As this kind of forum gets more popular, will clusters of people like Slashdot fall prey to the idiots, trolls, spammers, etc. that ruined usenet? Are the few moderation mechanisms in the system good enough to keep such garbage out?
There is a big difference between news sites such as Slashdot and Usenet newsgroups: each news site is controled by a small group of people (sometimes only one webmaster) that can change and adjust the moderation software whenever necessary. This is not possible with the newsgroups: changing the moderation policy on a newsgroup is a nightmare because it is driven by a process (RFD/CFV) prevents any quick changes in case of emergency. This is due to some architectural differences between the WWW and Usenet (one WWW server versus thousands of NNTP servers replicating each other) as well as some political differences (guidelines for creating Usenet newsgroups).
Another difference is that all news sites have their own interface for submitting new comments, and they implement their own moderation policies. This diversity makes it very difficult for spammers and trolls to ruin many news sites in a short time, as they can do with the newsgroups. It is trivial to write some software that can post the same spam to thousands of newsgroups.
Today I find it pretty much useless, save for a very few highly technical groups.
In this respect, news sites such as Slashdot may be similar to Usenet newsgroups. You will probably find hundreds or thousands of news sites on the Internet, but only a handful of them are relevant for you. So from your (or my) point of view, the WWW-based news sites will be 99% crap and 1% interesting (note that others may choose a 1% that is different from yours). This is not necessarily a bad thing, as long as you can find the 1% that suits you.
While they remain in a state of infancy, the rise of sites such as Slashdot and FreeRepublic herald the demise of traditional information flows.
Traditional information flows will change, but traditional information sources will not disappear. The community information filters would be slower and less efficient without theses sources. For example, most of the news posted on Slashdot are links to some of the traditional news sources. Calling Slashdot a "community information filter" implies that it is usually a filter, not a source.
But to some extent, the community-driven sites are also a source of information (the comments are often more informative than the original article). They could eventually become the dominant source of information, although the transition will be slow and there are some pitfalls on the way. One of the dangers is that the communities may be disconnected from other sources of information. Sometimes, important news come from places that are outside the community's focus, yet they have an impact on the community. Having some members subscribed to some "generalist" sources of information (that may be biased) ensures that the important news will eventually reach the community.
That model could be extended to a world in which most of the information comes from community-driven sources, as long as each community has enough members participating in other communities or collecting the information on their own. But without the traditional news flows, that model is only viable for (very) large communities. Or small ones that are mostly spin-offs from larger ones (thus relying on larger communities as their source of information).
I also doubt that such a technique will really help anyone to try something before buying it.
Even if that mouse was providing very good and accurate feedback (just like having the object in your hands), what would prevent the vendor from cheating? It is exactly the same problem as when you buy an object that looks very nice on the web, but is actually not as good as you saw on the advertisements.
Virtual feelings (with this feedback mouse) can be fooled as easily as virtual sight. As long as someone can get between you and the object, and modify the feedback that you get, they can fool you easily and make you buy something that is not as you "saw" or "felt" it.
Well, I don't have a Palm, but I have a Psion 5 instead. I prefer the Psion 5 because it has a real keyboard and can do a bit more stuff than the Palm, but both of them share more or less the same basic concepts so this message applies to all PDAs.
These devices can do much more than a paper notebook. For example, I have yet to see a paper notebook that beeps (or plays a fancy tune) to remind me about an appointement. Or one that wakes me up in the morning. Thanks to its keyboard, the Psion 5 is also very useful for taking notes. Sure, I could take these notes on paper, but then I would have to re-type them later on my PC anyway, because very often I have to include these notes in a document or to send them by e-mail to some colleagues. Instead of doing this, I can simply connect the little machine to my PC, transfer the notes and edit them.
I can also connect the machine (Psion or Palm) to a mobile phone and access some information remotely while I am on a business trip. I admit that it takes a while if the documents are big, but this is definitely faster than trying to find a way to get these documents faxed to you.
Sure, the usefulness of these devices depends very much on what you do for a living. Maybe some people would be more efficient with a pencil and paper. But when used properly, these tools are invaluable.
IIS already does this. Its free it runs on your desktop (9*,NTW). What can change?
I know that IIS already has this kind of extensions. The difference is that it does not currently come pre-installed with the Win9x machines (i.e. the ones that almost everybody uses). Internet Explorer was also available as a free download for Win95, but it did not take off until it came pre-installed on every user's desktop.
You will not even have to look for a web server and install it: it will already be there. So why bother with other servers (like Apache) if you already have one and the others do not offer significantly more features?
There can be another consequence to this: if every PC comes with a web server pre-installed, Microsoft and other companies can start developing services that rely on these servers and on some MS-specific features included in these servers. This would be attractive for companies that have many computers on an intranet: these web servers could be part of a distributed document management infrastructure, they could be tightly coupled with Office 2000 or with Exchange.
Just an example: you want to send a document (created with Word 2000) to your colleagues. Instead of sending it as an attachment to your mail message, Outlook 2000 could automatically send your announcement (only a short text) to all recipients and include a link to the document that is then published by your local web server. It could even take care of some security issues, such as making sure that only the recipients of the message can access the file. This would be a nice feature because it would reduce the load on the Exchange server (which is not so good under high loads) and at the same time it would lock all people into this proprietary solution.
They are applying their usual strategy to web servers: embrace and extend.
As a Win2000 user, you will get a free web server on your desktop. It will be very easy for you to publish some pages on the web, thanks to this nice tool. Maybe you will even get a few pre-installed scripts and utilities with it, such as a hit counter. Maybe also some FrontPage extensions so that it is easier for you to publish and manage your documents.
By coincidence, you will find the same set of utilities plus much more in IIS. So once you reach the limits of the free server, the only sensible way to upgrade your server will be to move to IIS. It will be easy, and you will not have to re-write any of your code.
On the other hand, upgrading to Apache will be much more difficult, because nobody told you that the nice features offered by the free server were MS-specific and were not available with Apache. So why would you take Apache if it would force you to modify several of your pages? Re-writing these pages could cost more than the price of Windows NT Server + IIS, so you will forget about this "free" Apache that requires more work from you.
And while you are at it, you will also use the nice goodies provided with the server, such as the ones that put active channels on the user's desktop. Never mind of some users cannot use these features because they have a non-standard browser (i.e. a browser that does not come pre-installed with the OS). They will be forced to upgrade sooner or later anyway, so why bother?
Well, it did not take long... I was browsing the site and suddenly I got several "connection reset by peer" in various frames. Did their Roxen server crash? Or maybe they do not have enough bandwidth to serve all curious/.'ers?
But still, securityfocus.com seems to be a very interesting site. From what I was able to see, they provide lots of useful infos and links to many useful tools.
Re:New... Athlon! With triple-cleaning power!
on
K7 Renamed "Athlon"
·
· Score: 2
Let's see what AMD could have done after the K6...
K7 - Is a common abbreviation for "cassette" in French (e.g. videotape).
K8 - Kate, where are you?
K9 - Will bite you!
K0 - Is your processor crashing frequently?
K12 - Hey kids, this processor is for you!
Hmmm... It's about time for AMD to rename their chip.
So if no other company has registered "Athlon" (in the computer industry), then AMD can be sure that nobody else will sell a processor with the same name.
Anything that is "too obvious" cannot be registered as a trademark. This includes:
A number (386, 486).
A single letter followed by a number, especially if it is part of some predictable series (i486, i586, K6, K7).
A common word or a combination of words that is too descriptive or that is already in common use in the area in which the trademark should be registered (for example, you could not register "fast computer" in the computer business, although you might be allowed to get that trademark for a door or a toilet seat).
So if no other company has registered "Athlon" (in the computer business), then AMD can be sure that nobody else will sell a processor called K7.
They will also have more control over the name. For example, if someone claims to have a device which is "Athlon-certified", AMD will have something to say about it (i.e. they could sue the guy if that claim is false). But if it was only "K7-certified", it would have been easy for anyone to say that they were refering to some other obscure chip that just happens to be also called K7 and they could get away with it.
So from a trademark point of view, having the K7 renamed "Athlon" makes a lot of sense. Whether or not the name "Athlon" is good and will help selling the chip is another story...
Problem is, that the higher level results of a lower level start configuration are emergent, which is why nanotechies keep talking about evolutionary systems.
You seem to assume that the system will evolve on its own, without any external influence. I think that the evolution can (and should) be guided in order to obtain what you want.
How do you design a starting configuration that will evolve into the desired high level result?
You don't. Instead, you guide the process at every step. You start by creating some basic building blocks (nanotubes, nanoswitches, etc.) and some basic tools for putting the blocks in place (nanotrains, etc.), then you use these to create some more advanced nanomachines that will in turn create some more complex building blocks. It is only when you reach a sufficient level of complexity that you can hope to have some parts of the system that are able to evolve on their own.
But as I wrote in another comment, I am still wondering about how you can control all these initial steps. How to tell thousands of nanotrains moving on the same surface to deliver their load in the right place?
Looks pretty cool. Where would they get the miniature traindrivers ?
I have been wondering about that too. Not about little men driving these trains, but about how such a system could be told to do complex tasks involving stopping and starting trains, changing tracks, telling the trains to drop their load, synchronizing trains, etc. This should not be too hard with a small number of trains (well, if you consider that building these trains is "simple"), but controlling thousands of these trains on the same surface would be a nightmare.
It would probably be necessary to have some kind of electronic circuits controlling the trains or the tracks. Maybe it could be possible to build a control grid under the glass layer? Of course that grid would have to be built by a previous generation of nano-machines...
They are only interested in the communication channels that can be intercepted easily.
Ummm...yeah. That'd be real useful. While we're at it, why don't we arrest only those criminals that happen to wander into jail cells.
Although the analogy is funny, it is not correct. You will probably never see a criminal wandering into jail, but it is very likely that you are using unprotected communication channels most of the time: fixed or mobile phone, connection to the Internet, etc. If any link between the sender and receiver can be easily tapped, the whole conversation is not private anymore. Do you know which links your phone company is using, and how secure they are? Do you know if they are using buried or aerial cables, microwave links, satellite relays, and so on?
So even if the spies do not have access to the switches of the telecom operators (which are the easiest and best source of information - but would not be unnoticed), they can already gather a lot of information by eavesdropping on a few selected links.
Assuming Echelon exists, big criminal/terrorist/whatever groups would know about it--and use unbreakable encryption. Which leaves only nobodies for Echelon to spy on.
It depends on the purpose of the system. If Echelon (or whatever exists) is designed to identify suspicious communications and analyze them further, then the simple act of using encryption can already put that criminal group on the suspicious list. If some people are known to exchange encrypted messages regularly and none of them are involved in electronic commerce or banking transactions, then the spies could have a closer look at them. Of course this depends on other factors such as the location of the parties, the frequency of these exchanges, and so on...
If you are trying to hide something, then the best thing to do is to act like if you had nothing to hide. So if I were a member of a criminal group, I would try to avoid encryption because it would draw the attention on me (steganography is a different matter, as I explained in another message.)
I think falling for this kind of thing betrays a shaky grasp of technology, not to mention logic.
But a system like Echelon can be built with today's technology and can be used to scan an enormous amount of information. As someone else explained in another thread, even the technology from more than 10 years ago would be sufficient to scan a fair amount of today's communications. Not all communications, but enough for all practical purposes.
Note that I am not sure if such a system is really existing today, and used to the extent described in various articles about Echelon. (And frankly, I don't really care.) But I think that I know enough about computers and telecommunications to be convinced that such a system can be build, and would not even be extremely expensive.
The question is how much processing capabilities do those dictionary computers have? Those stations have to be capturing terabytes of data every day and I'm not sure how well the computers can process the data.
The processing power required to search for some words in a data stream is reasonable and certainly affordable for NSA and the like.
Let's assume that they have a farm of computers scanning for words in huge amounts of clear text messages (encrypted messages are a separate case). You could compare the tasks performed by these computers to the ones done by routers and switches:
A router or a switch receives millions of IP packets (or ATM cells) per second, and looks at their header to decide on which link they have to go.
A dictionary computer receives a comparable amount of data and decides if a message should be discarded or not by looking for some pre-defined patterns in the data stream.
The routers have some specialized hardware for extracting the destination address from the packets and matching this against a routing table. Nothing prevents the spooks from having the same kind of thing for pattern matching.
If some telecom operators and ISPs can afford to deploy Gigabit or Terabit switches in their backbone network, think about what the spooks can do with their budget. Scanning a few Terabytes of (clear text) data is actually quite easy to do with modern technology.
Scanning encrypted data (including steganograpy) is of course harder to do, by several orders of magnitude. But they probably only attempt to decrypt the "interesting" messages (depending on the sender, receiver or carrier). Also, the simple fact of detecting that you are exchanging encrypted messages with someone is already interesting for the spooks. And if you exclude steganography, detecting encrypted messages in a data stream is as easy as scanning for words, which is reasonably easy to do, as explained above.
1) Collection: The various agencies involved would have to have their fingers in many many pies to cover all the ways people could communicate with each other. Phone lines (voice and data), banks, radio, cells, satellites, etc, etc, etc.
They are not trying to cover all the ways people could communicate with each other. They are only interested in the communication channels that can be intercepted easily. In this case, "easily" means anything that does not require the cooperation of the sender, receiver or carrier of information. Radio transmissions from satellites are an obvious candidate, as well as signals from any cables that are not usually accessed for maintainance (e.g. the cables that are deep in the ocean). They do not want to be detected when they are spying on us, so they are only scanning a few communication channels. But that is already a lot of information.
3) Secrecy: They've been doing for 50 years without a hitch? When they'd obviously need an army of techs/programmers/spooks? Not to mention all the accomplices necessary (phone companies, computer/software makers, etc)?
Obviously they do not want to have any accomplices. That's how they keep the secret. So they do not spy the communication channels that would require the cooperation of any third party.
Number 3 has some additional points: If this conspiracy is so vast, evil and secret, how come Joe Blow from Australia was allowed to blow the whistle?
Well, that particular Joe Blow had some reasons to know about this, given his position.
On the other hand, the fact that the Australian Government is willing to admit this publicly might imply that there is some bigger secret that they are trying to hide by "revealing" a secret that they could not hold for much longer. Don't you love all these conspiracy theories?
Sure, you can try to fool the program that is scanning for keywords. But this program is only the first step in the filtering. And it is probably not the only thing that can trigger a more detailled analysis of your conversations or e-mail messages.
So if you are not extremely careful about what you write, you might forget some keyword once (remember that we do not know which keywords they are looking for, and there may be much more than the obvious ones). And if some agent spots that message and decides to have a closer look at all your other messages, then you might be caught.
Personally, I do not care too much about what they do for hunting terrorists, because I am not making bombs or selling illegal drugs in my spare time and I do not think that I would get caught for any illegal activities (although I can never be too sure about that). And hopefully, none of these smart spies would have an interest in displaying some bits of my private life in public. But the gathering of commercial intelligence is an interesting issue...
Slashdot Mirror
There is a big difference between news sites such as Slashdot and Usenet newsgroups: each news site is controled by a small group of people (sometimes only one webmaster) that can change and adjust the moderation software whenever necessary. This is not possible with the newsgroups: changing the moderation policy on a newsgroup is a nightmare because it is driven by a process (RFD/CFV) prevents any quick changes in case of emergency. This is due to some architectural differences between the WWW and Usenet (one WWW server versus thousands of NNTP servers replicating each other) as well as some political differences (guidelines for creating Usenet newsgroups).
Another difference is that all news sites have their own interface for submitting new comments, and they implement their own moderation policies. This diversity makes it very difficult for spammers and trolls to ruin many news sites in a short time, as they can do with the newsgroups. It is trivial to write some software that can post the same spam to thousands of newsgroups.
In this respect, news sites such as Slashdot may be similar to Usenet newsgroups. You will probably find hundreds or thousands of news sites on the Internet, but only a handful of them are relevant for you. So from your (or my) point of view, the WWW-based news sites will be 99% crap and 1% interesting (note that others may choose a 1% that is different from yours). This is not necessarily a bad thing, as long as you can find the 1% that suits you.
Traditional information flows will change, but traditional information sources will not disappear. The community information filters would be slower and less efficient without theses sources. For example, most of the news posted on Slashdot are links to some of the traditional news sources. Calling Slashdot a "community information filter" implies that it is usually a filter, not a source.
But to some extent, the community-driven sites are also a source of information (the comments are often more informative than the original article). They could eventually become the dominant source of information, although the transition will be slow and there are some pitfalls on the way. One of the dangers is that the communities may be disconnected from other sources of information. Sometimes, important news come from places that are outside the community's focus, yet they have an impact on the community. Having some members subscribed to some "generalist" sources of information (that may be biased) ensures that the important news will eventually reach the community.
That model could be extended to a world in which most of the information comes from community-driven sources, as long as each community has enough members participating in other communities or collecting the information on their own. But without the traditional news flows, that model is only viable for (very) large communities. Or small ones that are mostly spin-offs from larger ones (thus relying on larger communities as their source of information).
I also doubt that such a technique will really help anyone to try something before buying it.
Even if that mouse was providing very good and accurate feedback (just like having the object in your hands), what would prevent the vendor from cheating? It is exactly the same problem as when you buy an object that looks very nice on the web, but is actually not as good as you saw on the advertisements.
Virtual feelings (with this feedback mouse) can be fooled as easily as virtual sight. As long as someone can get between you and the object, and modify the feedback that you get, they can fool you easily and make you buy something that is not as you "saw" or "felt" it.
Well, I don't have a Palm, but I have a Psion 5 instead. I prefer the Psion 5 because it has a real keyboard and can do a bit more stuff than the Palm, but both of them share more or less the same basic concepts so this message applies to all PDAs.
These devices can do much more than a paper notebook. For example, I have yet to see a paper notebook that beeps (or plays a fancy tune) to remind me about an appointement. Or one that wakes me up in the morning. Thanks to its keyboard, the Psion 5 is also very useful for taking notes. Sure, I could take these notes on paper, but then I would have to re-type them later on my PC anyway, because very often I have to include these notes in a document or to send them by e-mail to some colleagues. Instead of doing this, I can simply connect the little machine to my PC, transfer the notes and edit them.
I can also connect the machine (Psion or Palm) to a mobile phone and access some information remotely while I am on a business trip. I admit that it takes a while if the documents are big, but this is definitely faster than trying to find a way to get these documents faxed to you.
Sure, the usefulness of these devices depends very much on what you do for a living. Maybe some people would be more efficient with a pencil and paper. But when used properly, these tools are invaluable.
I know that IIS already has this kind of extensions. The difference is that it does not currently come pre-installed with the Win9x machines (i.e. the ones that almost everybody uses). Internet Explorer was also available as a free download for Win95, but it did not take off until it came pre-installed on every user's desktop.
You will not even have to look for a web server and install it: it will already be there. So why bother with other servers (like Apache) if you already have one and the others do not offer significantly more features?
There can be another consequence to this: if every PC comes with a web server pre-installed, Microsoft and other companies can start developing services that rely on these servers and on some MS-specific features included in these servers. This would be attractive for companies that have many computers on an intranet: these web servers could be part of a distributed document management infrastructure, they could be tightly coupled with Office 2000 or with Exchange.
Just an example: you want to send a document (created with Word 2000) to your colleagues. Instead of sending it as an attachment to your mail message, Outlook 2000 could automatically send your announcement (only a short text) to all recipients and include a link to the document that is then published by your local web server. It could even take care of some security issues, such as making sure that only the recipients of the message can access the file. This would be a nice feature because it would reduce the load on the Exchange server (which is not so good under high loads) and at the same time it would lock all people into this proprietary solution.
They are applying their usual strategy to web servers: embrace and extend.
