The OP was under the impression that net-wide opportunistic encryption was already solved by ipsec oe (rfc 4322). I was merely pointing out the problem that rfc4322 requires end-user control of dns for key distribution.
dns is obviously not the only way of distributing keys, ipsec itself does not depend on dns but the "solution" the OP mentioned does.
The purpose of this thing is to enable regular home users to avoid the dragnet filtering that the swedes are implementing. Forging replies for every tcp/udp connection crossing the swedish border would make that filtering a lot more expensive.
How many users do you know that (a) even knows what dns is (b) controls the dns name for their ip (c) is able to configure said dns to include their public key?
OE works fine for geeks, but is too heavy if the goal is to get average home users encrypted.
Thus, for those who are having problems with 4.0, the problem really lies with the user, as the user would have had to make the choice to move to 4.0 in the first place.
And I suppose none of the blame is to be placed at the people who decided to call it "KDE 4.0"?
While the KDE people have never said that KDE4 is mature, most users won't dig through the KDE website and mailinglists. The normal way of doing things is that version x+1 is better than x, how can you blame people when KDE decides to break with common practice with regards to versioning?
So I learned my lesson: Always research before you bid on eBay, even if the bid's not serious.;-)
I rather think the lesson should be: Don't bid on eBay unless you are serious, and especially don't post about it on/. trying to paint yourself as the good guy.
As far as great scams go, I think the greatest one is social and not technical. That is, people's innate ability to twist a story around in order to make themselves believe that they are in the right.
I fail to see the point of this phone for the general public, most people will never touch the OS source code
Exactly.
I wish the general media would shut up about this phone, at the current state of OpenMoko FreeRunner should realistically not even be a blip on the radar except for geeks that like to play with code. It is sort of equivalent to pushing "Desktop Linux" in the age of kernel version 0.95.
If we're doing car analogies, CPU freq would be more in line with engine RPM.
Freq is only a measure of how many ticks (revolutions) per second the CPU is capable of. Without knowing how much computing it can do per tick, you can't translate freq to computing power.
What I should have initially asked is what US providers that use GSM will allow you to bring your own phone to the table and shove one of their sim cards in it and make calls.
The fact that you even feel the need to ask this question makes me wonder how much the US cell carriers have managed to brainwash the populace.
Here in EU it is simply taken as a matter of fact that you can take any sim card you want and have it work in an unlocked phone. If carriers even considered making that impossible, the public would go instant pitchfork mob and the EU equivalent of the FCC would come down so hard on the carrier that tsunamis would flood the US east coast.
To be honest, that's perhaps a slight exaggeration. But only slightly; that you can use any sim in an unlocked phone is Simply The Way It Is.
If you are of the opinion that check marks on a feature comparison list is more important than Stallman's 4 freedoms then this phone is quite simply not for you.
I'd advise non-geek people to stay far away from this phone, there are better and cheaper alternatives everywhere.
If you are the kind of person that thinks phrases like "Linux in your pocket", "software packaging", "porting and cross-compiling", "bugtesting" sounds like fun and you want to help make true open Linux on phones a reality then this is the phone for you. If you are not that kind of person, then buying a FreeRunner now will only be a frustrating experience; please come back later when the platform is more mature.
I suspect that smiley is intended to show that you are joking, but in case it was a serious question: No. Never.
The cell carrier free phone on contract business plan depends on the customer staying with the network, in other words SIM-lock. More often than not they also do modifications to the phone software to flog the carrier's services ("click here to go to our mobile music store!") and disable phone functions that the carrier doesn't like (more so in the.us, but happens in.uk and other places too).
That is exactly opposite to what the FreeRunner is, a niche market linux-geek phone running a completely open and user-modifiable Linux software stack.
I agree. In terms of hardware (and especially on the cellular data side) this isn't anything to get excited about.
On the other hand, this handset's market segment is the "Linux in your pocket" people. Which meant that when they picked the components, the important question was "is this chip supported by Linux" and not "does this chip support the latest technology".
I'm kinda surprised myself that they couldn't find a GSM module that supports at least EDGE. But if the alternative was a binary blob driver (or more likely, unavailable or only-nda-available AT command set documentation, errata etc) well...:-/
Also remember that OpenMoko has a lot of rough edges still. The basic stuff is said to be working, but it is certainly not suitable for Aunt Tilly. The main market for this device is people who absolutely want true open Linux on their phone and who are willing to contribute (writing software, bug reports, porting, testing, etc) to make that happen. So, hardware ain't hot but it is as open as they could find; and except for camera it has all the hardware components needed (bt, cellular data, usb, wifi, etc) to enable people to test and develop the software needed for a fully featured handset.
