Depends heavily on your area. I go through Cox@Home in San Diego, and I'm so far pretty satisfied with the speeds. Downloads normally run about 300-400K/sec ( nominal downstream speed is about 3 megabits ). There's occasional dips to 100-200K/sec, but they don't last. The really bad speeds, down in the 20-40K/sec range, are usually associated with specific sites and I think indicate either network problems outside @Home's network or bandwidth or server problems at the site I'm accessing.
Other areas served by Cox seem to have different results. Some are as good as me, some are horrible. So far, though, the worst rates I've seen are still somewhat better than the max rate of a 56k modem, and I never got 56k on a modem.
Perhaps, but two things. First, do you want someone who doesn't know anything about Web servers running your business-critical e-commerce site? Isn't that sort of like having someone who doesn't know anything about flying in the pilot's seat of a 747 full of people at 30,000 feet? Second, if it's proper software you only configure it once. Which costs more for 10 years wear, a pair of boots that cost $100 and last 10 years, or boots that cost $20 a pair and have to be replaced every year?
Which differs from Windows help files how?:)
MS is correcting this problem.:)
One word: skins. You off-load the UI design from the program logic. Then again, I haven't found most O/S programs significantly harder to use than most Windows programs. Yeah Windows sometimes looks prettier, but buries the useful options under 10 levels of menus or on an obscure dialog reachable only from a place completely unrelated to what you want to do. And the UI changes every bloody year.
I think the real reason Dendrite and other companies don't like this ruling is that it turns on defamation. The judge basically said that they've got to show the statements were defamatory before they can use defamation as grounds to discover a defendant's identity. Since the companies aren't interested in proving defamation, just shutting someone up, they don't like this requirement. Tough for them in my book.
As for the ones saying the Constitution doesn't protect anonymous speech, all I can say is "Wrong!". Nowhere in the First Amendment does it require anyone to prove their true identity before they speak. If you want to add that requirement, you know the procedure for amending the Constitution. If you can show the statements made were defamatory or otherwise illegal then the First Amendment no longer applies and you can require proof of identity, but there's this other little provision in the Constitution that says you can't presume someone's guilty until proven otherwise.
Personally I don't think so. Yeah it'll be a pain for those of us who upgrade regularly, but that's the minor part. The real killer's the fact that you have to reactivate every time you reinstall the OS. Think about it. How often is the "solution" to a Windows problem "Reboot, and if that doesn't solve it get your Windows CD and reinstall."? Joe Sixpack's not gonna like it when he's gotta go through the activation every time something goes sour on that shiny new system he paid good money for dammit.
Yep, I've had that with Rat Shack. Odd thing, though. When I've got folding green in my hand and I tell them "I don't see why you need to know that. Now, either let me pay for my stuff or tell me you don't want my business and I'll cheerfully go down to Fry's from now on where they'll be happy to take my money no questions asked.", the cash register suddenly becomes quite happy to take the sale without any information. Money talks, loudly.
They want me to pay blind. In a bookstore or such I can look at the actual book I'm thinking of buying and decide whether I want it or not. A lot of online content providers want me to pay to even be able to find out what information they have. I may be willing to do that for specialized sources where the reputation of the source is enough, but not for things like general news.
Often the charge is higher than the information is worth. Why should I pay $15/month for access to hardware reviews from one of the big magazines when I can get more reliable reviews from other sources for less? A lot of the time this seems to be linked to size: the big sites are charging one fee for access to everything, when all I want is one small part. Mercenary, I know.
The payment systems they want to use are often cumbersome, and require disclosing far too much information for my comfort. If I'm paying for the right to search for and download articles, why do they need to know what my job is or how much I make every year? No real-world business asks that kind of stuff before they'll let me pay for things.
True, but I was also thinking along the lines of "Sure you can pay for one for each one, but that's not a license requirement so if we happen to need a dozen extra PCs configured overnight we can just install the stuff and not have to worry about whether we've bought licenses for them.". Nice point given a certain software company's increased paranoia about licenses.
So, use the price of a distributon CD if they want dollars to assign to it.
Beancounter: So, how much would this new system cost?
Sysadmin: Thirty dollars for the CD with the software on it.
BC: Thirty dollars per seat? That's cheap.
