face.
Of course, there are cultural conservatives (falsely so-called) who are all too happy to do anything they can to prevent your daughter from accessing information on birth control, teen pregnancy, and the like. "Abstainance Only" is the cry for these folks, and it is precisely this kind of legislation that they seek.
(Of course, their politics aren't conservative at all. In fact they're as big-government as it gets.)
Unfournattly that means that who ever is doing the training (even if it isn't overfit etc) is making a judgment call on what is pornographic. What one person/company considers "lewd" is totally subjective. This is also a problem of human backed blacklists.
Right. So part of building a good blacklist, even with AIs, is making sure that you have people with good judgement and a good definition to work from. This is hard (but tractible) for pornography. It's practically impossible for general obscenity, political incorrectness, etc.
There is no existing AI technology that can filter pornography.
Wrong. Pornographic web pages have certain common features that an AI can pick up on, that distinguish them from non-pornographic pages. I can't get into the numbers, but the resulting filters are impressively thorough (low false negatives) and specific (low false positives).
Pornographic photos are an entirely different matter. It'll be ten years before we can filter those.
I submit that you cannot effectivly block porn without also blocking substantial non-porn, unless you have a human-verified blacklist, in which case it merely becomes hugely expensive and unpractical. There's a pretty large amount of porn/sexually explicit content out there without any of the "obvious" triggers that commercial porn sites have, and there's plenty of non-porn sites that will have those triggers.
It's a good guess, but you're mistaken. You can "automate" the human judgement process though neural nets, like AOL does. There are tricks to this that I can't reveal, but the basic jist of it is to let human beings train the 'net and then let the 'net go to town. You can get very good results with this method, which is, by the way, patented.
The example that you cite shouldn't be pornographic, since it's not obviously lewd and obscene. Quite the opposite in fact: it's informational. The display of genitalia alone should not be enough to constitute pornography, since this would be an over-reaching definition, blocking things like photos of Michaelangelo's David.
The problem is, this depends on the product you choose, who created the definitions, how good their QA is, etc. The filters that I worked on were created to let this kind of content through, although on very rare (I can't get into numbers, unfortunately) occasions the pornography filters censored non-pornographic websites.
This fails to address the crux of librarians' and civil libertarians' concerns: the filtering is over-reaching. The worst filters filter by gross word-selection and/or have a political or religious agenda; others just have ill-defined categories for filtered web content. Even the best, however, censor pages that they shouln't, and this is a fundamental problem of the (pattern recognition) technology used to create these filters. The mistakes are potentially embarrassing.
I know all this because I used to work on building these filters.
In short, this strikes me as a poor compromise solution. A much better solution would be for libraries to filter for pornography on ALL computers. Pornography is much easier to filter than most other kinds of objectionable content (again, this is my professional experience in the matter). It also strikes me as a good political solution: on the one hand, even the hardest core civil libertarian can't argue that tax dollars used to support libraries should pay for pornography; on the other hand, this allows just about every other kind of content, including the most controversial (guns, abortion, homosexuality) through, so long as it isn't pornographic.
I think you're underestimating what the party can do. The main thing they need to do is present a more attractive vision than Bush's police state, and convince the voters that they can do it. That's really not as hard as it looks. If the economy continues its slump, or there's a change in the Supreme Court, then that makes their job all the easier.
I agree that they can't afford to lose more votes to the Greens. It killed them in Florida in 2000, and it probably did even more damage in 2002 (as the liberals were not inclined to support the moderate New Democrat agenda). The party can go Green without much difficulty, though.
I don't see them calling for election reform. By all accounts this was the cleanest election we had in years. I don't see things getting worse in that regard between now and 2004.
USAPATRIOT passed a Democratic-controlled Senate that still followed the Clinton strategy of "move the party to the Right."
I think in the future, you will not see the Democrats so ready to cooperate. Of course, that depends on what happens within the party in the next few weeks.
Privacy is also the cornerstone of the right to abortion and birth control. Regardless of how you feel about these things, the significance of loss of privacy in this matter can't be overestimated.
Slashdot Mirror
Sweet. I'm doing the same thing more or less for my own church, except they can't pay me squat. :)
Of course, for everything else, including games, email, publishing, music, UNIX development and web surfing, there's Linux.
(VMWare would be nice but there's that $300 price tag.)
My employer may have to lay off a lot of employees in the next few months. Personally, if I get laid off again, I'm acting on the assumption that programming work (in the Portland area) will be VERY hard for me to find. I'm teaching myself Visual Basic and C#, getting certifications (some people actually look for them .. I can't afford to get overlooked by these people), and looking for opportunities to work menial type jobs so I can get insurance until the "real job" comes along. Or until I start my own business. Or until I win the lottery.
