I say that watermarks will always be easy to defeat. They have to already be at a low level in order to not interfere with the content enjoyment. Eventually, it will be known exactly how these watermarks are formed. But that doesn't even have to be known in detail to defeat them. And this defeat only needs to be done for the analog to digital step, unless watermark detection is also implemented elsewhere (the current proposal being to detect it at the analog to digital stage).
An example of such a watermark might be a very slight change of brightness level between frames or groups of frames with sufficient redundancy over time to eventually be detected despite scene brightness change. The codings might even be specifically "fitted" to the brightness changes in content in order to be detected. The key to making this approach work is that over the course of some time, but not over the entire content, there is enough brightness stability averaged out to allow the signal through.
To defeat this, a radical noise signal can be added. It would have spectral components at all frequencies so as to have a greater probability of impacting a watermark of unknown specifications. While this might not be as practical with a hidden camcorder in a theatre, it certainly can be in the confines of your home. Still, an LED injecting luminance behind the lens in a little hole drilled into the camera can do the trick even on a small scale.
Once the recording is made and the watermark detection fooled, you now have to remove the "countermark". But this is easy if the original noise is known. If the noise generator uses a pseudo random number generator with a known initial state, it can be precisely reproduced by replaying from the same initial state. Just get it syncronized (not hard to do in software) and subtract it back out. If linearity/gamma of the analog pickup is an issue, you can simply play the whole "countermark" against a white or gray solid background to measure and compensate for the effects.
Doing this for audio would be similar, but with a different range of spectra and signal levels.
The goal the ultimate watermark protection scheme has to have is to not only prevent attacks like that, but also do so for 100% of those attempting it. Let just one accomplish the task and P2P will see traffic.
You mean kid safe content like disney.kids.us telling kids that being able to copy your own files on your own computer unchecked promotes piracy because that same program can be used to copy copyrighted music or movies?
Last time I was there, the library did not have a policeman preventing kids from exploring other sections of the library. While that may not yet be the case with kids.us, it could easily become so in the next wave of lawmakers on the loose. What if the next law requires parents to identify to an ISP when an account might be used by a child, and penalizes ISPs who allow such accounts to access any site besides kids.us?
At what age do you think that your kids should be free to roam about outside of the kids.us domain? While I certainly think a 7 year old has no need to access outside of what I think kids.us will be, a 14 year old, OTOH, certainly needs full latitude. The problem still exists that to let a child out to access the "whole library", they can also get access to "the alley out back" [goatse.cx].
Probably slightly easier than alternativelifestyles.kids.us. Probably a lot easier than sexualissues.kids.us. Probably a lot harder than disney.kids.us (which could very easily use that site to promote their restrictive legal agendas unopposed).
Who are we going to allow access to having sites in the kids.us domain? Will controversial issues be allowed? Or will corporate policies be allowed to indoctrinate children unimpeded by opposing views? Will disney.kids.us be allowed to present their views on making most open source software illegal, unopposed?
Who is going to decide what is appropriate for kids.us? Would it be appropriate for CERT (as in www.cert.org) to have cert.kids.us? It doesn't really look very kids oriented to me, but some kids might be better prepared for their future careers if they can visit. And what about sites that provide information about issues kids might have to deal with, such as suicide and sexuality? If kids.us becomes the norm for all parents to restrict their children to, then it does become defacto censorship. And what's to prevent the next step in a couple years where parents and/or ISPs are required to deploy the restriction under legal threats? Would you or I as parents even be allowed then to let our children access a site some other parent would not?
Sad, isn't it, that so many managers only hire people they know, and still end up whining (even in this economy) that finding good people is hard to do. Maybe they should start looking for people on the basis of qualifications for a change.
Back when there was a big demand for people and it was hard to find them, the managers doing the hiring didn't do the contact thing you are suggesting. There were in fact, quite many people available, but the communication wasn't there, so they just didn't connect. Maybe the situation today is similar, where so many people are not connecting because they don't go out and pursue contacts.
The fundamental problem, though, is that neither side is really doing it. In good times or bad, communication is the key, and it's not working very well... but mostly because people don't really try to do it. I've been to user groups and vendor demos. There's plenty of people looking for work, but still no jobs. Back when the situation was reversed, there were managers begging for people, and no one willing to change jobs. Still, you gotta try.