SA: Well, if you want a physical CD for each PC, but you only need one CD for the entire company.
BC: Thirty dollars for an enterprise-wide license? How much do they hit us for upgrades?
SA: Well, minor stuff we just download the upgraded packages from the vendor. Major updates we buy another thirty-dollar CD, but those only happen about once a year.
BC: And we need to pay $BIGNUM per seat for MS software, every year...
SA: Plus retraining people on the new way the new versions work, plus upgrading all the other software that won't work right on the new systems, plus upgrading the hardware because what we've got isn't supported on the new versions...
BC:...
I wouldn't mind text or graphical ads on web sites, banner or box, where the ad graphic was kept reasonably light, say under 40K in size and the total size for ads was less than half the size of the non-ad part of the page.
I don't like ads that take longer to load than the page itself does. I'm looking for the information on the page, so keeping me from getting what I'm looking for is a sure-fire way to send me elsewhere. Ask any shopkeeper in the Real World whether it's a good idea to keep people from getting to the product in the store.
I don't like ads that track who I am and what sites I've visited. If someone physically follows me around all day writing down where I've been, he can expect a visit from the cops before too long.
I don't like ads that jerk control away from me by opening another window. I'm doing something, and that's just as rude as a stranger walking up to you, jerking you around and shoving their favorite screed in your face. Again, ask an RL shopkeeper whether this is a good way to treat customers or not.
I don't like ads that hide themselves. It makes me thing the people behind them feel they have to hide from me. I don't do business with people who I know are hiding things from me. What they're trying to hide might come back to bite me.
Two things strike me here. First, if the lawyers sending the letter are working for Adobe, it's Adobe's bill. The lawyers weren't hired by the person they're billing, that person isn't responsible for paying them. Second, only Adobe can take action about one of their trademarks. If the lawyers are representing Adobe, then aren't they required to indicate this on the legal paperwork? And if they haven't been retained by Adobe and aren't acting for it, then they've no legal basis for their cease-and-desist, right?
The whole thing smells somewhat of an extortion attempt.
The main problem is that it's the spammer's computer that decides ( theoretically based on MX records ) which mail server to connect to. The only way to make this happen is to use multiple hostnames for users, eg. x@open.domain.net vs. x@orbs.domain.net, instead of vanilla x@domain.net. You'd need the servers configured with lists of users they should accept mail for, and some way for users to maintain that list. I'd prefer headers myself, but I'm a firm believer that if an ISP doesn't do procmail I don't want to use that ISP for mail.
The recipients typically can't block mail from open relays. Doing that requires rulesets in the mail server that process based on the IP address the incoming SMTP connection is coming from. That requires root access to the ISP's mail servers. Few ISPs give that access to ordinary users, and gods help the ones that do. And it'd require a mailserver for each user. The best you can do is have the ISP use services like MAPS and ORBS and add headers to the message that users can use to reject mail, and that depends on the users being able to set up procmail or something similar, which isn't feasible for Windows-based users.
Gilmore's own argument works against him. If ISPs have a right to transport mail, then they have a right to not transport mail. Gilmore's going beyond advocating free speech and into the unacceptable to me area of requiring a third party to pay for the hall for the spammers to excercise their free speech in.
Yes, MAPS and ORBS do cut off legitimate mail. If they didn't, then there's be no incentive for anyone to clean up the spam. The recipients of spam who're complaining about it typically aren't customers of the ISPs being used to send the spam, so the ISP loses no money by ignoring their complaints. Only when their customers start complaining because all mail from that ISP is being rejected do the ISPs feel any pressure to shut down the spammers. It'd be nice if it were otherwise and ISPs acted politely, but reality is they don't and we have to live with it.
Well then, I guess you wouldn't have any problem with us recording you and publishing a record of every single place you went every day and how long you spent there, right? After all, there's no camera in your house and you still have your privacy, right? I'm sure you wouldn't mind your wife knowing where you went to pick up her birthday gift, or your boss knowing that you went to the competition for stuff.
