I wish MS updated their base system more than once every 10 years.
It's quite possible that Apple provides less longevity than MS in terms of OS updates, and I suspect that the future will bear out that supposition, but you can hardly compare the XP->Vista stretch to competitors that released several major updates in the same period.
We can tell you're lazy, because apparently you haven't bothered to Google for "cold boot attacks" yourself.
So now you should Google for "cold boot attacks", read until you understand them, then come back and explain how "standby" (i.e. suspend-to-RAM) is more dangerous than simply be on and running normally (hint: it's not).
Whether the machine is running normally or in standby mode an attacker can power-cycle the machine and access the encryption key from the uncleared RAM. There's literally no difference in the attack if the machine is standby versus normal operating mode. And if you have 10-minute boot times I guarantee that your users are not powering off their machine every time they leave it to get coffee.
The only power mode you have to worry about it suspend-to-disk, because a poorly designed encryption system might not dismount the encrypted volumes before the RAM is copied to disk, which would result in on-disk encryption keys. So if you wanted to disable that power mode, and/or verify that your encryption solution properly dismounts volumes before hibernating, that would be reasonable.
Low internal resistance allows NiMH cells to deliver a near-constant voltage until they are almost completely discharged. This will result in a battery level indicator to overstate the remaining charge if it was designed to read only the voltage curve of alkaline cells. The voltage of alkaline cells decreases steadily during most of the discharge cycle.
The voltage/capacity curve for rechargeable batteries is *much* flatter than for alkaline batteries -- alkaline batteries have a pretty steep slope with a fairly linear relationship between "current voltage" and "remaining power", and devices that expect alkaline batteries and have a low-power indicator calibrated for that curve. Rechargeable batteries have a much flatter curve and the low-power indicator never lights because the trigger voltage is not reached until there's far too little power.
So presumably, devices with built-in rechargeable batteries are properly calibrated for those, and produce a low-power warning at the appropriate time.
You apparently didn't bother to read the link in the parent. It specifically refutes your example and the method under which you claim it operates.
If you'd like to refute the link feel free, but please cite credentials at least as authoritative as C. Wu, Science News, Vol. 153, No. 22, May 30, 1998, p. 341 or Zanotto, E.D. 1998. Do cathedral glasses flow? American Journal of Physics 66(May):392, as the linked page does.
And `dd` can still read EBCDIC, and ASCII maps just fine into UTF-8. So with exactly one very-well-documented conversion via a tool that comes installed on every unix-y system available, you can still read files from the dawn of modern computing. It's been 47 since we invented EBCDIC, and I see no reason to believe that text files will become more difficult to read over the next 50 years. (Not to mention the IBM support for UTF-16 on EBCDIC systems, and limited support in Unicode for UTF-EBCDIC).
Plus the realistic lifespan of these is more like 10 years. You might keep the originals around if you were compelled to do so for legal reasons, but if you haven't transferred to new media and done any requisite format updates by then you don't really care about the data, and wouldn't really mind losing it.
You'd be guilty of manslaughter or negligent homicide, not murder. Murder requires intent. Manslaughter/negligent homicide only requires a willful disregard for life/criminal negligence.
99% of what? The viruses they have definitions for? There's not a product on the market that catches 99% of all viruses.
You might make a comparison of the number of entries in their definitions library, or the different techniques each has available to match the various types of obfuscation in use, but a claim of catching 99% is both meaningless and unsupportable.
The problem with SeaQuest DSV was season 2. Season 1 was promising, but season 2 went all mid-90s-telepathy on us, like B5 season 5. As you noted season 3 wasn't the same show at all, but IMHO it was by far the best. I think they should have re-launched it as a spin-off instead of pretending it was the same show.
We have portable email addresses -- just buy a domain name and setup email hosting. You can buy such services for less than $20/year without any technical expertise, which should not be a significant expense for anyone who cares about keeping their email address indefinitely.
It's not *any* better? I seems about 4000 times better to me.
With plain HTTP anyone in range of my WiFi network can sit passively and catch my credentials without ever even sending a packet. And they could do the same for every single user on my network without any additional work. With encrypted but unauthenticated HTTPS an attacker would need to actively insert themselves into my stream and fake both sides of the transaction to intercept my credentials. And they'd have to maintain a separate session for each additional user on my network.
I agree that protecting against MitM attacks is also a worth goal, but to claim that encryption is useless without authentication is like claiming that locking your door it worthless because anyone with a key could unlock it.
But just adding an untrusted cert would raise the bar from "passive eavesdropping" to "man in the middle attack". It's still an improvement.
Plus self-signed certs can be perfectly safe if you verify the fingerprint. Not that setting up a CA is a bad idea, particularly if you'll have multiple certificates, but it's not strictly necessary.
While I agree that setting up a CA is a superior solution, there's a big difference between "passive eavesdropping" and "man in the middle attack". Conflating the two is at least as bad as pretending that all non-Verisign signatures are self-signatures.
Plus if you're willing to do the "distributed to users" step you could simply distribute the fingerprint from the self-signed key rather than the CA cert. Until and unless you generate additional certificates with the same CA it's actually less work and provides the same security.
It's only the optimal solution for you. If the client choose HTTPS and you change back to HTTP then *you're* deciding that their content shouldn't be encrypted, even if they think it should be. You can choose not to offer HTTPS if you think the burden is too high on your end, but you're lying to yourself by calling it the "optimal" solution for both sides.
You might not care that your web browsing is encrypted. But I might be on a monitored network and don't want my overlords to know that I downloaded a cheesecake recipe because it would ruin their surprise birthday party. That or any of 1,000 other scenarios might lead me to desire encrypted communications even for information that you don't consider worthy of encryption.
Frankly I think *all* communications should be routinely encrypted just to discourage eavesdropping. Plus if encryption became the status quo your browser could offer sane warning messages about unencrypted transfers, rather than putting up no warning for unencrypted transfers and then freaking out when you have an encrypted but unauthenticated transfer.
And decent phones do. On a BlackBerry, for example, you have to specifically authorize each application to access to the voice radio, IP connections (as a whole or per-domain), GPS, address book, etc. It's easy to use and provides great protection, not to mention the instant insight into what a program is actually doing (i.e. "Why does this free calculator want to connect to warez.ru"). Why WindowsCE doesn't do such things is a complete mystery.
To the best of my knowledge, there's no credible evidence that non-ionizing radiation (i.e. radiation that doesn't break bonds), causes anything other than heating/motion. If you absorb some radiant energy but not enough to break bonds you're just going to end up with faster-moving (i.e. hotter) molecules. This is the basic principle behind all forms of radiant heating, from the sun down to your microwave oven.
It's possible that certain cells in certain situations are unusually sensitive to heating in a way that's not obvious, and that such a sensitive might make the related organism respond poorly specific applications of RF radiation, but since everything on the surface of Earth is regularly exposed to a quite a bit of RF radiation from the sun this is typically not a concern, because the sun would have killed such organisms long before we discovered radio.
I was annoyed by the design of the test too (ignoring the obvious methodology flaws in the number of samples/etc.) Why did the inactive cell phones need to be dummies instead of just "off"? What if the bees are simply allergic to the batteries in the real cell phones? The test is obviously intended to examine the effects of the radio waves, since bees are not often in close proximity to cell phones themselves -- wouldn't a better test be to put in identical phones and simply disable the *radio* amplifier in one of the phones, so that the other conditions are as close to identical as practical? Or as you suggested, to simply pound the hives with radio sans any local electronics installation?
As far as we know with other examples of non-ionizing radiation, there are virtually no effects, immediate, delayed, substantial or otherwise.
Even in the case of ionizing radiation, the effects *are* immediate. One might not notice the effects right away if they are mild, but the tissue damage happens when you're exposed, not some time later via radiation time-delay magic.
But your CMS requires you to perform a POST operation before things are actually changed, right? Google doesn't POST anything when they're crawling or checking pages, so as long as your CMS follow the HTTP RFC and uses POST operations for non-idempotent requests, everything should be fine.
Couldn't you also argue their intent was to see what the edit form looked like, not to actually perform an edit? Or more generally, that their intent was simply to index your website and present it to users, not to edit anything in your CMS, even if that was the result of their actions. As you say, it's all about intent, and it's not clear to me at all that Google demonstrated any intent to change your website.
In many cases, the detailed design is appropriate. Consider the case where statute allows a factory polluting a river to be shut down. Under exactly which conditions can this be done? What pollutants count? Which don't? At what levels? What procedure must the agency follow to implement the shut down? Are there exceptions, say, in the interest of national security? Is there an appeals process? If so, what documentation must be submitted and on what schedule? Absent the detailed Congressional design, the agency and/or the courts are going to make it all up as they go along.
Except in reality we create independent agencies -- like the EPA -- to set the specific regulations within the guidelines of their authorizing legislation. And they do make it up as the go along. But that's not necessarily a bad thing -- in many ways, the constitution is authorizing legislation for all three branches of the federal government, and it seems to work pretty well. Is there some reason the same sort of solution wouldn't work in other areas?
More to the point, things like "what count as pollutants" should absolutely not be enshrined in law. Laws take far too long to change and any sort of detailed, prescriptive solutions are bound to be outdated long before the law is repealed or updated. Which is a good part of the reason we've created independent agencies in the first place. It's important that the scope of their authorizing legislation is appropriately narrow, but there's no reason to codify specific requirements about chemicals or radio frequencies or any other such thing.
Finally, I'd suggest that anything too complicated to describe accurately and comprehensively in less than two pages is probably too complex a law to reasonably expect the average citizen to follow, and therefore worthless and perhaps even abusive. I'd be perfectly willing to give you extra space to spell out penalties, remedies, and other post-judgement details, or to provide tabular or other scientific data that might be required for specific compliance, but if the description of what behaviors are prohibited takes more than two pages you're making a law that reasonable people will likely be unable to follow simply due to it's complexity, which is always bad policy.
I find that argument odd -- if the state legislature is corrupt there's already a problem in the state and forcing direct elections can't possibly fix it because the same people you're expecting to choose federal senators already picked the corrupt state legislature. What makes you think they'd do a better job with the federal government than they do in their own state?
I think you're also confusing "governor" with "state legislature", but that's rather beside the point; the same basic logic applies.
Doesn't have the memory/performance to handle a large number of TCP connections, IPSec, gigabit, more than a couple of isolated Ethernet interfaces, etc. There are lots of reasons that a 200 MHz CPU with 8 MB of RAM might not be enough router. They don't necessarily apply to every network or user, but it's absurd to suggest that no such situation exists on home networks.
Or he might have been referring to the economic theory under which we subsidize higher education -- that the money we spend sending people to school will be made up in their additional tax liability and increased economic output after they graduate. That theory only holds if the people we educate actually go on to make more money and have increased economic output, which is not true if they work as a lumberjack.
But hey, if you want randomly bash people without bothering to understand what they're talking about who I am I to stop you.
Some people would consider typing the letters to be "easy access" compared to the click-drag sort of interface the Sibelius offers. And it wouldn't be terribly difficult to rig up a MIDI device to "type" the letters if you wanted to integrate a keyboard. It's not necessarily the right choice for everyone or every project, but it's silly to dismiss it as useless.
You're also missing the possibility of, for example, allowing Sibelius to export to HTML pages for easy sharing with people that don't have the $$$ to drop on such a program. I know you can print from Sibeilus, or export to MIDI, but it doesn't have any option to export an editable file that includes all the typography markup without also requiring an expensive software license.
As for playing sounds, that's not the point of this project, but being able to translate "b5" into a noise is something that's pretty trivial and could easily be added -- it's supported by most BASIC environments as far back as multi-tonal audio was available.
Slashdot Mirror
I wish MS updated their base system more than once every 10 years.
It's quite possible that Apple provides less longevity than MS in terms of OS updates, and I suspect that the future will bear out that supposition, but you can hardly compare the XP->Vista stretch to competitors that released several major updates in the same period.
We can tell you're lazy, because apparently you haven't bothered to Google for "cold boot attacks" yourself.
So now you should Google for "cold boot attacks", read until you understand them, then come back and explain how "standby" (i.e. suspend-to-RAM) is more dangerous than simply be on and running normally (hint: it's not).
Whether the machine is running normally or in standby mode an attacker can power-cycle the machine and access the encryption key from the uncleared RAM. There's literally no difference in the attack if the machine is standby versus normal operating mode. And if you have 10-minute boot times I guarantee that your users are not powering off their machine every time they leave it to get coffee.
The only power mode you have to worry about it suspend-to-disk, because a poorly designed encryption system might not dismount the encrypted volumes before the RAM is copied to disk, which would result in on-disk encryption keys. So if you wanted to disable that power mode, and/or verify that your encryption solution properly dismounts volumes before hibernating, that would be reasonable.
Or to use the better language for Wikipedia:
http://en.wikipedia.org/wiki/Nickel-metal_hydride_battery#Comparison_with_other_battery_types
Low internal resistance allows NiMH cells to deliver a near-constant voltage until they are almost completely discharged. This will result in a battery level indicator to overstate the remaining charge if it was designed to read only the voltage curve of alkaline cells. The voltage of alkaline cells decreases steadily during most of the discharge cycle.
The voltage/capacity curve for rechargeable batteries is *much* flatter than for alkaline batteries -- alkaline batteries have a pretty steep slope with a fairly linear relationship between "current voltage" and "remaining power", and devices that expect alkaline batteries and have a low-power indicator calibrated for that curve. Rechargeable batteries have a much flatter curve and the low-power indicator never lights because the trigger voltage is not reached until there's far too little power.
So presumably, devices with built-in rechargeable batteries are properly calibrated for those, and produce a low-power warning at the appropriate time.
http://www.powerstream.com/AA-tests.htm
You apparently didn't bother to read the link in the parent. It specifically refutes your example and the method under which you claim it operates.
If you'd like to refute the link feel free, but please cite credentials at least as authoritative as C. Wu, Science News, Vol. 153, No. 22, May 30, 1998, p. 341 or Zanotto, E.D. 1998. Do cathedral glasses flow? American Journal of Physics 66(May):392, as the linked page does.
And `dd` can still read EBCDIC, and ASCII maps just fine into UTF-8. So with exactly one very-well-documented conversion via a tool that comes installed on every unix-y system available, you can still read files from the dawn of modern computing. It's been 47 since we invented EBCDIC, and I see no reason to believe that text files will become more difficult to read over the next 50 years. (Not to mention the IBM support for UTF-16 on EBCDIC systems, and limited support in Unicode for UTF-EBCDIC).
Plus the realistic lifespan of these is more like 10 years. You might keep the originals around if you were compelled to do so for legal reasons, but if you haven't transferred to new media and done any requisite format updates by then you don't really care about the data, and wouldn't really mind losing it.
You'd be guilty of manslaughter or negligent homicide, not murder. Murder requires intent. Manslaughter/negligent homicide only requires a willful disregard for life/criminal negligence.
99% of what? The viruses they have definitions for? There's not a product on the market that catches 99% of all viruses.
You might make a comparison of the number of entries in their definitions library, or the different techniques each has available to match the various types of obfuscation in use, but a claim of catching 99% is both meaningless and unsupportable.
The problem with SeaQuest DSV was season 2. Season 1 was promising, but season 2 went all mid-90s-telepathy on us, like B5 season 5. As you noted season 3 wasn't the same show at all, but IMHO it was by far the best. I think they should have re-launched it as a spin-off instead of pretending it was the same show.
Those who fail to understand SMTP are doomed to re-implement it. Poorly.
We have portable email addresses -- just buy a domain name and setup email hosting. You can buy such services for less than $20/year without any technical expertise, which should not be a significant expense for anyone who cares about keeping their email address indefinitely.
It's not *any* better? I seems about 4000 times better to me.
With plain HTTP anyone in range of my WiFi network can sit passively and catch my credentials without ever even sending a packet. And they could do the same for every single user on my network without any additional work. With encrypted but unauthenticated HTTPS an attacker would need to actively insert themselves into my stream and fake both sides of the transaction to intercept my credentials. And they'd have to maintain a separate session for each additional user on my network.
I agree that protecting against MitM attacks is also a worth goal, but to claim that encryption is useless without authentication is like claiming that locking your door it worthless because anyone with a key could unlock it.
But just adding an untrusted cert would raise the bar from "passive eavesdropping" to "man in the middle attack". It's still an improvement.
Plus self-signed certs can be perfectly safe if you verify the fingerprint. Not that setting up a CA is a bad idea, particularly if you'll have multiple certificates, but it's not strictly necessary.
While I agree that setting up a CA is a superior solution, there's a big difference between "passive eavesdropping" and "man in the middle attack". Conflating the two is at least as bad as pretending that all non-Verisign signatures are self-signatures.
Plus if you're willing to do the "distributed to users" step you could simply distribute the fingerprint from the self-signed key rather than the CA cert. Until and unless you generate additional certificates with the same CA it's actually less work and provides the same security.
It's only the optimal solution for you. If the client choose HTTPS and you change back to HTTP then *you're* deciding that their content shouldn't be encrypted, even if they think it should be. You can choose not to offer HTTPS if you think the burden is too high on your end, but you're lying to yourself by calling it the "optimal" solution for both sides.
You might not care that your web browsing is encrypted. But I might be on a monitored network and don't want my overlords to know that I downloaded a cheesecake recipe because it would ruin their surprise birthday party. That or any of 1,000 other scenarios might lead me to desire encrypted communications even for information that you don't consider worthy of encryption.
Frankly I think *all* communications should be routinely encrypted just to discourage eavesdropping. Plus if encryption became the status quo your browser could offer sane warning messages about unencrypted transfers, rather than putting up no warning for unencrypted transfers and then freaking out when you have an encrypted but unauthenticated transfer.
And decent phones do. On a BlackBerry, for example, you have to specifically authorize each application to access to the voice radio, IP connections (as a whole or per-domain), GPS, address book, etc. It's easy to use and provides great protection, not to mention the instant insight into what a program is actually doing (i.e. "Why does this free calculator want to connect to warez.ru"). Why WindowsCE doesn't do such things is a complete mystery.
To the best of my knowledge, there's no credible evidence that non-ionizing radiation (i.e. radiation that doesn't break bonds), causes anything other than heating/motion. If you absorb some radiant energy but not enough to break bonds you're just going to end up with faster-moving (i.e. hotter) molecules. This is the basic principle behind all forms of radiant heating, from the sun down to your microwave oven.
It's possible that certain cells in certain situations are unusually sensitive to heating in a way that's not obvious, and that such a sensitive might make the related organism respond poorly specific applications of RF radiation, but since everything on the surface of Earth is regularly exposed to a quite a bit of RF radiation from the sun this is typically not a concern, because the sun would have killed such organisms long before we discovered radio.
I was annoyed by the design of the test too (ignoring the obvious methodology flaws in the number of samples/etc.) Why did the inactive cell phones need to be dummies instead of just "off"? What if the bees are simply allergic to the batteries in the real cell phones? The test is obviously intended to examine the effects of the radio waves, since bees are not often in close proximity to cell phones themselves -- wouldn't a better test be to put in identical phones and simply disable the *radio* amplifier in one of the phones, so that the other conditions are as close to identical as practical? Or as you suggested, to simply pound the hives with radio sans any local electronics installation?
As far as we know with other examples of non-ionizing radiation, there are virtually no effects, immediate, delayed, substantial or otherwise.
Even in the case of ionizing radiation, the effects *are* immediate. One might not notice the effects right away if they are mild, but the tissue damage happens when you're exposed, not some time later via radiation time-delay magic.
But your CMS requires you to perform a POST operation before things are actually changed, right? Google doesn't POST anything when they're crawling or checking pages, so as long as your CMS follow the HTTP RFC and uses POST operations for non-idempotent requests, everything should be fine.
Couldn't you also argue their intent was to see what the edit form looked like, not to actually perform an edit? Or more generally, that their intent was simply to index your website and present it to users, not to edit anything in your CMS, even if that was the result of their actions. As you say, it's all about intent, and it's not clear to me at all that Google demonstrated any intent to change your website.
In many cases, the detailed design is appropriate. Consider the case where statute allows a factory polluting a river to be shut down. Under exactly which conditions can this be done? What pollutants count? Which don't? At what levels? What procedure must the agency follow to implement the shut down? Are there exceptions, say, in the interest of national security? Is there an appeals process? If so, what documentation must be submitted and on what schedule? Absent the detailed Congressional design, the agency and/or the courts are going to make it all up as they go along.
Except in reality we create independent agencies -- like the EPA -- to set the specific regulations within the guidelines of their authorizing legislation. And they do make it up as the go along. But that's not necessarily a bad thing -- in many ways, the constitution is authorizing legislation for all three branches of the federal government, and it seems to work pretty well. Is there some reason the same sort of solution wouldn't work in other areas?
More to the point, things like "what count as pollutants" should absolutely not be enshrined in law. Laws take far too long to change and any sort of detailed, prescriptive solutions are bound to be outdated long before the law is repealed or updated. Which is a good part of the reason we've created independent agencies in the first place. It's important that the scope of their authorizing legislation is appropriately narrow, but there's no reason to codify specific requirements about chemicals or radio frequencies or any other such thing.
Finally, I'd suggest that anything too complicated to describe accurately and comprehensively in less than two pages is probably too complex a law to reasonably expect the average citizen to follow, and therefore worthless and perhaps even abusive. I'd be perfectly willing to give you extra space to spell out penalties, remedies, and other post-judgement details, or to provide tabular or other scientific data that might be required for specific compliance, but if the description of what behaviors are prohibited takes more than two pages you're making a law that reasonable people will likely be unable to follow simply due to it's complexity, which is always bad policy.
I find that argument odd -- if the state legislature is corrupt there's already a problem in the state and forcing direct elections can't possibly fix it because the same people you're expecting to choose federal senators already picked the corrupt state legislature. What makes you think they'd do a better job with the federal government than they do in their own state?
I think you're also confusing "governor" with "state legislature", but that's rather beside the point; the same basic logic applies.
Doesn't have the memory/performance to handle a large number of TCP connections, IPSec, gigabit, more than a couple of isolated Ethernet interfaces, etc. There are lots of reasons that a 200 MHz CPU with 8 MB of RAM might not be enough router. They don't necessarily apply to every network or user, but it's absurd to suggest that no such situation exists on home networks.
Or he might have been referring to the economic theory under which we subsidize higher education -- that the money we spend sending people to school will be made up in their additional tax liability and increased economic output after they graduate. That theory only holds if the people we educate actually go on to make more money and have increased economic output, which is not true if they work as a lumberjack.
But hey, if you want randomly bash people without bothering to understand what they're talking about who I am I to stop you.
Some people would consider typing the letters to be "easy access" compared to the click-drag sort of interface the Sibelius offers. And it wouldn't be terribly difficult to rig up a MIDI device to "type" the letters if you wanted to integrate a keyboard. It's not necessarily the right choice for everyone or every project, but it's silly to dismiss it as useless.
You're also missing the possibility of, for example, allowing Sibelius to export to HTML pages for easy sharing with people that don't have the $$$ to drop on such a program. I know you can print from Sibeilus, or export to MIDI, but it doesn't have any option to export an editable file that includes all the typography markup without also requiring an expensive software license.
As for playing sounds, that's not the point of this project, but being able to translate "b5" into a noise is something that's pretty trivial and could easily be added -- it's supported by most BASIC environments as far back as multi-tonal audio was available.