"terrorists hate our intrusive foreign policy that installs dictatorships, topples governments, crushes dissent, exploits and degrades the region, and prevents autonomous governance."
Well, Firefox OS is the only one only using Web technologies. With the other two, you might have been using one of the other options for any number of reasons, like for instance that C++ was faster, or your boss thought that JavaScript was a toy language, or that the API you actually needed for the app was missing in the web app version, and then that cut you out of reuse.
What would be even better is if all these manufacturers got together and worked out a common API. Then developers wouldn't have to write a different app for each device. It can't really be that hard.
I don't want to piss on your parade, but actually that is pretty hard and nobody has really done any kind of work on that yet. All that they have done (and what this project promises) is to do digital modelling of what we know to be analogue systems. There's no talk of actually designing analogue microprocessors and hooking them up in massively parallel, biologically plausible configurations. All the work on analogue systems stopped in the 60s. The huge progress we've made on von Neumann machines doesn't actually buy us anything here.
It represents approximately 1% of the number of neurons, nodes in the system. In the human brain each neuron is connected to, on average, 7000 other neurons. So you would not have to build a computer 100x faster in 10 years. You would have to build a computer approximately 10000x faster. This is what I mean when I say there is a scaling factor involved, and we might not be quite there with the hardware yet.
As far as I am aware current VLSI technology can be used to model on the order of 10 billion synapses. The human brain has on the order of 100 trillion synapses. Unless Henry Markham has also invented a radically new kind of supercomputer, we are still somewhat behind.
I can't decompose (decompile?) it because I can't download it.
I can't download it because downloading it would mean making a copy of it.
I can't make a copy of it because that would infringe the author's copyright since they have not published it under a licence which gives me the right to make such a copy.
As far as contacting the authors is concerned, I see little point. I'm hardly going to trust my security to a developer who can't even install an HTTPS certificate correctly, let alone distribute under a licence that allows distribution.
James Clapper, Director of National Intelligence, says that...the NSA tried to penetrate and compromise Tor, but it was only because terrorists and criminals use it, too...
Well, he's right. As far as that goes.
Terrorists and criminals also breathe air and drink water.
So you think that I have something that I wish to communicate secretly, and I am going to do so using a piece of software that I can't see the source code to?
Not sure what millenium you're living in there bro.
So some website has hundreds of mugshot photos up on their site. What kind of person is going to be interested in leafing through those the whole time? I can't see how they could generate any interest whatsoever, let alone enough that you risk being shunned by anyone you actually know.
Your anonymity in a dead drop system depends on the dead drop location being known only to you and to the person with whom you want to exchange the secret.
As soon as you publish the location of the dead drop anyone can observe it and you have no anonymity whatsoever.
How about, if you have pictures of another person, given with reasonable expectations that it was for your private use (i.e. you do not have any signed permission to the contrary), why not just make it illegal to make these public with intent to bully, defame, humiliate, or shame?
I take a picture of you covertly dumping toxic waste into a reservoir.
I take a picture of a policeman using excessive force on a witness.
I take a picture of a politician accepting a bribe to close a park and turn it into condos.
Unfortunately it is now illegal for me to publish any of these photos since they might humiliate the person they depict.
In this world, people don't negotiate contracts for each communication they make with one another. It would hardly be scalable to do so.
Also, I don't think that many people on here are considering the fact many women have pictures taken of them that they are unaware of (drunk, sleeping, hidden cam, etc.), and the type of person that would take those kinds of pictures is the exact type of person that would post them.
That's not actually the issue with revenge porn. The item under consideration is where someone takes a compromising selfie and sends it to their partner while they are in a relationship. Then the relationship ends badly and the recipient, feeling hard done by, publishes the selfie with the intent to distress the sender.
'Revenge porn' is nearly always copyright violation. Clearly the 'model' has not signed off on this usage, and in the case of 'sexting' in particular, the 'model' is usually the 'photographer' too and therefore has ALL publication and distribution rights to the image, not just the model rights.
In the "GMail wiretapping" case most Slashdotters seemed to be of the opinion that once you have sent someone a message, the message belongs to the recipient not the author.