Re:Comments from a Bugnosis author
on
Web Bug Detector
·
· Score: 2
NAT makes the IP address ambiguous, yes. That's why I specified that triple instead. To make the triple ambiguous you need to have two people behind the same NAT box hit the same URL within a second or so of each other. That is a lot less likely, and if they hit different URLs then I can match the referrer in my logs against the URL in their logs and disambiguate the sources. Ditto if they hit it at different times. As far as time synchronization, see NTP. Time sync within a few hundred milliseconds isn't hard at all.
As for what the bug's there for, that's the whole point. "Page X on your site was viewed N times by M unique people." is a perfectly legitimate Web-site statistic. So is "Q people followed this path through your site and abandoned it at page B.". In fact a lot of sites could use smacked over the head with that latter statistic, to prove to the salescritters that huge Flash delays, overly-busy and confusing index pages and disruptive intersitial advertisements do indeed make people go elsewhere. Then one comes to Doubleclick and such, who use the same methods to record things like "Person Z browsed these pages on these sites today.". That's getting way past the bounds of acceptable, but it's being done by the same technique.
Just calling it "info transfer" and then saying that all info transfer is bad because some of it's bad is missing the point. The problem isn't that information is being transferred, it's what information is transferred and what's done with it. Dropping the OSDN image, where no personal information or tracking data can be transferred through it because of the way it's coded, into the same category as Doubleclick's bugs, which do transfer a tracking cookie back to a company that's said flat-out that tracking personal habits is their business, is at best disingenuous.
As for why the images are small and transparent, let me ask this: if the only purpose of the image is in fact to collect legitimate site statistics, what purpose does it serve to have it taking up more real-estate on the page and more bandwidth on the network than it absolutely has to to do it's job? Which leads right back to the same problem, that the logical, minimally-disruptive way of doing something legitimate is on the surface identical to what you'd do if you were trying to conceal evil intent. For myself I tend to be quick to block things I don't know, but it annoys me that I have to block or interfere with legitimate things in order to keep out the slime. I'd much rather LART the abusers off the net.
Re:Comments from a Bugnosis author
on
Web Bug Detector
·
· Score: 2
Only point, though: if a site's coding custom Javascript to transfer their cookies to a third-party site, they're planning on synchronizing information in advance already. That or the ad site's handing them cut-and-paste code to use and they aren't checking it, and that can be seen in the HTML source. Pulled-in scripts where the JS in question doesn't appear in the page source won't work, because browsers enforce cookie-domain rules based on the source of the script, not the page pulling it in.
As for difficulty in synchronizing, think about the trio of timestamp, source IP address and referring URL. Off the cuff estimate, I could probably get 95% accuracy from those on any given hit, and over the course of a few hits I'd get effectively 100% accuracy for any given surfer. All automatic once it's created, no effort needed on the part of the operators once the software's installed.
As for what you call Web bugs being only for the info transfer, that depends on what sort of info transfer you're talking about. I can tell you right now that the OSDN 'bug' on Slashdot's pages doesn't do what you're suggesting, so right off there's a counterexample to your assertion. Ditto at least Hitbox's stuff. The only problem is that the illegitimate tracking ones and the legitimage statistics ones look almost identical in the code, until you start really digging into the Javascript ( if any ) and the back-end systems. That's a job that's too complicated usually for a simple plug-in.
As for HTML e-mail bugs, that assumes that a) the user's using a mail reader that renders HTML and b) that mail reader's dumb enough to pull in content not contained in the e-mail message. If your mail reader's a Web browser, then you're obviously open to all the exploits that can be applied to a Web browser. That's why I don't use a Web browser to read mail.:)
Re:Comments from a Bugnosis author
on
Web Bug Detector
·
· Score: 2
Well, yes. But then again, Slashdot could add a module to the Web server that logged their cookie info along with the hit data and timestamps into a file and e-mail that file to anyone they felt like, too. Some shell scripting and a cron job and it'd be completely automatic. That's not Web bugs leaking the information to a third party, that's the main site deliberately giving that information to a third party. I may have concerns about the main site doing that, but Web bugs don't add anything to that concern IMHO seeing as the conduit exists without Web bugs.
And yes, I have thought about that sort of Apache ( and possibly IIS ) module. It's got applications for legitimate site statistics, not just unethical tracking.
Re:Managing cookies in Mozilla 0.9.1
on
Web Bug Detector
·
· Score: 2
Pull up the Tasks menu, Security and Privacy, Cookie Manager, and hit the Cookie Sites tab. Find the sites you want to allow cookies for and remove them from the list of blocked sites.
Re:Comments from a Bugnosis author
on
Web Bug Detector
·
· Score: 2
Actually that about the cookies isn't right. Looking at the OSDN image on Slashdot's page, OSDN can't pick up any slashdot.org cookies from it. Not unless the browser is failing to apply the same-domain rule, that is. You can do some things with Javascript to put osdn.com cookie information into the image request query string, but the OSDN code doesn't do that.
There's also another point. All those Web bugs look identical from an HTML/HTTP point of view, but they're radically different from a data-collection point of view. Hitbox, for example, uses those bugs solely for site statistics. They can tell when two hits were from the same person and can tell a site things like how many people followed a given path through it, but they've no idea who a given person is and don't store any information on which paths a particular person followed in the database the sites access.
Disclaimer: I only program the systems for Hitbox/WebSideStory. I don't represent them or their opinions, they pay the executives to do that.
My take: search engines don't modify my content, (Not)Smart Tags do. Indexing my Web site is fine, as long as you index what I wrote without modifying it. Going in and adding what the user sees as marking in my copyrighted content that I didn't put there is altering my work, which is not legal unless you've negotiated with me and gotten permission.
Now if it's under the user's control, ie. the user and not the software is specifying which words to mark up and where to point the resulting links, I'll call that fair use just like annotating a textbook. If it's done out-of-band, ie. the user marks a word and tells the software to mark it up for them, that's fair use. But when the editorial control over what gets marked up and where the links point to does not rest with the person viewing my site and is not at their direct request, then that's a third party distributing an altered version of my work without permission.
Fundamental difference. Netscape's links are buttons in the browser toolbars and such. They don't intrude on the content. Microsoft's tags aren't buttons in their toolbars, they're presented as markup in my content.
Well, if they use GPL'd code in their application itself, then the license by which they're allowed to use the GPL'd code requires them to put their code under the GPL as well. Same as the license for Microsoft's run-time stuff requires you to follow certain conditions if you want to ( or need to ) distribute it with their applications. You need to keep GPL'd code at arm's length from your application, or follow the code owner's license terms same as with any commercial software ( well, actually most commercial software will force you to follow their terms even if you keep it at arm's length )l.
Then again, I don't think even Ballmer has any problem with being able to impose license terms on code you own. He just doesn't like it when other people require him to abide by those license terms.
The first is a cop-out for engineers too lazy to design the equipment to survive the crash. Such a crash could just as easily leave the GPS signaller functioning but sever the wires that would indicate 'airbag deployed', resulting in a false negative. Better to design the system to continue working after any occupant-survivable crash.
For the second, that assumes that the phone or the central site knows me better than I do. If I'm not hungry, I'm not going to be looking at it's recommendations anyway. If I ask it for Mexican and it responds with Thai, I'm more likely to be annoyed that it wouldn't answer the question I asked it. Just because I love Thai doesn't mean I want Thai right now, and assuming otherwise is pure unwarranted effrontry.
The problem with his article is that most of his examples are invalid. The ambulance-attendant one is the only valid one. The others can be accomplished without giving up any personal information. The car, for example, could transmit it's GPS location only when the airbag did deploy. The phone doesn't need to know my preference in restaurants when I can just ask it for the list of Mexican restaurants within a certain distance of this street corner whenever I'm in the mood for Mexican food.
He starts out by asking what personal information you need to give up to an outside entity to let them do a certain thing. It'd be better to start by asking whether you need to give up any personal information to have them answer a particular question. More often than Scott would apparently like, the answer will be "No, I don't.".
Slashdot Mirror
Depends heavily on your area. I go through Cox@Home in San Diego, and I'm so far pretty satisfied with the speeds. Downloads normally run about 300-400K/sec ( nominal downstream speed is about 3 megabits ). There's occasional dips to 100-200K/sec, but they don't last. The really bad speeds, down in the 20-40K/sec range, are usually associated with specific sites and I think indicate either network problems outside @Home's network or bandwidth or server problems at the site I'm accessing.
Other areas served by Cox seem to have different results. Some are as good as me, some are horrible. So far, though, the worst rates I've seen are still somewhat better than the max rate of a 56k modem, and I never got 56k on a modem.