As a Win2000 user, you will get a free web server on your desktop. It will be very easy for you to publish some pages on the web, thanks to this nice tool. Maybe you will even get a few pre-installed scripts and utilities with it, such as a hit counter. Maybe also some FrontPage extensions so that it is easier for you to publish and manage your documents.
By coincidence, you will find the same set of utilities plus much more in IIS. So once you reach the limits of the free server, the only sensible way to upgrade your server will be to move to IIS. It will be easy, and you will not have to re-write any of your code.
On the other hand, upgrading to Apache will be much more difficult, because nobody told you that the nice features offered by the free server were MS-specific and were not available with Apache. So why would you take Apache if it would force you to modify several of your pages? Re-writing these pages could cost more than the price of Windows NT Server + IIS, so you will forget about this "free" Apache that requires more work from you.
And while you are at it, you will also use the nice goodies provided with the server, such as the ones that put active channels on the user's desktop. Never mind of some users cannot use these features because they have a non-standard browser (i.e. a browser that does not come pre-installed with the OS). They will be forced to upgrade sooner or later anyway, so why bother?
Sigh...
Well, it did not take long... I was browsing the site and suddenly I got several "connection reset by peer" in various frames. Did their Roxen server crash? Or maybe they do not have enough bandwidth to serve all curious /.'ers?
But still, securityfocus.com seems to be a very interesting site. From what I was able to see, they provide lots of useful infos and links to many useful tools.
Let's see what AMD could have done after the K6...
Hmmm... It's about time for AMD to rename their chip.
Err... The paragraph in the middle should read:
So if no other company has registered "Athlon" (in the computer industry), then AMD can be sure that nobody else will sell a processor with the same name.
Yes, this is probably a trademark issue.
Anything that is "too obvious" cannot be registered as a trademark. This includes:
So if no other company has registered "Athlon" (in the computer business), then AMD can be sure that nobody else will sell a processor called K7.
They will also have more control over the name. For example, if someone claims to have a device which is "Athlon-certified", AMD will have something to say about it (i.e. they could sue the guy if that claim is false). But if it was only "K7-certified", it would have been easy for anyone to say that they were refering to some other obscure chip that just happens to be also called K7 and they could get away with it.
So from a trademark point of view, having the K7 renamed "Athlon" makes a lot of sense. Whether or not the name "Athlon" is good and will help selling the chip is another story...
You seem to assume that the system will evolve on its own, without any external influence. I think that the evolution can (and should) be guided in order to obtain what you want.
You don't. Instead, you guide the process at every step. You start by creating some basic building blocks (nanotubes, nanoswitches, etc.) and some basic tools for putting the blocks in place (nanotrains, etc.), then you use these to create some more advanced nanomachines that will in turn create some more complex building blocks. It is only when you reach a sufficient level of complexity that you can hope to have some parts of the system that are able to evolve on their own.
But as I wrote in another comment, I am still wondering about how you can control all these initial steps. How to tell thousands of nanotrains moving on the same surface to deliver their load in the right place?
I have been wondering about that too. Not about little men driving these trains, but about how such a system could be told to do complex tasks involving stopping and starting trains, changing tracks, telling the trains to drop their load, synchronizing trains, etc. This should not be too hard with a small number of trains (well, if you consider that building these trains is "simple"), but controlling thousands of these trains on the same surface would be a nightmare.
It would probably be necessary to have some kind of electronic circuits controlling the trains or the tracks. Maybe it could be possible to build a control grid under the glass layer? Of course that grid would have to be built by a previous generation of nano-machines...
Although the analogy is funny, it is not correct. You will probably never see a criminal wandering into jail, but it is very likely that you are using unprotected communication channels most of the time: fixed or mobile phone, connection to the Internet, etc. If any link between the sender and receiver can be easily tapped, the whole conversation is not private anymore. Do you know which links your phone company is using, and how secure they are? Do you know if they are using buried or aerial cables, microwave links, satellite relays, and so on?
So even if the spies do not have access to the switches of the telecom operators (which are the easiest and best source of information - but would not be unnoticed), they can already gather a lot of information by eavesdropping on a few selected links.
It depends on the purpose of the system. If Echelon (or whatever exists) is designed to identify suspicious communications and analyze them further, then the simple act of using encryption can already put that criminal group on the suspicious list. If some people are known to exchange encrypted messages regularly and none of them are involved in electronic commerce or banking transactions, then the spies could have a closer look at them. Of course this depends on other factors such as the location of the parties, the frequency of these exchanges, and so on...
If you are trying to hide something, then the best thing to do is to act like if you had nothing to hide. So if I were a member of a criminal group, I would try to avoid encryption because it would draw the attention on me (steganography is a different matter, as I explained in another message.)
But a system like Echelon can be built with today's technology and can be used to scan an enormous amount of information. As someone else explained in another thread, even the technology from more than 10 years ago would be sufficient to scan a fair amount of today's communications. Not all communications, but enough for all practical purposes.
Note that I am not sure if such a system is really existing today, and used to the extent described in various articles about Echelon. (And frankly, I don't really care.) But I think that I know enough about computers and telecommunications to be convinced that such a system can be build, and would not even be extremely expensive.
The processing power required to search for some words in a data stream is reasonable and certainly affordable for NSA and the like.
Let's assume that they have a farm of computers scanning for words in huge amounts of clear text messages (encrypted messages are a separate case). You could compare the tasks performed by these computers to the ones done by routers and switches:
- A router or a switch receives millions of IP packets (or ATM cells) per second, and looks at their header to decide on which link they have to go.
- A dictionary computer receives a comparable amount of data and decides if a message should be discarded or not by looking for some pre-defined patterns in the data stream.
The routers have some specialized hardware for extracting the destination address from the packets and matching this against a routing table. Nothing prevents the spooks from having the same kind of thing for pattern matching.If some telecom operators and ISPs can afford to deploy Gigabit or Terabit switches in their backbone network, think about what the spooks can do with their budget. Scanning a few Terabytes of (clear text) data is actually quite easy to do with modern technology.
Scanning encrypted data (including steganograpy) is of course harder to do, by several orders of magnitude. But they probably only attempt to decrypt the "interesting" messages (depending on the sender, receiver or carrier). Also, the simple fact of detecting that you are exchanging encrypted messages with someone is already interesting for the spooks. And if you exclude steganography, detecting encrypted messages in a data stream is as easy as scanning for words, which is reasonably easy to do, as explained above.
They are not trying to cover all the ways people could communicate with each other. They are only interested in the communication channels that can be intercepted easily. In this case, "easily" means anything that does not require the cooperation of the sender, receiver or carrier of information. Radio transmissions from satellites are an obvious candidate, as well as signals from any cables that are not usually accessed for maintainance (e.g. the cables that are deep in the ocean). They do not want to be detected when they are spying on us, so they are only scanning a few communication channels. But that is already a lot of information.
Obviously they do not want to have any accomplices. That's how they keep the secret. So they do not spy the communication channels that would require the cooperation of any third party.
Well, that particular Joe Blow had some reasons to know about this, given his position.
On the other hand, the fact that the Australian Government is willing to admit this publicly might imply that there is some bigger secret that they are trying to hide by "revealing" a secret that they could not hold for much longer. Don't you love all these conspiracy theories?
Sure, you can try to fool the program that is scanning for keywords. But this program is only the first step in the filtering. And it is probably not the only thing that can trigger a more detailled analysis of your conversations or e-mail messages.
So if you are not extremely careful about what you write, you might forget some keyword once (remember that we do not know which keywords they are looking for, and there may be much more than the obvious ones). And if some agent spots that message and decides to have a closer look at all your other messages, then you might be caught.
Personally, I do not care too much about what they do for hunting terrorists, because I am not making bombs or selling illegal drugs in my spare time and I do not think that I would get caught for any illegal activities (although I can never be too sure about that). And hopefully, none of these smart spies would have an interest in displaying some bits of my private life in public. But the gathering of commercial intelligence is an interesting issue...