We do need something better, though. Trouble is, monster and dice and the like are not doing it. Maybe slashdot could? Who knows. You got any better ideas?
I just cancelled my Earthlink/Mindspring account
on
Disconnecting
·
· Score: 2
I just cancelled my Earthlink/Mindspring account. I called 800-719-4660 and went through 2 levels of voice menu, waited about 20 seconds, and got a live person. After providing my email address, code word, and verifying my address, I asked them to cancel my account. When asked why, I told them it was an issue with tech support (actually there were several such issues but I didn't go into detail). They told me the account was now closed and my credit card removed from it. The call went through smoothly, so we'll see if it actually works.
By comparison, tech support is a mess. The last time I called, the wait time was quoted to be 46 minutes but I was unwilling to wait on the phone that long (maybe I should have, as it was a toll-free line, but I had other things to do). I emailed tech support, and the response I got back over the course of about 4 exchanges suggested the person (different each time) either did not read what I said, or could not read English. That's not an uncommon problem that email tech support has for most companies. Even when talking to people by voice, I after have to go into explanations for them (some actually end up understanding it when I do that). Unfortunately with email support, there is no opportunity to gauge their level of (lack of) understanding, especially because the reply ends up being handled by a different person.
My Earthlink/Mindspring account was not my only account. I have almost a dozen dialup accounts. So it was no big loss (and in fact a savings of $21.95 a month) to drop one. It was my only nationwide dialup account, but I found another company that provides them, so I will probably next try out these guys. They are slightly lower cost, list Linux in their support, and even offer shell accounts. So maybe the move will be worthwhile if these guys are honest.
College is not needed for most computer/IT/internet careers... at least not to be able to do them. I've seen too many "idiots with degrees" to ever believe that college makes much of a difference. And this applies whether the career is programming or system administration or network administration.
What college will do for you is:
Let you "earn" a degree which is useful in the first 2 or 3 years of your career in "getting in the door".
Let you bank some more non-employed, and maybe
even some employed, experience, which can be a plus on your resume.
Let you sit out the economic doldrums we are in right now where all the money people have basically "screwed the pooch" for everyone else. Hopefully things will be better in 4 years.
I may be hiring by the end of the year in a business I'm still trying to get launched. I can tell you this. I'd much rather hire someone coming out of the military than someone coming out of college. Good people can learn new technology. Bad people are stuck in what they managed to learn some of. Learning takes discipline, and you get far more of that in the military than in college. Not everyone coming out would be worth hiring, but even fewer coming out of college will be... and fewer still coming out of high school. I'll be looking more for solidity in understanding basic logic and strategy, then in understanding any particular system. I'd rather hire a smart person with an MCSE (which is otherwise worthless) than a dumb person with an RHCE for doing Linux administration.
That can be done by simple BGP trickery. I know, because I've done it before. What they have is redundant paths through those 3 upstreams, which is certainly better than none. But it just doesn't say one way or the other what locations these are in. Can one fire take out all of.us because 3 pipes converge in the same building? The thing is, we don't know.
It's not Neustar having problems. My changes put in at Verisign (I have one.us domain there) go through quickly just fine. But the changes at register.com (also one.us domain there) are not going through at all. I think register.com is having problesm deploying their software. They did offer to update manually, but so far it hasn't been done correctly.
All the 1 letter and 2 letter domains are reserved, as are many others. I found the list somewhere (forgot where, but I'm sure it's somewhere at Neustar) and put it online here (compressed text is here).
The root servers delegate the.us zone to the Neustar servers, which are:
a.gtld.biz - 209.173.53.162
b.gtld.biz - 209.173.57.162
c.gtld.biz - 209.173.60.65
Recursive lookup starts a one of the servers in root-servers.net, and goes to the above, then finally to your registered DNS servers (yours, or you ISP's, or the registrar's).
Now there are still some problems:
There are only 3 servers handling.us whereas there are 13 for.com, etc.
All 3 servers are in the same netblock, and this suggests maybe even the same physical location.
The TTL (time to live) is set to only 15 minutes at the 3 servers (compared to 2 days at the.com servers... a ratio of 1:192).
If you don't use a.us domain as your DNS servers, then there's an extra lookup because the A-records with the server IPs don't get added in the response (this is how DNS is supposed to work).
I registered 2 domains in.us, one via register.com and the other via Verisign/Netsol. I, too, had trouble getting my DNS servers set up via register.com. They finally said they could do it manually, but then they goofed that up by entering them wrong after giving me an excuse that some of them had no IPs (they did... but they weren't the ones I wanted used, though). OTOH, with Verisign, I was able to change my DNS servers just fine, and it went through in 6 hours (while everyone says it can take 24-72 hours). My only beef with Verisign is the limit of 4 DNS servers. I've had another person also tell me eNom is working OK for them.
Why can't we have (besides the fact that **AA's don't like it) services that offer downloads for a subscription fee, but host all the files themselves? Sure, that does mean they have more costs, and maybe their service will be more expensive. But perhaps some people would prefer that over getting some crappy rip encoded with some crappy software that dumbed down the quality a bit to tighten the file. Compulsory licensing should be just as applicable to this business model, to allow it to be a viable choice in the market and see if it will make it in the market.
Unfortunately, this can break some web based chat sites. But there are so many porn sites that snare browsers with Javascript (technically, the browser is broken if it can success... but we already know IE and NS are broken beyond all hope) that even if a kid is savvy enough to immediately back out of an accidentally encountered porn site, he/she may end up being snatched back in.
My brother currently has the ear of a senator and two congressmen regarding some small businesses issues. I've talked with him about the IP issues in the past, and we've concluded what is needed is a lot more evidence, details, and stories, of what is going on. Substantiated is better, but even if not, it is possible to contact the parties involved and get information. But what is needed is to show a pattern of how the current law and patent issuing process fails. If anyone has collected a list of these cases (who sued who over what patent number) that could be a big help, especially if the numbers are in the thousands.
These things should be done by dlopen(), dlsym(), call a specific function to filter/check data, and dlclose()... instead of fork() and execve(). I've already discovered that dynamic web pages built from shared object files run much faster on the server than corresponding CGI even with the CGI program written in C.
I've been blocking flowgo and euniverse for over a year. These people are totally slimy, and they haven't figured that mail bounces. These people give the human race a bad name.
Slashdot Mirror
I say that watermarks will always be easy to defeat. They have to already be at a low level in order to not interfere with the content enjoyment. Eventually, it will be known exactly how these watermarks are formed. But that doesn't even have to be known in detail to defeat them. And this defeat only needs to be done for the analog to digital step, unless watermark detection is also implemented elsewhere (the current proposal being to detect it at the analog to digital stage).
An example of such a watermark might be a very slight change of brightness level between frames or groups of frames with sufficient redundancy over time to eventually be detected despite scene brightness change. The codings might even be specifically "fitted" to the brightness changes in content in order to be detected. The key to making this approach work is that over the course of some time, but not over the entire content, there is enough brightness stability averaged out to allow the signal through.
To defeat this, a radical noise signal can be added. It would have spectral components at all frequencies so as to have a greater probability of impacting a watermark of unknown specifications. While this might not be as practical with a hidden camcorder in a theatre, it certainly can be in the confines of your home. Still, an LED injecting luminance behind the lens in a little hole drilled into the camera can do the trick even on a small scale.
Once the recording is made and the watermark detection fooled, you now have to remove the "countermark". But this is easy if the original noise is known. If the noise generator uses a pseudo random number generator with a known initial state, it can be precisely reproduced by replaying from the same initial state. Just get it syncronized (not hard to do in software) and subtract it back out. If linearity/gamma of the analog pickup is an issue, you can simply play the whole "countermark" against a white or gray solid background to measure and compensate for the effects.
Doing this for audio would be similar, but with a different range of spectra and signal levels.
The goal the ultimate watermark protection scheme has to have is to not only prevent attacks like that, but also do so for 100% of those attempting it. Let just one accomplish the task and P2P will see traffic.
You mean kid safe content like disney.kids.us telling kids that being able to copy your own files on your own computer unchecked promotes piracy because that same program can be used to copy copyrighted music or movies?
Oh, you mean they're going to have to re-hire some of the web developers they have let go over the past couple of years?
Last time I was there, the library did not have a policeman preventing kids from exploring other sections of the library. While that may not yet be the case with kids.us, it could easily become so in the next wave of lawmakers on the loose. What if the next law requires parents to identify to an ISP when an account might be used by a child, and penalizes ISPs who allow such accounts to access any site besides kids.us?
At what age do you think that your kids should be free to roam about outside of the kids.us domain? While I certainly think a 7 year old has no need to access outside of what I think kids.us will be, a 14 year old, OTOH, certainly needs full latitude. The problem still exists that to let a child out to access the "whole library", they can also get access to "the alley out back" [goatse.cx].
Probably slightly easier than alternativelifestyles.kids.us. Probably a lot easier than sexualissues.kids.us. Probably a lot harder than disney.kids.us (which could very easily use that site to promote their restrictive legal agendas unopposed).
Who are we going to allow access to having sites in the kids.us domain? Will controversial issues be allowed? Or will corporate policies be allowed to indoctrinate children unimpeded by opposing views? Will disney.kids.us be allowed to present their views on making most open source software illegal, unopposed?
Who is going to decide what is appropriate for kids.us? Would it be appropriate for CERT (as in www.cert.org) to have cert.kids.us? It doesn't really look very kids oriented to me, but some kids might be better prepared for their future careers if they can visit. And what about sites that provide information about issues kids might have to deal with, such as suicide and sexuality? If kids.us becomes the norm for all parents to restrict their children to, then it does become defacto censorship. And what's to prevent the next step in a couple years where parents and/or ISPs are required to deploy the restriction under legal threats? Would you or I as parents even be allowed then to let our children access a site some other parent would not?
Sad, isn't it, that so many managers only hire people they know, and still end up whining (even in this economy) that finding good people is hard to do. Maybe they should start looking for people on the basis of qualifications for a change.
Back when there was a big demand for people and it was hard to find them, the managers doing the hiring didn't do the contact thing you are suggesting. There were in fact, quite many people available, but the communication wasn't there, so they just didn't connect. Maybe the situation today is similar, where so many people are not connecting because they don't go out and pursue contacts.
The fundamental problem, though, is that neither side is really doing it. In good times or bad, communication is the key, and it's not working very well ... but mostly because people don't really try to do it. I've been to user groups and vendor demos. There's plenty of people looking for work, but still no jobs. Back when the situation was reversed, there were managers begging for people, and no one willing to change jobs. Still, you gotta try.
We do need something better, though. Trouble is, monster and dice and the like are not doing it. Maybe slashdot could? Who knows. You got any better ideas?
I just cancelled my Earthlink/Mindspring account. I called 800-719-4660 and went through 2 levels of voice menu, waited about 20 seconds, and got a live person. After providing my email address, code word, and verifying my address, I asked them to cancel my account. When asked why, I told them it was an issue with tech support (actually there were several such issues but I didn't go into detail). They told me the account was now closed and my credit card removed from it. The call went through smoothly, so we'll see if it actually works.
By comparison, tech support is a mess. The last time I called, the wait time was quoted to be 46 minutes but I was unwilling to wait on the phone that long (maybe I should have, as it was a toll-free line, but I had other things to do). I emailed tech support, and the response I got back over the course of about 4 exchanges suggested the person (different each time) either did not read what I said, or could not read English. That's not an uncommon problem that email tech support has for most companies. Even when talking to people by voice, I after have to go into explanations for them (some actually end up understanding it when I do that). Unfortunately with email support, there is no opportunity to gauge their level of (lack of) understanding, especially because the reply ends up being handled by a different person.
My Earthlink/Mindspring account was not my only account. I have almost a dozen dialup accounts. So it was no big loss (and in fact a savings of $21.95 a month) to drop one. It was my only nationwide dialup account, but I found another company that provides them, so I will probably next try out these guys. They are slightly lower cost, list Linux in their support, and even offer shell accounts. So maybe the move will be worthwhile if these guys are honest.
College is not needed for most computer/IT/internet careers ... at least not to be able to do them. I've seen too many "idiots with degrees" to ever believe that college makes much of a difference. And this applies whether the career is programming or system administration or network administration.
What college will do for you is:
I may be hiring by the end of the year in a business I'm still trying to get launched. I can tell you this. I'd much rather hire someone coming out of the military than someone coming out of college. Good people can learn new technology. Bad people are stuck in what they managed to learn some of. Learning takes discipline, and you get far more of that in the military than in college. Not everyone coming out would be worth hiring, but even fewer coming out of college will be ... and fewer still coming out of high school. I'll be looking more for solidity in understanding basic logic and strategy, then in understanding any particular system. I'd rather hire a smart person with an MCSE (which is otherwise worthless) than a dumb person with an RHCE for doing Linux administration.
That can be done by simple BGP trickery. I know, because I've done it before. What they have is redundant paths through those 3 upstreams, which is certainly better than none. But it just doesn't say one way or the other what locations these are in. Can one fire take out all of .us because 3 pipes converge in the same building? The thing is, we don't know.
It's not Neustar having problems. My changes put in at Verisign (I have one .us domain there) go through quickly just fine. But the changes at register.com (also one .us domain there) are not going through at all. I think register.com is having problesm deploying their software. They did offer to update manually, but so far it hasn't been done correctly.
All the 1 letter and 2 letter domains are reserved, as are many others. I found the list somewhere (forgot where, but I'm sure it's somewhere at Neustar) and put it online here (compressed text is here).
The root servers delegate the .us zone to the Neustar servers, which are:
Recursive lookup starts a one of the servers in root-servers.net, and goes to the above, then finally to your registered DNS servers (yours, or you ISP's, or the registrar's).
Now there are still some problems:
I registered 2 domains in .us, one via register.com and the other via Verisign/Netsol. I, too, had trouble getting my DNS servers set up via register.com. They finally said they could do it manually, but then they goofed that up by entering them wrong after giving me an excuse that some of them had no IPs (they did ... but they weren't the ones I wanted used, though). OTOH, with Verisign, I was able to change my DNS servers just fine, and it went through in 6 hours (while everyone says it can take 24-72 hours). My only beef with Verisign is the limit of 4 DNS servers. I've had another person also tell me eNom is working OK for them.
Why can't we have (besides the fact that **AA's don't like it) services that offer downloads for a subscription fee, but host all the files themselves? Sure, that does mean they have more costs, and maybe their service will be more expensive. But perhaps some people would prefer that over getting some crappy rip encoded with some crappy software that dumbed down the quality a bit to tighten the file. Compulsory licensing should be just as applicable to this business model, to allow it to be a viable choice in the market and see if it will make it in the market.
Unfortunately, this can break some web based chat sites. But there are so many porn sites that snare browsers with Javascript (technically, the browser is broken if it can success ... but we already know IE and NS are broken beyond all hope) that even if a kid is savvy enough to immediately back out of an accidentally encountered porn site, he/she may end up being snatched back in.
It also says microprocessor. So put your site on an IBM zSeries mainframe running Linux :-)
My brother currently has the ear of a senator and two congressmen regarding some small businesses issues. I've talked with him about the IP issues in the past, and we've concluded what is needed is a lot more evidence, details, and stories, of what is going on. Substantiated is better, but even if not, it is possible to contact the parties involved and get information. But what is needed is to show a pattern of how the current law and patent issuing process fails. If anyone has collected a list of these cases (who sued who over what patent number) that could be a big help, especially if the numbers are in the thousands.
If the universe is expanding, then doesn't that mean there is some measure (even if we don't know what it is) of the size?
It was THE worst linux distro I'd ever seen.
These things should be done by dlopen(), dlsym(), call a specific function to filter/check data, and dlclose() ... instead of fork() and execve(). I've already discovered that dynamic web pages built from shared object files run much faster on the server than corresponding CGI even with the CGI program written in C.
Yes, that would do the job of keeping the mail on the server, and giving a 5XX to the spamware or the open relay. Any word on a Postfix version?
I've been blocking flowgo and euniverse for over a year. These people are totally slimy, and they haven't figured that mail bounces. These people give the human race a bad name.