I think the real reason Dendrite and other companies don't like this ruling is that it turns on defamation. The judge basically said that they've got to show the statements were defamatory before they can use defamation as grounds to discover a defendant's identity. Since the companies aren't interested in proving defamation, just shutting someone up, they don't like this requirement. Tough for them in my book.
As for the ones saying the Constitution doesn't protect anonymous speech, all I can say is "Wrong!". Nowhere in the First Amendment does it require anyone to prove their true identity before they speak. If you want to add that requirement, you know the procedure for amending the Constitution. If you can show the statements made were defamatory or otherwise illegal then the First Amendment no longer applies and you can require proof of identity, but there's this other little provision in the Constitution that says you can't presume someone's guilty until proven otherwise.
Personally I don't think so. Yeah it'll be a pain for those of us who upgrade regularly, but that's the minor part. The real killer's the fact that you have to reactivate every time you reinstall the OS. Think about it. How often is the "solution" to a Windows problem "Reboot, and if that doesn't solve it get your Windows CD and reinstall."? Joe Sixpack's not gonna like it when he's gotta go through the activation every time something goes sour on that shiny new system he paid good money for dammit.
Yep, I've had that with Rat Shack. Odd thing, though. When I've got folding green in my hand and I tell them "I don't see why you need to know that. Now, either let me pay for my stuff or tell me you don't want my business and I'll cheerfully go down to Fry's from now on where they'll be happy to take my money no questions asked.", the cash register suddenly becomes quite happy to take the sale without any information. Money talks, loudly.
True, but I was also thinking along the lines of "Sure you can pay for one for each one, but that's not a license requirement so if we happen to need a dozen extra PCs configured overnight we can just install the stuff and not have to worry about whether we've bought licenses for them.". Nice point given a certain software company's increased paranoia about licenses.
So, use the price of a distributon CD if they want dollars to assign to it.
Beancounter: So, how much would this new system cost? ...
Sysadmin: Thirty dollars for the CD with the software on it.
BC: Thirty dollars per seat? That's cheap.
SA: Well, if you want a physical CD for each PC, but you only need one CD for the entire company.
BC: Thirty dollars for an enterprise-wide license? How much do they hit us for upgrades?
SA: Well, minor stuff we just download the upgraded packages from the vendor. Major updates we buy another thirty-dollar CD, but those only happen about once a year.
BC: And we need to pay $BIGNUM per seat for MS software, every year...
SA: Plus retraining people on the new way the new versions work, plus upgrading all the other software that won't work right on the new systems, plus upgrading the hardware because what we've got isn't supported on the new versions...
BC:
I wouldn't mind text or graphical ads on web sites, banner or box, where the ad graphic was kept reasonably light, say under 40K in size and the total size for ads was less than half the size of the non-ad part of the page.
I don't like ads that take longer to load than the page itself does. I'm looking for the information on the page, so keeping me from getting what I'm looking for is a sure-fire way to send me elsewhere. Ask any shopkeeper in the Real World whether it's a good idea to keep people from getting to the product in the store.
I don't like ads that track who I am and what sites I've visited. If someone physically follows me around all day writing down where I've been, he can expect a visit from the cops before too long.
I don't like ads that jerk control away from me by opening another window. I'm doing something, and that's just as rude as a stranger walking up to you, jerking you around and shoving their favorite screed in your face. Again, ask an RL shopkeeper whether this is a good way to treat customers or not.
I don't like ads that hide themselves. It makes me thing the people behind them feel they have to hide from me. I don't do business with people who I know are hiding things from me. What they're trying to hide might come back to bite me.
Two things strike me here. First, if the lawyers sending the letter are working for Adobe, it's Adobe's bill. The lawyers weren't hired by the person they're billing, that person isn't responsible for paying them. Second, only Adobe can take action about one of their trademarks. If the lawyers are representing Adobe, then aren't they required to indicate this on the legal paperwork? And if they haven't been retained by Adobe and aren't acting for it, then they've no legal basis for their cease-and-desist, right?
The whole thing smells somewhat of an extortion attempt.
The main problem is that it's the spammer's computer that decides ( theoretically based on MX records ) which mail server to connect to. The only way to make this happen is to use multiple hostnames for users, eg. x@open.domain.net vs. x@orbs.domain.net, instead of vanilla x@domain.net. You'd need the servers configured with lists of users they should accept mail for, and some way for users to maintain that list. I'd prefer headers myself, but I'm a firm believer that if an ISP doesn't do procmail I don't want to use that ISP for mail.
Well then, I guess you wouldn't have any problem with us recording you and publishing a record of every single place you went every day and how long you spent there, right? After all, there's no camera in your house and you still have your privacy, right? I'm sure you wouldn't mind your wife knowing where you went to pick up her birthday gift, or your boss knowing that you went to the competition for stuff.
NAT makes the IP address ambiguous, yes. That's why I specified that triple instead. To make the triple ambiguous you need to have two people behind the same NAT box hit the same URL within a second or so of each other. That is a lot less likely, and if they hit different URLs then I can match the referrer in my logs against the URL in their logs and disambiguate the sources. Ditto if they hit it at different times. As far as time synchronization, see NTP. Time sync within a few hundred milliseconds isn't hard at all.
As for what the bug's there for, that's the whole point. "Page X on your site was viewed N times by M unique people." is a perfectly legitimate Web-site statistic. So is "Q people followed this path through your site and abandoned it at page B.". In fact a lot of sites could use smacked over the head with that latter statistic, to prove to the salescritters that huge Flash delays, overly-busy and confusing index pages and disruptive intersitial advertisements do indeed make people go elsewhere. Then one comes to Doubleclick and such, who use the same methods to record things like "Person Z browsed these pages on these sites today.". That's getting way past the bounds of acceptable, but it's being done by the same technique.
Just calling it "info transfer" and then saying that all info transfer is bad because some of it's bad is missing the point. The problem isn't that information is being transferred, it's what information is transferred and what's done with it. Dropping the OSDN image, where no personal information or tracking data can be transferred through it because of the way it's coded, into the same category as Doubleclick's bugs, which do transfer a tracking cookie back to a company that's said flat-out that tracking personal habits is their business, is at best disingenuous.
As for why the images are small and transparent, let me ask this: if the only purpose of the image is in fact to collect legitimate site statistics, what purpose does it serve to have it taking up more real-estate on the page and more bandwidth on the network than it absolutely has to to do it's job? Which leads right back to the same problem, that the logical, minimally-disruptive way of doing something legitimate is on the surface identical to what you'd do if you were trying to conceal evil intent. For myself I tend to be quick to block things I don't know, but it annoys me that I have to block or interfere with legitimate things in order to keep out the slime. I'd much rather LART the abusers off the net.
Only point, though: if a site's coding custom Javascript to transfer their cookies to a third-party site, they're planning on synchronizing information in advance already. That or the ad site's handing them cut-and-paste code to use and they aren't checking it, and that can be seen in the HTML source. Pulled-in scripts where the JS in question doesn't appear in the page source won't work, because browsers enforce cookie-domain rules based on the source of the script, not the page pulling it in.
As for difficulty in synchronizing, think about the trio of timestamp, source IP address and referring URL. Off the cuff estimate, I could probably get 95% accuracy from those on any given hit, and over the course of a few hits I'd get effectively 100% accuracy for any given surfer. All automatic once it's created, no effort needed on the part of the operators once the software's installed.
As for what you call Web bugs being only for the info transfer, that depends on what sort of info transfer you're talking about. I can tell you right now that the OSDN 'bug' on Slashdot's pages doesn't do what you're suggesting, so right off there's a counterexample to your assertion. Ditto at least Hitbox's stuff. The only problem is that the illegitimate tracking ones and the legitimage statistics ones look almost identical in the code, until you start really digging into the Javascript ( if any ) and the back-end systems. That's a job that's too complicated usually for a simple plug-in.
As for HTML e-mail bugs, that assumes that a) the user's using a mail reader that renders HTML and b) that mail reader's dumb enough to pull in content not contained in the e-mail message. If your mail reader's a Web browser, then you're obviously open to all the exploits that can be applied to a Web browser. That's why I don't use a Web browser to read mail. :)
Well, yes. But then again, Slashdot could add a module to the Web server that logged their cookie info along with the hit data and timestamps into a file and e-mail that file to anyone they felt like, too. Some shell scripting and a cron job and it'd be completely automatic. That's not Web bugs leaking the information to a third party, that's the main site deliberately giving that information to a third party. I may have concerns about the main site doing that, but Web bugs don't add anything to that concern IMHO seeing as the conduit exists without Web bugs.
And yes, I have thought about that sort of Apache ( and possibly IIS ) module. It's got applications for legitimate site statistics, not just unethical tracking.
Pull up the Tasks menu, Security and Privacy, Cookie Manager, and hit the Cookie Sites tab. Find the sites you want to allow cookies for and remove them from the list of blocked sites.
Actually that about the cookies isn't right. Looking at the OSDN image on Slashdot's page, OSDN can't pick up any slashdot.org cookies from it. Not unless the browser is failing to apply the same-domain rule, that is. You can do some things with Javascript to put osdn.com cookie information into the image request query string, but the OSDN code doesn't do that.
There's also another point. All those Web bugs look identical from an HTML/HTTP point of view, but they're radically different from a data-collection point of view. Hitbox, for example, uses those bugs solely for site statistics. They can tell when two hits were from the same person and can tell a site things like how many people followed a given path through it, but they've no idea who a given person is and don't store any information on which paths a particular person followed in the database the sites access.
Disclaimer: I only program the systems for Hitbox/WebSideStory. I don't represent them or their opinions, they pay the executives to do that.
My take: search engines don't modify my content, (Not)Smart Tags do. Indexing my Web site is fine, as long as you index what I wrote without modifying it. Going in and adding what the user sees as marking in my copyrighted content that I didn't put there is altering my work, which is not legal unless you've negotiated with me and gotten permission.
Now if it's under the user's control, ie. the user and not the software is specifying which words to mark up and where to point the resulting links, I'll call that fair use just like annotating a textbook. If it's done out-of-band, ie. the user marks a word and tells the software to mark it up for them, that's fair use. But when the editorial control over what gets marked up and where the links point to does not rest with the person viewing my site and is not at their direct request, then that's a third party distributing an altered version of my work without permission.
Fundamental difference. Netscape's links are buttons in the browser toolbars and such. They don't intrude on the content. Microsoft's tags aren't buttons in their toolbars, they're presented as markup in my content.
And www.geocaching.com produces a "Server too busy" error too. Pity about Ed's site, but I can't have too much sympathy for Geocaching. Snicker.
Well, if they use GPL'd code in their application itself, then the license by which they're allowed to use the GPL'd code requires them to put their code under the GPL as well. Same as the license for Microsoft's run-time stuff requires you to follow certain conditions if you want to ( or need to ) distribute it with their applications. You need to keep GPL'd code at arm's length from your application, or follow the code owner's license terms same as with any commercial software ( well, actually most commercial software will force you to follow their terms even if you keep it at arm's length )l.
Then again, I don't think even Ballmer has any problem with being able to impose license terms on code you own. He just doesn't like it when other people require him to abide by those license terms.
The first is a cop-out for engineers too lazy to design the equipment to survive the crash. Such a crash could just as easily leave the GPS signaller functioning but sever the wires that would indicate 'airbag deployed', resulting in a false negative. Better to design the system to continue working after any occupant-survivable crash.
For the second, that assumes that the phone or the central site knows me better than I do. If I'm not hungry, I'm not going to be looking at it's recommendations anyway. If I ask it for Mexican and it responds with Thai, I'm more likely to be annoyed that it wouldn't answer the question I asked it. Just because I love Thai doesn't mean I want Thai right now, and assuming otherwise is pure unwarranted effrontry.
The problem with his article is that most of his examples are invalid. The ambulance-attendant one is the only valid one. The others can be accomplished without giving up any personal information. The car, for example, could transmit it's GPS location only when the airbag did deploy. The phone doesn't need to know my preference in restaurants when I can just ask it for the list of Mexican restaurants within a certain distance of this street corner whenever I'm in the mood for Mexican food.
He starts out by asking what personal information you need to give up to an outside entity to let them do a certain thing. It'd be better to start by asking whether you need to give up any personal information to have them answer a particular question. More often than Scott would apparently like, the answer will be "No, I don't.".