1) I apologize for saying you were wrong. 2) I will not pay you and I expect no money from you, even if you get 0%. However, I doubt you will get larger than a 2% overblock rate for porn. No, I'm not going to wager on this. The only thing you have to gain by doing this is the right to publish your methology and results, and make me look like an idiot.
Not really, I didn't have time, but that's not a good excuse.
Last night I did some digging around and found that the software in question *was* using RS technology. However, that still doesn't answer the question, since the sample was ***by PeaceFire's own admission*** not a scientific sample, and because the technology has evolved since then. Thus, the 34% number is grossly misleading.
A better review of the technology would be to take a randomly selected sample of non-porn websites and compare it against a randomly selected sample of porn web sites, and see how AOL's parental controls perform. That's the best test of the technology that I could imagine.
We do that in Oregon too. Only here, we call them ballot measures.
Don't get me wrong, I had a lot of bones to pick with Bob Barr, particularly when it came to the religious freedom of our Armed Forces. But I'm sad to see him go, because now we need privacy advocates more than ever before.
Why, hello mister Vice Admiral!
Try again later.
Two jobs ago I was coding Java servlets. Got laid off.
One job ago I was coding Perl and C++. Got laid off.
This job I'm coding COBOL (using Synon) on miniframes.
Trust me, COBOL isn't going anywhere.
Bah .. sorry for the double posting, but here's some new information to make up for it. This information is more current than what PeaceFire has. It basically undermines your argument, but on the other hand, it doesn't contain anything to support my claims. Oh well.
If you really want to test their technology, your best bet would be to do a survey of the web with AOL's parental filters in place. They're not perfect but I've seen the numbers and I doubt that you'll get as high as 34% for false negatives.
Sheesh.
Anyway, if you can dig up hard numbers on AOL, feel free to post them .. they've been using RS's technology since mid-May of 2001. As far as I know that's where RS presents its best face.
(Of course, their politics aren't conservative at all. In fact they're as big-government as it gets.)
Right. So part of building a good blacklist, even with AIs, is making sure that you have people with good judgement and a good definition to work from. This is hard (but tractible) for pornography. It's practically impossible for general obscenity, political incorrectness, etc.
Wrong. Pornographic web pages have certain common features that an AI can pick up on, that distinguish them from non-pornographic pages. I can't get into the numbers, but the resulting filters are impressively thorough (low false negatives) and specific (low false positives).
Pornographic photos are an entirely different matter. It'll be ten years before we can filter those.
It's a good guess, but you're mistaken. You can "automate" the human judgement process though neural nets, like AOL does. There are tricks to this that I can't reveal, but the basic jist of it is to let human beings train the 'net and then let the 'net go to town. You can get very good results with this method, which is, by the way, patented.
The problem is, this depends on the product you choose, who created the definitions, how good their QA is, etc. The filters that I worked on were created to let this kind of content through, although on very rare (I can't get into numbers, unfortunately) occasions the pornography filters censored non-pornographic websites.
I know all this because I used to work on building these filters.
In short, this strikes me as a poor compromise solution. A much better solution would be for libraries to filter for pornography on ALL computers. Pornography is much easier to filter than most other kinds of objectionable content (again, this is my professional experience in the matter). It also strikes me as a good political solution: on the one hand, even the hardest core civil libertarian can't argue that tax dollars used to support libraries should pay for pornography; on the other hand, this allows just about every other kind of content, including the most controversial (guns, abortion, homosexuality) through, so long as it isn't pornographic.
If you think the GOP won't support this piece of anti-consumer, pro-big-business legislation, then you probably haven't been paying attention.
bareman++, and shame on both parties for chickening out and not taking this stand.
Yes, except that the little girl tells you to buy Windows XP.
I agree that they can't afford to lose more votes to the Greens. It killed them in Florida in 2000, and it probably did even more damage in 2002 (as the liberals were not inclined to support the moderate New Democrat agenda). The party can go Green without much difficulty, though.
I don't see them calling for election reform. By all accounts this was the cleanest election we had in years. I don't see things getting worse in that regard between now and 2004.
I think in the future, you will not see the Democrats so ready to cooperate. Of course, that depends on what happens within the party in the next few weeks.
Privacy is also the cornerstone of the right to abortion and birth control. Regardless of how you feel about these things, the significance of loss of privacy in this matter can't be overestimated.