The problem with btrfs snapshots is that if data is corrupted at any point, that data silently remains corrupted in future snapshots because they are copy-on-write. Suppose, for example, that your backup of your httpd.conf file gets corrupted at any point. If you take a snapshot, the "new" version is actually the old version - still corrupted. You then rsync from the source. Rsync looks at the mtime and because the file isn't marked as changed on the source, it doesn't get synced. All backups are still corrupted.
I've seen people come up with the same issue using stuff like rdiff-backup, too -- in both cases, you're making an incremental backup, but doing the necessary fakery (either within the FS for btrfs snapshots or with hard-links on top of the filesystem for rdiff-backup) to make each incremental backup look like a full backup.
This is not a problem with using btrfs snapshots to perform incremental backups -- it's a problem with incremental backups. Note that, in the usual tape-backup scheme, there's two reasons for doing full backups periodically: one is that restoring the last full backup, plus each of the 500 incremental backups since then, is prohibitively time-consuming -- this doesn't apply with fake-full backups. The other is because any corruption in one full backup is no obstacle to correctly recovering from any later full backup, whereas a corruption anywhere in a chain of incremental backups is liable to show up as corruption in the result of recovering from that entire chain (unless you're so lucky that all affected files have been modified since, and thus there are new clean versions to overwrite the corrupt versions) -- and this is exactly the problem you're talking about. The shorter you keep the chain of incremental backups, the less likely you get bit! So the solution, if you simply keep in mind that even though ls shows all your files there, it's not a full backup, is obviously the same as it's always been: you've got to periodically perform full backups as well as your frequent incremental backups.
And you can take advantage of those fake-full backups to do a full backup without the bandwidth consumption. When it's time for a full backup, just do an incremental backup as usual, then do an rsync into that incremental backup with --ignore-times. This will force each file to be re-transfered, but still uses minimal bandwidth since it only transfers the hashes it needs to identify any differences and the full text of those differences, whether due to a corrupted backup or to changes between the incremental backup and the rsync.
I agree that rsyncrypto sounds exactly like the poster child of cryptosystems invented by people who don't understand cryptography.
But I can't resist:
Uhm, that property is exactly what you DON'T want in an encryption algorithm.
Eh, one-time pad has exactly that property: if you use the same key to encrypt similar files, you get similar output. And nobody complains about one-time pad.;)
The Neo900.org front page has gone through some changes since last I remembered -- I'm pretty sure that a couple weeks ago, the USB port being sturdier was mentioned on the front page, but it's not now.
Anyway, it is a well-documented fact that they're quite aware of that issue, and will be avoiding it, but sorry for calling RTFA.
Nitpick: I'm with the AC in objecting to your use of the term "bricking". As long as you don't scribble over the bootloader and initrd (and thus need to reflash it, for which USB is the only easy approach), and you have some way to charging the battery, it's still far more serviceable than a brick. I rigged an inductive charging unit into my N900 and used it for over a year with a busted USB port. Crippled, sure, but not bricked.
VNC, take your pick of which variant. How could anyone remotely in contact with the open source world not know about VNC?
From TFS (emphasis mine):
Recently I switched to Meraki Systems Manager because it is free — and it uses VNC — but unfortunately it isn't intended for the one-time-use type support I'll be offering.
So... yeah. He actually does "know about VNC". He's just wanting a good, user-friendly wrapper/front-end suited to what he's doing with it. As you say, it's a case of "take your pick of which variant" -- and he's asking for recommendations on which variant to pick.
GSM - free GSM module doesn't exists, replacing sw means revoke of certification and using non-certified device on public network is illegal
What a bunch of FUD. On your shitty network in a country with broken laws? Maybe.
dos1 is from Poland. I'm pretty sure that's not the broken-lawed country you were thinking of.
In the real world out there, operators don't "certify" devices.
And he never said they did. However, the FCC in the US, and the corresponding authority in other countries, do certify cellular terminals, and do make it illegal to operate uncertified ones (in the normal way, on a public GSM network).
They provide a SIM and the SIM is used in the whatever device the customer pleases.
Yes, and if you build yourself an uncertified GSM terminal, you are practically free to slot your SIM and use it -- you'll be breaking the law, but as long as your equipment really does conform to the specifications, nobody will notice or care.
However, you can't presently get someone to sell you a prebuilt GSM module with open firmware, because (at least in USA; I'm not sure how the policies of radio comms regulators in other countries compare, but they're unlikely to be much better) the FCC will not certify such a device, because open means the user could load modified firmware that would cause the device to not conform to specifications. (If firmware is only distributed as a binary, that's all ok, because apparently FCC thinks reverse-engineering doesn't happen...)
AFAIK there's not a big enough market for such things to motivate some Chinese factory beyond the FCC's reach to develop and produce open-firmware GSM modules, but even if there was, that only helps if each user is getting their own parts through customs and assembling it. If GolDeliCo imports such a (necessarily uncertified) module and uses it in a phone, that phone will then have to be certified, which not only is a ridiculously expensive process, but also will be impossible (at least in the USA) because the FCC still applies the policy that such firmware cannot be open. And selling an uncertified mobile phone for use on the public GSM networks will get you in trouble in a way that building your own for yourself won't.
Hi, I'm in the military, specifically the Army and began my career as an Infantryman. 1. We have no "definition" of an "assault rifle".
Others have pointed to at least one definition. If you don't have a current definition contradicting that, I think that makes me correct.
2. A 7.62 mm machine IS a crew served weapon, i.e.: the M240B or the M60 that preceded it.
It was likely before your time, but 7.62 cartridges were rather popular in the LMG role before the Minimi and RPK-74. BAR? Bren? RPK? (Indeed, despite the RPK-74 in 5.45x39, RPK variants in 7.62x39 are still widely used in third-world militaries.) Yes, these days most significant militaries use LMGs in 5.56 or 5.45x39, but that doesn't mean the classic LMGs have ceased to exist, or have been "promoted" to GPMGs by virtue of their caliber's unpopularity.
I make no claims that my knowledge is infallible or even particularly well-researched*, but it seems you're assuming anything that doesn't align with what you see today in the U.S. Army is due to "poor research", rather than considering that some people might have broader knowledge of military terminology by drawing on the recent past of militaries throughout the anglosphere.
*If you'd told me that the Army does have a definition for "assault rifle", but it's different from the one I'm familiar with, I'd accept the correction -- I'm not in today's Army, and you are.
yes, but why not make it better? why not make it better and dodge a lawsuit at the same time?
but why not make it better in hw too then, why stick with 800x480 resistive...
Because the projected volume is on the order of 1000 units. In that sort of volume, selecting a new keyboard, new screen, redesigning casework for them, and tooling up to produce it becomes prohibitively expensive.
And I really don't understand what you mean by "dodge a lawsuit" -- AFAIK Nokia doesn't have any design patents on the N900, and even if they do, since they can't be bothered to pursue the makers and sellers of the parts that this will be made of, it seems really unlikely they'd bother coming after the Neo900 crew for (1) a tiny production run with (2) a retired OS lacking in general consumer appeal and (3) a last-gen CPU (though substantially upgraded from the original N900). It's not like this sort of thing is realistically costing them anything in lost sales.
Well, GP is blabbering on without a clear idea of what he's talking about, but given that the military (and IMO only legitimate) definition of the "assault rifle" is a rifle or carbine that is (1) select-fire and (2) chambered for an intermediate cartridge, he's not quite wrong about the deer-hunting thing. (Most states actually forbid use of legally-owned select-fire weapons for hunting, but we'll focus on part 2.) Note however, that "small arms" in military parlance includes everything from handguns to sniper rifles to 7.62 machine guns -- the word "small" is meant in comparison to crew-served weapons -- so if he's right about the cartridges it's more due to luck than knowledge.
OTOH, the news and entertainment media generally don't know or don't care much about firearms, and the general public (including witnesses of this attack) learn from there, so there's no telling what the term "assault rifle" actually refers to, so trying to figure out what cartridge he was using from the use of that term is a singularly foolish pursuit.
Can you elaborate, and provide sources. I am not aware of specific rifles being banned for deer hunting. Nor am I aware of cartridge rules, other than a centerfire requirement.
Of course cartridge legality varies state to state. Many states have rules based on some combination (rational or not) of one or more of: caliber, velocity, case length, and energy; the effect in most states is to prohibit modern assault rifle cartridges (e.g. 5.56 NATO and 5.45x39). Sometimes 7.62x39 is OK, sometimes not.
I can't be bothered to make sure it's right, but a brief google for your state turned up this, which suggests that Oregon is more permissive and simple than most, apparently permitting any.22 caliber or larger centerfire for taking deer, while taking e.g. elk requires.24 caliber or larger.
(Note that, if Oregon law is as simple as that page suggests, it's so loose as to defeat the point of having cartridge restrictions in the first place, as it seems to permit a carbine in.25 ACP for taking elk, while forbidding.220 Swift for elk or.17 Remington for deer.)
If a specific cartridge, say the.223 were banned for its inability to kill, why is a bow legal?
Well, there's two answers to that. The physics-geek one is to do some math about momentum and/or kinetic energy of arrows (which are much slower, but much, much heavier) vs. bullets, maybe throw in some notion of the difference between the wound channels produced by broadheads and bullets, and argue that some arrows are rather more lethal than some assault-rifle rounds (which, despite not having done the math, I'm pretty sure is true).
But the real answer is: because laws aren't rational. Bowhunting is, at this point, a traditional sport pursued for fun, and banning it on humane grounds (whether or not this would make sense from a physics perspective) will be perceived as a killjoy assault on tradition, and will not be politically viable. As I said, I'm pretty sure some arrows exceed the performance of the rifle rounds in question anyway, so I suspect a more rational approach would involve requiring bowhunting setups to meet a minimum momentum or energy standard. However, this would be effectively banning bowhunting for those too small and/or weak to draw a bow with enough energy to meet the requirements, and would be portrayed as an assault on tradition and discrimination against women bowhunters, and would likewise be politically unviable. By contrast, a restriction on rifle chamberings just means a few hunters need to buy a new rifle, then go on hunting -- some will grumble, some will see it as a good excuse to buy the gun they wanted but their wife wouldn't let them buy, but it won't be near as disruptive to the sport.
"WOW - an Opera Web Surfer Please Sing for the Internet!" pointless javascript is pointless.
Whereas an Opera user latching onto any excuse to mention to the world that he uses Opera is... also pointless, but kind of amusing for the rest of us.
Since a regular microphone can't record sound above 20000 Hz, you will be limited to 20kb/s, probably much less.
You assert 1 bit/s/Hz as the limit, why?
That does fall out of the Shannon-Hartley theorem if we assume the SNR is unity (0dB). Some justification might be in order, or at the very least, state that you're making that assumption. As it stands, there's no indication you've ever heard of the Shannon-Hartley theorem, which seems to indicate you're speaking with no theoretical basis.
Anyway, regular microphones of the sort found in ordinary laptops, tablets, and mobiles, have no trouble recording ultrasound at 24 kHz (the maximum generatable by a sound card at 48 kHz sample rate), just as laptop etc. speakers have no trouble producing such sounds. There's some attenuation (at both ends), but it's still way above the noise floor. Your denial of this (which I, and numerous others posting in this thread, have all verified personally) indicates you're speaking with no practical basis.
Since you can't or won't even give the slightest impression that your claim has any basis, theoretical or practical, why should anyone care how many bit/s you think is likely?
That gives you at least 4kHz of bandwidth above the limits of human hearing right there.
Actually only 2KHz because of sampling theory.
The upper limit of human hearing is generally considered to be 20 kHz, mainly because that's a nice round number (it's substantially above the mean for all humans, but a decent estimate for children). If you mean to assert that we should be considering the limit to be 22kHz, you might want to make that assertion explicitly, and be prepared to back it up; if so, do consider the demographics of the people any such malware is at risk of detection from (IT workers, sysadmins, security researchers, but generally not grade-school children).
GP specifically mentioned 48 kHz sample-rate, giving a maximum representable frequency of 24 kHz. So his claim of 4 kHz (the band from 20 kHz to 24 kHz) above the limits of human hearing is rather accurate. (Perhaps you were thinking of the red book CD sample-rate of 44.1 kHz, instead of the one GP actually mentioned?)
p.s. SI. Learn it! "kHz" means 1000 Hz. "KHz" means 1 kelvin * Hz.
Now mobile phones will be allowed in planes, is another security breach.
Yeah, actually you're on crack.
Mobile phones have always been allowed on airplanes. They've also been allowed to be turned on (e.g. to use as an ebook reader, mp3 player, etc.) in part of the flight, provided the cellular radio is turned off. They're still required to have the cellular radio turned off.
The only difference is that you previously weren't permitted to have them powered on during takeoff, climb, approach and landing.
It will only show on the scope if the signal deposits significant energy in a narrow band. If the signal is spread spectrum or frequency hopping, it very well may not show. See Cinavia for an example of a currently undetectable and non-separable (by pirates thus far) hidden channel.
(Note: my knowledge of Cinavia is entirely derived from reading the paper purporting that it is in fact separable by pirates (or, in theory, by those seeking to make legal copies under fair use, who would never, ever dream of violating copyright law) and explaining how it is removed.) Cinavia works because it's embedded in media, so there's plenty of data for it to hide in. Since the pirates don't know the bit-for-bit original without Cinavia, it's not easy to isolate Cinavia. (If you did know the original, you could use that to first reverse the phase/speed errors Cinavia introduces to obfuscate the data channel, then (in the frequency domain) divide the original by the phase-fixed version of the protected track, and the result is the Cinavia modulation. As it is, a bloody mess of heuristics is required.)
This, on the other hand, is applied on top of whatever the speaker's currently supposed to be outputting, which is not only known to the user, but under the user's control. Therefore, you can set it to output silence, and see whatever's left. Even if it's spread across the entire reproducible frequency range of the sound card, there will be some signal (even if it has the appearance of broadband, low-amplitude noise) where there should be none.
If this malware worked, like Cinavia, by modulating the existing audio signal rather than superimposing its own signal, it would be undetectable when the computer is supposed to be outputting silence (so the above method doesn't work), but it also would be incapable of communicating under the same circumstances. Even if this were the case, we could substitute any known signal (e.g. a pure sine wave at a given frequency) and monitor for variations from the expected output.
Bust out an oscilloscope and a logic analyzer and start looking at these signals.
As long as you have a microphone that can work at those frequencies.
1. As several posters have already pointed out, that's not much of a hurdle. 2. Who said anything about a microphone? The easy way is to connect your scope directly to the electrical signal that drives the speaker -- one channel to each of the compromised machines, and you get the signal you want with no hassle. Unless the software is being particularly clever, you won't even have to crack open the case -- just plug in a Y-cable with an external speaker on one leg and a pigtail with bare wires to clip the scope probes to on the other.
Using a microphone to pick up the signal is just senselessly complicating stuff -- now you've got both sides of a presumably duplex connection* and the sound of you coughing, the click of each knob on the 'scope, all coming in through one mic, and have to figure out what frequency is what? No thanks.
*And if there's more than two compromised machines, you've either got multiple simultaneous duplex connections, or the whole thing runs CSMA like ethernet.
I like to use this ABM1 - Passive Air Band Monitor when flying. I keep it discreet as I'm sure most flight crews won't understand how it's different from a typical radio receiver. I regularly hear that "bzz bzz bzz" of cell phones with this device. I then ask my girlfriend sitting next to me if she put her phone in airplane mode. If she hadn't and does it the noise usually goes away. If she had her's in airplane mode then I assume it's someone else sitting near me.
Correct so far.
Phones do cause interference in the aircraft frequency bands (at least at short range).
And... you go off the rails.
GSM phones cause interference in audio-frequency circuits because the phone transmits in regular bursts every 4.62 ms (this is why it doesn't affect CDMA, UMTS, etc., only 2G GSM/GPRS/EDGE, because they use TDMA). It turns out audio amplifiers generally tend to serve as decent wide-band AM receivers, so this is very easily picked up as a 217Hz buzz. However, this is all happening on the audio-frequency side, so calling it "interference in the aircraft frequency bands" is just plain wrong -- it will affect practically any unshielded or insufficiently-shielded device (it's only a couple watts or so transmitter, but within a meter or less the inverse-square law says you do need better shielding than a lot of consumer electronics have) with an amplified audio output, including all sorts of radios, and non-radio devices from MP3 players to cassette players.
Fortunately, it's a really easy problem to solve -- just keep your phone out of the cockpit. Thanks to the inverse-square law, it's really only a practical issue at very short range.
This technology... only becomes useful when it is the only way to identify the location of the car. At which point, the drivers will have already bailed.
implies that you believe criminals will routinely either (1) decelerate to a complete stop physics-defyingly fast when out of sight of police or (2) bail out of a moving car.
In reality, it's possible for a pursuing vehicle to maintain a close enough distance to be able to move in when the GPS tracker indicates the fleeing vehicle is slowing, in order to apprehend anyone exiting it, but not staying close enough to avoid the risk of breaking visual contact long enough for the fleeing vehicle to escape (e.g. an intersection after a blind curve, where a single pursuing unit has 50% or less chance of correctly guessing which route the fleeing vehicle took). In this case, the GPS is useful, both to tell when you need to catch back up, and to prevent them from escaping.
Now noting that there is some use for this isn't the same as saying it's worth the expense (both hardware costs and developing procedures and policies for its use), and the present high-speed chase behavior of several local police agencies in my state makes it obvious to me that they absolutely should not have any new toys to distract them, but it's not as dumb a concept as you claim.
I would also like to understand what's the definition of "unstable code".
Unstable code is code such that, when you make an arbitrarily small change, you end up rewriting the entire thing.
Stable code, by contrast, is code such that when you make an arbitrarily small change, the code ends up being restored to its original state, or perhaps engaging in a bounded oscillation, where you and another coder keep changing it back and forth with every release.
So, open up cellphones to get their LCD displays to build your own heads-up display?
No, you don't do that, because it makes no bloody sense. You buy LCDs* from (a middleman who buys them from) the same factories phone makers buy them from. Whether you go through a middleman or talk to the factory depends on the volume you're buying and whether or not you have a presence in China.
*note: don't say "Liquid Crystal Display displays" unless you work for your employer, the Department of Redundancy Bureau.
No. 20 years ago we had the Casio databank watch... A calculator, world time clock, and personal contact manager. It wasn't cool back then either, it made you a nerd.
I had a Casio calculator watch (just a boring calculator watch, without the databank stuff) when I was a kid, and I can assure you it didn't make me a nerd. In fact, the causality is exactly opposite; being a nerd made me get a calculator watch.
Slashdot Mirror
The problem with btrfs snapshots is that if data is corrupted at any point, that data silently remains corrupted in future snapshots because they are copy-on-write. Suppose, for example, that your backup of your httpd.conf file gets corrupted at any point. If you take a snapshot, the "new" version is actually the old version - still corrupted. You then rsync from the source. Rsync looks at the mtime and because the file isn't marked as changed on the source, it doesn't get synced. All backups are still corrupted.
I've seen people come up with the same issue using stuff like rdiff-backup, too -- in both cases, you're making an incremental backup, but doing the necessary fakery (either within the FS for btrfs snapshots or with hard-links on top of the filesystem for rdiff-backup) to make each incremental backup look like a full backup.
This is not a problem with using btrfs snapshots to perform incremental backups -- it's a problem with incremental backups. Note that, in the usual tape-backup scheme, there's two reasons for doing full backups periodically: one is that restoring the last full backup, plus each of the 500 incremental backups since then, is prohibitively time-consuming -- this doesn't apply with fake-full backups. The other is because any corruption in one full backup is no obstacle to correctly recovering from any later full backup, whereas a corruption anywhere in a chain of incremental backups is liable to show up as corruption in the result of recovering from that entire chain (unless you're so lucky that all affected files have been modified since, and thus there are new clean versions to overwrite the corrupt versions) -- and this is exactly the problem you're talking about. The shorter you keep the chain of incremental backups, the less likely you get bit! So the solution, if you simply keep in mind that even though ls shows all your files there, it's not a full backup, is obviously the same as it's always been: you've got to periodically perform full backups as well as your frequent incremental backups.
And you can take advantage of those fake-full backups to do a full backup without the bandwidth consumption. When it's time for a full backup, just do an incremental backup as usual, then do an rsync into that incremental backup with --ignore-times. This will force each file to be re-transfered, but still uses minimal bandwidth since it only transfers the hashes it needs to identify any differences and the full text of those differences, whether due to a corrupted backup or to changes between the incremental backup and the rsync.
I agree that rsyncrypto sounds exactly like the poster child of cryptosystems invented by people who don't understand cryptography.
But I can't resist:
Uhm, that property is exactly what you DON'T want in an encryption algorithm.
Eh, one-time pad has exactly that property: if you use the same key to encrypt similar files, you get similar output. And nobody complains about one-time pad. ;)
Oh, I'm quite wrong. Sorry about that.
The Neo900.org front page has gone through some changes since last I remembered -- I'm pretty sure that a couple weeks ago, the USB port being sturdier was mentioned on the front page, but it's not now.
Anyway, it is a well-documented fact that they're quite aware of that issue, and will be avoiding it, but sorry for calling RTFA.
Nitpick: I'm with the AC in objecting to your use of the term "bricking". As long as you don't scribble over the bootloader and initrd (and thus need to reflash it, for which USB is the only easy approach), and you have some way to charging the battery, it's still far more serviceable than a brick. I rigged an inductive charging unit into my N900 and used it for over a year with a busted USB port. Crippled, sure, but not bricked.
VNC, take your pick of which variant. How could anyone remotely in contact with the open source world not know about VNC?
From TFS (emphasis mine):
Recently I switched to Meraki Systems Manager because it is free — and it uses VNC — but unfortunately it isn't intended for the one-time-use type support I'll be offering.
So... yeah. He actually does "know about VNC". He's just wanting a good, user-friendly wrapper/front-end suited to what he's doing with it. As you say, it's a case of "take your pick of which variant" -- and he's asking for recommendations on which variant to pick.
permit a carbine in .25 ACP for taking elk, while forbidding .220 Swift for elk or .17 Remington for deer
But that's safe, because a .25 isn't even going to get through the elk's skin except at point-blank range.
Heh, fair point. ;)
RTFA; yes they'll avoid that.
Please. Calm down.
GSM - free GSM module doesn't exists, replacing sw means revoke of certification and using non-certified device on public network is illegal
What a bunch of FUD. On your shitty network in a country with broken laws? Maybe.
dos1 is from Poland. I'm pretty sure that's not the broken-lawed country you were thinking of.
In the real world out there, operators don't "certify" devices.
And he never said they did. However, the FCC in the US, and the corresponding authority in other countries, do certify cellular terminals, and do make it illegal to operate uncertified ones (in the normal way, on a public GSM network).
They provide a SIM and the SIM is used in the whatever device the customer pleases.
Yes, and if you build yourself an uncertified GSM terminal, you are practically free to slot your SIM and use it -- you'll be breaking the law, but as long as your equipment really does conform to the specifications, nobody will notice or care.
However, you can't presently get someone to sell you a prebuilt GSM module with open firmware, because (at least in USA; I'm not sure how the policies of radio comms regulators in other countries compare, but they're unlikely to be much better) the FCC will not certify such a device, because open means the user could load modified firmware that would cause the device to not conform to specifications. (If firmware is only distributed as a binary, that's all ok, because apparently FCC thinks reverse-engineering doesn't happen...)
AFAIK there's not a big enough market for such things to motivate some Chinese factory beyond the FCC's reach to develop and produce open-firmware GSM modules, but even if there was, that only helps if each user is getting their own parts through customs and assembling it. If GolDeliCo imports such a (necessarily uncertified) module and uses it in a phone, that phone will then have to be certified, which not only is a ridiculously expensive process, but also will be impossible (at least in the USA) because the FCC still applies the policy that such firmware cannot be open. And selling an uncertified mobile phone for use on the public GSM networks will get you in trouble in a way that building your own for yourself won't.
Hi, I'm in the military, specifically the Army and began my career as an Infantryman. 1. We have no "definition" of an "assault rifle".
Others have pointed to at least one definition. If you don't have a current definition contradicting that, I think that makes me correct.
2. A 7.62 mm machine IS a crew served weapon, i.e.: the M240B or the M60 that preceded it.
It was likely before your time, but 7.62 cartridges were rather popular in the LMG role before the Minimi and RPK-74. BAR? Bren? RPK? (Indeed, despite the RPK-74 in 5.45x39, RPK variants in 7.62x39 are still widely used in third-world militaries.) Yes, these days most significant militaries use LMGs in 5.56 or 5.45x39, but that doesn't mean the classic LMGs have ceased to exist, or have been "promoted" to GPMGs by virtue of their caliber's unpopularity.
I make no claims that my knowledge is infallible or even particularly well-researched*, but it seems you're assuming anything that doesn't align with what you see today in the U.S. Army is due to "poor research", rather than considering that some people might have broader knowledge of military terminology by drawing on the recent past of militaries throughout the anglosphere.
*If you'd told me that the Army does have a definition for "assault rifle", but it's different from the one I'm familiar with, I'd accept the correction -- I'm not in today's Army, and you are.
yes, but why not make it better? why not make it better and dodge a lawsuit at the same time?
but why not make it better in hw too then, why stick with 800x480 resistive...
Because the projected volume is on the order of 1000 units. In that sort of volume, selecting a new keyboard, new screen, redesigning casework for them, and tooling up to produce it becomes prohibitively expensive.
And I really don't understand what you mean by "dodge a lawsuit" -- AFAIK Nokia doesn't have any design patents on the N900, and even if they do, since they can't be bothered to pursue the makers and sellers of the parts that this will be made of, it seems really unlikely they'd bother coming after the Neo900 crew for (1) a tiny production run with (2) a retired OS lacking in general consumer appeal and (3) a last-gen CPU (though substantially upgraded from the original N900). It's not like this sort of thing is realistically costing them anything in lost sales.
Well, GP is blabbering on without a clear idea of what he's talking about, but given that the military (and IMO only legitimate) definition of the "assault rifle" is a rifle or carbine that is (1) select-fire and (2) chambered for an intermediate cartridge, he's not quite wrong about the deer-hunting thing. (Most states actually forbid use of legally-owned select-fire weapons for hunting, but we'll focus on part 2.) Note however, that "small arms" in military parlance includes everything from handguns to sniper rifles to 7.62 machine guns -- the word "small" is meant in comparison to crew-served weapons -- so if he's right about the cartridges it's more due to luck than knowledge.
OTOH, the news and entertainment media generally don't know or don't care much about firearms, and the general public (including witnesses of this attack) learn from there, so there's no telling what the term "assault rifle" actually refers to, so trying to figure out what cartridge he was using from the use of that term is a singularly foolish pursuit.
Can you elaborate, and provide sources. I am not aware of specific rifles being banned for deer hunting. Nor am I aware of cartridge rules, other than a centerfire requirement.
Of course cartridge legality varies state to state. Many states have rules based on some combination (rational or not) of one or more of: caliber, velocity, case length, and energy; the effect in most states is to prohibit modern assault rifle cartridges (e.g. 5.56 NATO and 5.45x39). Sometimes 7.62x39 is OK, sometimes not.
I can't be bothered to make sure it's right, but a brief google for your state turned up this, which suggests that Oregon is more permissive and simple than most, apparently permitting any .22 caliber or larger centerfire for taking deer, while taking e.g. elk requires .24 caliber or larger.
(Note that, if Oregon law is as simple as that page suggests, it's so loose as to defeat the point of having cartridge restrictions in the first place, as it seems to permit a carbine in .25 ACP for taking elk, while forbidding .220 Swift for elk or .17 Remington for deer.)
If a specific cartridge, say the .223 were banned for its inability to kill, why is a bow legal?
Well, there's two answers to that. The physics-geek one is to do some math about momentum and/or kinetic energy of arrows (which are much slower, but much, much heavier) vs. bullets, maybe throw in some notion of the difference between the wound channels produced by broadheads and bullets, and argue that some arrows are rather more lethal than some assault-rifle rounds (which, despite not having done the math, I'm pretty sure is true).
But the real answer is: because laws aren't rational. Bowhunting is, at this point, a traditional sport pursued for fun, and banning it on humane grounds (whether or not this would make sense from a physics perspective) will be perceived as a killjoy assault on tradition, and will not be politically viable. As I said, I'm pretty sure some arrows exceed the performance of the rifle rounds in question anyway, so I suspect a more rational approach would involve requiring bowhunting setups to meet a minimum momentum or energy standard. However, this would be effectively banning bowhunting for those too small and/or weak to draw a bow with enough energy to meet the requirements, and would be portrayed as an assault on tradition and discrimination against women bowhunters, and would likewise be politically unviable. By contrast, a restriction on rifle chamberings just means a few hunters need to buy a new rifle, then go on hunting -- some will grumble, some will see it as a good excuse to buy the gun they wanted but their wife wouldn't let them buy, but it won't be near as disruptive to the sport.
"WOW - an Opera Web Surfer
Please Sing for the Internet!"
pointless javascript is pointless.
Whereas an Opera user latching onto any excuse to mention to the world that he uses Opera is... also pointless, but kind of amusing for the rest of us.
Since a regular microphone can't record sound above 20000 Hz, you will be limited to 20kb/s, probably much less.
You assert 1 bit/s/Hz as the limit, why?
That does fall out of the Shannon-Hartley theorem if we assume the SNR is unity (0dB). Some justification might be in order, or at the very least, state that you're making that assumption. As it stands, there's no indication you've ever heard of the Shannon-Hartley theorem, which seems to indicate you're speaking with no theoretical basis.
Anyway, regular microphones of the sort found in ordinary laptops, tablets, and mobiles, have no trouble recording ultrasound at 24 kHz (the maximum generatable by a sound card at 48 kHz sample rate), just as laptop etc. speakers have no trouble producing such sounds. There's some attenuation (at both ends), but it's still way above the noise floor. Your denial of this (which I, and numerous others posting in this thread, have all verified personally) indicates you're speaking with no practical basis.
Since you can't or won't even give the slightest impression that your claim has any basis, theoretical or practical, why should anyone care how many bit/s you think is likely?
That gives you at least 4kHz of bandwidth above the limits of human hearing right there.
Actually only 2KHz because of sampling theory.
The upper limit of human hearing is generally considered to be 20 kHz, mainly because that's a nice round number (it's substantially above the mean for all humans, but a decent estimate for children). If you mean to assert that we should be considering the limit to be 22kHz, you might want to make that assertion explicitly, and be prepared to back it up; if so, do consider the demographics of the people any such malware is at risk of detection from (IT workers, sysadmins, security researchers, but generally not grade-school children).
GP specifically mentioned 48 kHz sample-rate, giving a maximum representable frequency of 24 kHz. So his claim of 4 kHz (the band from 20 kHz to 24 kHz) above the limits of human hearing is rather accurate. (Perhaps you were thinking of the red book CD sample-rate of 44.1 kHz, instead of the one GP actually mentioned?)
p.s. SI. Learn it!
"kHz" means 1000 Hz.
"KHz" means 1 kelvin * Hz.
Now mobile phones will be allowed in planes, is another security breach.
Yeah, actually you're on crack.
Mobile phones have always been allowed on airplanes. They've also been allowed to be turned on (e.g. to use as an ebook reader, mp3 player, etc.) in part of the flight, provided the cellular radio is turned off. They're still required to have the cellular radio turned off.
The only difference is that you previously weren't permitted to have them powered on during takeoff, climb, approach and landing.
It will only show on the scope if the signal deposits significant energy in a narrow band. If the signal is spread spectrum or frequency hopping, it very well may not show. See Cinavia for an example of a currently undetectable and non-separable (by pirates thus far) hidden channel.
(Note: my knowledge of Cinavia is entirely derived from reading the paper purporting that it is in fact separable by pirates (or, in theory, by those seeking to make legal copies under fair use, who would never, ever dream of violating copyright law) and explaining how it is removed.)
Cinavia works because it's embedded in media, so there's plenty of data for it to hide in. Since the pirates don't know the bit-for-bit original without Cinavia, it's not easy to isolate Cinavia. (If you did know the original, you could use that to first reverse the phase/speed errors Cinavia introduces to obfuscate the data channel, then (in the frequency domain) divide the original by the phase-fixed version of the protected track, and the result is the Cinavia modulation. As it is, a bloody mess of heuristics is required.)
This, on the other hand, is applied on top of whatever the speaker's currently supposed to be outputting, which is not only known to the user, but under the user's control. Therefore, you can set it to output silence, and see whatever's left. Even if it's spread across the entire reproducible frequency range of the sound card, there will be some signal (even if it has the appearance of broadband, low-amplitude noise) where there should be none.
If this malware worked, like Cinavia, by modulating the existing audio signal rather than superimposing its own signal, it would be undetectable when the computer is supposed to be outputting silence (so the above method doesn't work), but it also would be incapable of communicating under the same circumstances. Even if this were the case, we could substitute any known signal (e.g. a pure sine wave at a given frequency) and monitor for variations from the expected output.
Bust out an oscilloscope and a logic analyzer and start looking at these signals.
As long as you have a microphone that can work at those frequencies.
1. As several posters have already pointed out, that's not much of a hurdle.
2. Who said anything about a microphone? The easy way is to connect your scope directly to the electrical signal that drives the speaker -- one channel to each of the compromised machines, and you get the signal you want with no hassle. Unless the software is being particularly clever, you won't even have to crack open the case -- just plug in a Y-cable with an external speaker on one leg and a pigtail with bare wires to clip the scope probes to on the other.
Using a microphone to pick up the signal is just senselessly complicating stuff -- now you've got both sides of a presumably duplex connection* and the sound of you coughing, the click of each knob on the 'scope, all coming in through one mic, and have to figure out what frequency is what? No thanks.
*And if there's more than two compromised machines, you've either got multiple simultaneous duplex connections, or the whole thing runs CSMA like ethernet.
I prefer my 2560x1600 screens in 10" form factor. 27"+ needs to be at least 3840x2160.
I prefer my 3840x2400 screens in 22" form factor. 27"+ needs to be... I dunno, something bigger.
On the bright side, modern 4k displays do have better frame-rate and more convenient inputs than the old beast.
I like to use this ABM1 - Passive Air Band Monitor when flying. I keep it discreet as I'm sure most flight crews won't understand how it's different from a typical radio receiver. I regularly hear that "bzz bzz bzz" of cell phones with this device. I then ask my girlfriend sitting next to me if she put her phone in airplane mode. If she hadn't and does it the noise usually goes away. If she had her's in airplane mode then I assume it's someone else sitting near me.
Correct so far.
Phones do cause interference in the aircraft frequency bands (at least at short range).
And... you go off the rails.
GSM phones cause interference in audio-frequency circuits because the phone transmits in regular bursts every 4.62 ms (this is why it doesn't affect CDMA, UMTS, etc., only 2G GSM/GPRS/EDGE, because they use TDMA). It turns out audio amplifiers generally tend to serve as decent wide-band AM receivers, so this is very easily picked up as a 217Hz buzz. However, this is all happening on the audio-frequency side, so calling it "interference in the aircraft frequency bands" is just plain wrong -- it will affect practically any unshielded or insufficiently-shielded device (it's only a couple watts or so transmitter, but within a meter or less the inverse-square law says you do need better shielding than a lot of consumer electronics have) with an amplified audio output, including all sorts of radios, and non-radio devices from MP3 players to cassette players.
(for more on this, google or start here)
Fortunately, it's a really easy problem to solve -- just keep your phone out of the cockpit. Thanks to the inverse-square law, it's really only a practical issue at very short range.
Same thermoelectric bracelet as two weeks ago.
I'm not saying you don't have a point, but
This technology ... only becomes useful when it is the only way to identify the location of the car. At which point, the drivers will have already bailed.
implies that you believe criminals will routinely either (1) decelerate to a complete stop physics-defyingly fast when out of sight of police or (2) bail out of a moving car.
In reality, it's possible for a pursuing vehicle to maintain a close enough distance to be able to move in when the GPS tracker indicates the fleeing vehicle is slowing, in order to apprehend anyone exiting it, but not staying close enough to avoid the risk of breaking visual contact long enough for the fleeing vehicle to escape (e.g. an intersection after a blind curve, where a single pursuing unit has 50% or less chance of correctly guessing which route the fleeing vehicle took). In this case, the GPS is useful, both to tell when you need to catch back up, and to prevent them from escaping.
Now noting that there is some use for this isn't the same as saying it's worth the expense (both hardware costs and developing procedures and policies for its use), and the present high-speed chase behavior of several local police agencies in my state makes it obvious to me that they absolutely should not have any new toys to distract them, but it's not as dumb a concept as you claim.
You do realize he was objecting to the characterization of the tracking bullets as "James-Bond-Style", right?
Try that with your Seburo!
I would also like to understand what's the definition of "unstable code".
Unstable code is code such that, when you make an arbitrarily small change, you end up rewriting the entire thing.
Stable code, by contrast, is code such that when you make an arbitrarily small change, the code ends up being restored to its original state, or perhaps engaging in a bounded oscillation, where you and another coder keep changing it back and forth with every release.
So, open up cellphones to get their LCD displays to build your own heads-up display?
No, you don't do that, because it makes no bloody sense. You buy LCDs* from (a middleman who buys them from) the same factories phone makers buy them from. Whether you go through a middleman or talk to the factory depends on the volume you're buying and whether or not you have a presence in China.
*note: don't say "Liquid Crystal Display displays" unless you work for your employer, the Department of Redundancy Bureau.
No. 20 years ago we had the Casio databank watch... A calculator, world time clock, and personal contact manager. It wasn't cool back then either, it made you a nerd.
I had a Casio calculator watch (just a boring calculator watch, without the databank stuff) when I was a kid, and I can assure you it didn't make me a nerd. In fact, the causality is exactly opposite; being a nerd made me get a calculator watch.