If the sending host is a zombie, it won't get through.
It will authenticate as the hijackee, so it will get through.
It's a chicken and egg problem with no obvious solution since the Internet lacks the ability to "require" any particular participation.
"The Internet" is not to blame. As I said before the solution is unclear because people are lazy and not motivated enough to fix the problem.
Here's a potential motivation: a massive player like Google decides to advertise a "spam free" version of GMail, or a consortium of such players. They only deliver from valid SPF sources with an audited policy of disconnecting any hosts found to participate in botnets. Want to be able to reach all those customers?
None of these solutions you proposed are even remotely practical or effective and there are far better ideas (I'll be it still impractical from an implementation standpoint) out there. In fact, the technology already exists, but isn't currently implemented universally enough to be turned on. Thinks like authenticated mail servers would allow for mail to be required to be sent from the authoritative server for that domain with cryptographic verification that the server is authentic, (based on the DNS records for the domain). A SPAM server would not have access to these and would have to setup their own domain (which could then be trivially blocked). The reason it isn't used is that not enough domains have it setup and we don't want to block a significant portion of the legitimate mail traffic on the Internet.
Right, the technology you are describing is called SPF. As you say it's not universally used, but I disagree that you wouldn't want to absolutely require it as a first step. It's totally trivial to use it and there's no reason not to, so just require it and any stragglers will soon fall in when they realize their mail won't be delivered otherwise.
A major problem with SPF is that it isn't going to have any effect if the sending host is a zombie. Keep those ideas coming!
Does Google do something more advanced than this for marketing purposes?
Unlikely. The current situation represents the state of the art; if Google had invented a better technology for this it would have patented it.
Do we really want to limit the ability to use a better SPAM filter in the future?
A better means of suppressing spam would be to prevent it or limit it at its source - that's the only way to free up the bandwidth, storage, and processing power that currently has to deal with it.
One possibility is making it cost a small amount of actual money to send an email - micropayments, say $0.0001 per email. This would add up to such tiny amounts for ordinary users that they wouldn't really notice. For businesses it would be covered by operating costs, but for spammers it would be crippling.
Another possibility is to make it computationally expensive to send an email to each recipient. This might not be very effective since spammers often hijack ordinary users' computers to do the spamming.
Yet another solution would be to require that all email be end to end encrypted: you'd have to have the public key of the person you want to send to. Obtaining people's public keys should be reasonably straightforward but not necessarily easily automatable - in that case the cost of discovery might be too high for spammers.
It's unlikely that any of these strategies will be implemented though as people are lazy and don't actually care enough about fixing the problem, otherwise it would have happened by now.
You have to extract meaning to perform SPAM filtering.
No you don't, and spam filters currently don't do this as it's extremely hard. Spam filters (that examine message contents) use probability theory based on individual instances of words in the corpus.
Slashdot Mirror
All the ones on Windows, anyway.
I see the ketamine is kicking in...
I doubt they enjoy the drone strikes either.
Well, Firefox OS is the only one only using Web technologies. With the other two, you might have been using one of the other options for any number of reasons, like for instance that C++ was faster, or your boss thought that JavaScript was a toy language, or that the API you actually needed for the app was missing in the web app version, and then that cut you out of reuse.
What would be even better is if all these manufacturers got together and worked out a common API. Then developers wouldn't have to write a different app for each device. It can't really be that hard.
At least with Firefox OS you stand a reasonable chance of being able to reuse some of your code in your web project as well as your mobile project.
I don't think Ford was interested in stopping WW2: he was one of Hitler's major foreign supporters.
Because scumbags.
I don't want to piss on your parade, but actually that is pretty hard and nobody has really done any kind of work on that yet. All that they have done (and what this project promises) is to do digital modelling of what we know to be analogue systems. There's no talk of actually designing analogue microprocessors and hooking them up in massively parallel, biologically plausible configurations. All the work on analogue systems stopped in the 60s. The huge progress we've made on von Neumann machines doesn't actually buy us anything here.
It represents approximately 1% of the number of neurons, nodes in the system. In the human brain each neuron is connected to, on average, 7000 other neurons. So you would not have to build a computer 100x faster in 10 years. You would have to build a computer approximately 10000x faster. This is what I mean when I say there is a scaling factor involved, and we might not be quite there with the hardware yet.