The OP does have a point, though. The Openmoko folks better make sure that gsmd (and whatever device node the serial interface has) has proper access control, so that a rogue program can't give you a nasty surprise when you get the phone bill.
In many phones (including the FreeRunner) the entire GSM stack is handled by a separate chip. As you say, it is required for type approval in a handset that is as open as this one.
Writing software to talk to one of those is quite deja vu if you were into modem stuff in the bbs days. Serial link, AT command set. From the point of view of the Linux software running on the phone, it is pretty much identical to an old PC connected to a serial modem. Retro-computing in your pocket.;-p
"You managed to make an ends-justify-the-means argument" Yes, I'm justifying AVG's actions.
Well, if it actually worked you might have had a somewhat valid point. Tell me, what exactly does AVG's shotgun approach do that could not be achieved by using a local filtering proxy instead?
The main objection is that this "feature" is just plain dumb and has a horrible implementation.
If it had 802.11n and a 4-port GigE switch I wouldn't complain, but the current hardware spec on this thing makes it just a clone of the good old wrt54gl. It is really nothing new or exciting at all, just a clone of a Linksys product.
Now, with some USB ports you can do all sorts of additional stuff. External harddisks. Printers. Scanners. NAS for your home network. uPnP media server. Network printer/scanner server. Look up all the things people have been using NSLU2s for and then imagine a device that has the capabilities of both the 54GL and the NSLU2.
The specs on this thing is suspiciously similar to the good old WRT54GL. Unless the price is lower, I really don't see what this thing brings to the table.
If it had just included a couple USB ports and upped the ram/flash a little bit, it would have improved the hackability considerably. Look at what people have been able to do with the NSLU2. With these fairly minor changes the WGR614L could supersede both the 54GL and NSLU2.
Slashdot Mirror
The OP was under the impression that net-wide opportunistic encryption was already solved by ipsec oe (rfc 4322). I was merely pointing out the problem that rfc4322 requires end-user control of dns for key distribution.
dns is obviously not the only way of distributing keys, ipsec itself does not depend on dns but the "solution" the OP mentioned does.
No, it is a shortcoming in spec. No amount of friendly user interface will fix the problem that the average home user does not control DNS for his IP.
That the hardware supports it does not mean that the driver supports it.
Then how do you defend against the mitm substituting his own keys?
Why do you think your web browser includes root certificates?
The purpose of this thing is to enable regular home users to avoid the dragnet filtering that the swedes are implementing. Forging replies for every tcp/udp connection crossing the swedish border would make that filtering a lot more expensive.
Excellent, grasshopper.
Now for the next step: how to share said keys in an efficient way.
How many users do you know that (a) even knows what dns is (b) controls the dns name for their ip (c) is able to configure said dns to include their public key?
OE works fine for geeks, but is too heavy if the goal is to get average home users encrypted.
Yes, we all know what a mitm attack is.
This is not to defend against mitm attacks targeted at specific users, this is to defend against net-wide dragnet filtering.
Thus, for those who are having problems with 4.0, the problem really lies with the user, as the user would have had to make the choice to move to 4.0 in the first place.
And I suppose none of the blame is to be placed at the people who decided to call it "KDE 4.0"?
While the KDE people have never said that KDE4 is mature, most users won't dig through the KDE website and mailinglists. The normal way of doing things is that version x+1 is better than x, how can you blame people when KDE decides to break with common practice with regards to versioning?
c|n>k
I was drinking coffee when I read that, you insensitive clod!
So I learned my lesson: Always research before you bid on eBay, even if the bid's not serious. ;-)
I rather think the lesson should be: Don't bid on eBay unless you are serious, and especially don't post about it on /. trying to paint yourself as the good guy.
As far as great scams go, I think the greatest one is social and not technical. That is, people's innate ability to twist a story around in order to make themselves believe that they are in the right.
Removes the need for a phone. Oh, and GPS.
But yeah, touché; N8x0 are excellent devices.
I fail to see the point of this phone for the general public, most people will never touch the OS source code
Exactly.
I wish the general media would shut up about this phone, at the current state of OpenMoko FreeRunner should realistically not even be a blip on the radar except for geeks that like to play with code. It is sort of equivalent to pushing "Desktop Linux" in the age of kernel version 0.95.
If we're doing car analogies, CPU freq would be more in line with engine RPM.
Freq is only a measure of how many ticks (revolutions) per second the CPU is capable of. Without knowing how much computing it can do per tick, you can't translate freq to computing power.
What I should have initially asked is what US providers that use GSM will allow you to bring your own phone to the table and shove one of their sim cards in it and make calls.
The fact that you even feel the need to ask this question makes me wonder how much the US cell carriers have managed to brainwash the populace.
Here in EU it is simply taken as a matter of fact that you can take any sim card you want and have it work in an unlocked phone. If carriers even considered making that impossible, the public would go instant pitchfork mob and the EU equivalent of the FCC would come down so hard on the carrier that tsunamis would flood the US east coast.
To be honest, that's perhaps a slight exaggeration. But only slightly; that you can use any sim in an unlocked phone is Simply The Way It Is.
If you are of the opinion that check marks on a feature comparison list is more important than Stallman's 4 freedoms then this phone is quite simply not for you.
I'd advise non-geek people to stay far away from this phone, there are better and cheaper alternatives everywhere.
If you are the kind of person that thinks phrases like "Linux in your pocket", "software packaging", "porting and cross-compiling", "bugtesting" sounds like fun and you want to help make true open Linux on phones a reality then this is the phone for you. If you are not that kind of person, then buying a FreeRunner now will only be a frustrating experience; please come back later when the platform is more mature.
I suspect that smiley is intended to show that you are joking, but in case it was a serious question: No. Never.
The cell carrier free phone on contract business plan depends on the customer staying with the network, in other words SIM-lock. More often than not they also do modifications to the phone software to flog the carrier's services ("click here to go to our mobile music store!") and disable phone functions that the carrier doesn't like (more so in the .us, but happens in .uk and other places too).
That is exactly opposite to what the FreeRunner is, a niche market linux-geek phone running a completely open and user-modifiable Linux software stack.
I agree. In terms of hardware (and especially on the cellular data side) this isn't anything to get excited about.
On the other hand, this handset's market segment is the "Linux in your pocket" people. Which meant that when they picked the components, the important question was "is this chip supported by Linux" and not "does this chip support the latest technology".
I'm kinda surprised myself that they couldn't find a GSM module that supports at least EDGE. But if the alternative was a binary blob driver (or more likely, unavailable or only-nda-available AT command set documentation, errata etc) well... :-/
Also remember that OpenMoko has a lot of rough edges still. The basic stuff is said to be working, but it is certainly not suitable for Aunt Tilly. The main market for this device is people who absolutely want true open Linux on their phone and who are willing to contribute (writing software, bug reports, porting, testing, etc) to make that happen. So, hardware ain't hot but it is as open as they could find; and except for camera it has all the hardware components needed (bt, cellular data, usb, wifi, etc) to enable people to test and develop the software needed for a fully featured handset.
The OP does have a point, though. The Openmoko folks better make sure that gsmd (and whatever device node the serial interface has) has proper access control, so that a rogue program can't give you a nasty surprise when you get the phone bill.
In many phones (including the FreeRunner) the entire GSM stack is handled by a separate chip. As you say, it is required for type approval in a handset that is as open as this one.
Writing software to talk to one of those is quite deja vu if you were into modem stuff in the bbs days. Serial link, AT command set. From the point of view of the Linux software running on the phone, it is pretty much identical to an old PC connected to a serial modem. Retro-computing in your pocket. ;-p
"You managed to make an ends-justify-the-means argument"
Yes, I'm justifying AVG's actions.
Well, if it actually worked you might have had a somewhat valid point. Tell me, what exactly does AVG's shotgun approach do that could not be achieved by using a local filtering proxy instead?
The main objection is that this "feature" is just plain dumb and has a horrible implementation.
What the hell do you need USB for?
If it had 802.11n and a 4-port GigE switch I wouldn't complain, but the current hardware spec on this thing makes it just a clone of the good old wrt54gl. It is really nothing new or exciting at all, just a clone of a Linksys product.
Now, with some USB ports you can do all sorts of additional stuff. External harddisks. Printers. Scanners. NAS for your home network. uPnP media server. Network printer/scanner server. Look up all the things people have been using NSLU2s for and then imagine a device that has the capabilities of both the 54GL and the NSLU2.
The specs on this thing is suspiciously similar to the good old WRT54GL. Unless the price is lower, I really don't see what this thing brings to the table.
If it had just included a couple USB ports and upped the ram/flash a little bit, it would have improved the hackability considerably. Look at what people have been able to do with the NSLU2. With these fairly minor changes the WGR614L could supersede both the 54GL and NSLU2.