As far as I am aware current VLSI technology can be used to model on the order of 10 billion synapses. The human brain has on the order of 100 trillion synapses. Unless Henry Markham has also invented a radically new kind of supercomputer, we are still somewhat behind.
I can't decompose (decompile?) it because I can't download it.
I can't download it because downloading it would mean making a copy of it.
I can't make a copy of it because that would infringe the author's copyright since they have not published it under a licence which gives me the right to make such a copy.
As far as contacting the authors is concerned, I see little point. I'm hardly going to trust my security to a developer who can't even install an HTTPS certificate correctly, let alone distribute under a licence that allows distribution.
Terrorists and criminals also breathe air and drink water.
So you think that I have something that I wish to communicate secretly, and I am going to do so using a piece of software that I can't see the source code to?
Not sure what millenium you're living in there bro.
So some website has hundreds of mugshot photos up on their site. What kind of person is going to be interested in leafing through those the whole time? I can't see how they could generate any interest whatsoever, let alone enough that you risk being shunned by anyone you actually know.
Your anonymity in a dead drop system depends on the dead drop location being known only to you and to the person with whom you want to exchange the secret.
As soon as you publish the location of the dead drop anyone can observe it and you have no anonymity whatsoever.
I take a picture of you covertly dumping toxic waste into a reservoir.
I take a picture of a policeman using excessive force on a witness.
I take a picture of a politician accepting a bribe to close a park and turn it into condos.
Unfortunately it is now illegal for me to publish any of these photos since they might humiliate the person they depict.
In this world, people don't negotiate contracts for each communication they make with one another. It would hardly be scalable to do so.
That's not actually the issue with revenge porn. The item under consideration is where someone takes a compromising selfie and sends it to their partner while they are in a relationship. Then the relationship ends badly and the recipient, feeling hard done by, publishes the selfie with the intent to distress the sender.
I'm so glad you provided evidence for that statement.
In the "GMail wiretapping" case most Slashdotters seemed to be of the opinion that once you have sent someone a message, the message belongs to the recipient not the author.
It will authenticate as the hijackee, so it will get through.
"The Internet" is not to blame. As I said before the solution is unclear because people are lazy and not motivated enough to fix the problem.
Here's a potential motivation: a massive player like Google decides to advertise a "spam free" version of GMail, or a consortium of such players. They only deliver from valid SPF sources with an audited policy of disconnecting any hosts found to participate in botnets. Want to be able to reach all those customers?
It's probably not that important, as Linus already pointed out.
Right, the technology you are describing is called SPF. As you say it's not universally used, but I disagree that you wouldn't want to absolutely require it as a first step. It's totally trivial to use it and there's no reason not to, so just require it and any stragglers will soon fall in when they realize their mail won't be delivered otherwise.
A major problem with SPF is that it isn't going to have any effect if the sending host is a zombie. Keep those ideas coming!
Unlikely. The current situation represents the state of the art; if Google had invented a better technology for this it would have patented it.
A better means of suppressing spam would be to prevent it or limit it at its source - that's the only way to free up the bandwidth, storage, and processing power that currently has to deal with it.
One possibility is making it cost a small amount of actual money to send an email - micropayments, say $0.0001 per email. This would add up to such tiny amounts for ordinary users that they wouldn't really notice. For businesses it would be covered by operating costs, but for spammers it would be crippling.
Another possibility is to make it computationally expensive to send an email to each recipient. This might not be very effective since spammers often hijack ordinary users' computers to do the spamming.
Yet another solution would be to require that all email be end to end encrypted: you'd have to have the public key of the person you want to send to. Obtaining people's public keys should be reasonably straightforward but not necessarily easily automatable - in that case the cost of discovery might be too high for spammers.
It's unlikely that any of these strategies will be implemented though as people are lazy and don't actually care enough about fixing the problem, otherwise it would have happened by now.
No you don't, and spam filters currently don't do this as it's extremely hard. Spam filters (that examine message contents) use probability theory based on individual instances of words in the corpus.
2 points: