Forensic scientitsts have been saying for a while that DNA evidence isn't absolute proof - and nor are fingerprints.
The reason is that only a few key DNA and fingerprint points are taken, so there's a small (but finite) probablity of matching on an innocent party.
This hasn't mattered in the past beause the amount of DNA on file has been small (which means that someone innocent could be prosecuted because he has a "bad" history - aka "round up the usual suspects") but as the DNA record pool grows the probability of a false match grows.
Ditto fingerprints. There have been a number of known cases where a match has been found to someone who died years ago. What's "good enough" for court evidence, isn't actually good enough anymore - much like your old DES crypts in a moder environment.
Speaking of fingerprints - identical twins have differing fingerprints, so if they have this dilemma then there are clearly none collected - which smacks of sloppy work in the first place.
Of course your average USA prosecutor doesn't want to know this stuff even if it results in false convictions and massive payouts later on. Hopefully the french beaks are somewhat more clued up.
There have already been at least 2 documented cases of internally carried explosives on a suicide bomber - whilst they didn't detonate on aircraft, at least one of them flew to his destination and the explosive weren't detected enroute.
You many not be able to tell where they are while sitting in your car, but any regular user of a road will tell you exactly where the black ice patches are most likely to form (as can the engineers).
Indicating high risk areas can already be done with roadside signs and in extreme cases roads have been known to be fitted with heating circuits. having the indicator on the road surface itself is a nice idea but I suspect it's not likely to have wide uptake.
As a bonus, scada virus won't result in rapidly escalating nuke exchanges. The scenario for such events now is actually worse than in the cold war days.
For all the USA military is sabre rattling about cyber-attacks. any such attack on a foreign nation which has ample resources to retaliate is a non-starter. It's much easier to simply arrange for top tier carriers to blackhole the ASNs involved.
This is already done routinely, although the usual target is hijacked netblocks and the ASNs advertising their routes. Moving to deal with attacks across the network, vs attacks on the integrityof the network will take a lot more pushing and shoving at high levels.
All it'd take to knock out North American (or European, or Asian, etc) infrastructure is a few well-aimed strikes at the power distribution inftrastructure.
The choices are:
Conventional precision strikes on major transformers (the same ones everyone's worrying which would be knocked out in a major solar flare)
A suitable yield stratospheric nuclear airburst - why create masses of fallout and civilian deaths when you can simply wipe out half a continent's worth of electrical grid in one go with a good-size EMP?
FWIW, (NOT in the USA) the differebce between Huawei 10Gb/s TRILL kit and the equivalently specified Cisco stuff is a factor of 6-8 after discounts are offered (and significantly higher if list prices are looekd at )
Huawei's performance levels are significantly higher. I am a CCNA by necessity but IMO Cisco trade on their name, not on their acual performance. It's a bit like the old "Nobody got fired for buying IBM" mantra.
Not that I agree with the vast majority of what they're used for, but reputation.com is one the outfits that springs to mind.
Definietely go on the trademarking of one's own name then yank the domain on a uDRP complaint.
If you want to bury HVDC lines, be my guest. It's not a trivial undertaking.
The issue isn''t burying urban cabling. That's a well-known, fairly straightforward job and deals nicely with localised issues.
Burying cross-country distribution cables and switchyards is a whole 'nother ballgame and not even the europeans do it, except in extreme cases. (I live in europe).
I can already buy SSD DOM (disk on module) only slightly larger than the SATA connector in sizes up to 128Gb. They aren't particularly cheap but they're a great solution for diskless setups (fewer cables) where the main requirement is a read-mostly device and most storage is on the LAN.
Which is why the first thing you must do after a sucessful revolution is kill the leader if he won't step down.
History has plenty of examples of what happens when you don't.
Krakatoa wasn't even close to being "super" - it only just qualifies as "major". Taupo was 20 times bigger when it blew ~1900 years ago and that was tiny compared to the eruption there ~20k years ago which created the caldera visible today. (The current lake in Taupo's caldera could easily accomodate Singapore Island and it only occupies about 1/3 of the caldera's area).
Rule of thumb: Cone-shaped volcanoes are relatively harmless as long as you're not in the immediate vicinity when they erupt.
Moving back to (A)GW, the reason that there are so many arguments is that the calculated warming from the known inputs don't match observations (what's observed is higher than it should be). Removing all recording stations where heat island effects have happened in the last 70 years (paving runways/aprons near airfield thermometers and rural airfields being swallowed up by suburbia) helps a bit, but the observations still don't match any of the theories closely enough.
It doesn't help that fossil evidence has shown the poles to have gone ice-free several times during the current ice-age cycles (last 12 million years), and the arctic has shown signs of previously going ice-free during cooling trends (Ice extent was believed to be a lot lower 1000 years ago, f'instance). We haven't been observing long enough to know if things are due to AGW or various long-term natural cycles (Atlantic hurricanes were unnaturally low in occurance and intensity for the 60 years up to 2005, f'instance. Is a return to "normal" because of AGW or just business as usual?)
Precautionary moves aren't a bad idea but there are extremists at both ends who end up discrediting the entire field.
Out of the hobbyist arena, money becomes available for hardware.
This leads to the problem of people who think IDL is a programming language (I have to work with them), especially in the astrophysics arena. There would be a lot of milage in forcing anyone who has to write code to actually take a few programming courses.
This discussion thread is about consumer PC motherboards, not servers. There are a whole universe of other issues in that arena (It's my dayjob) which most "enthusiasts" (more-or-less an industry euphemism for "gamers") wouldn't even think of.
For single processes you're right, however Operating Systems are seldom a single process - unless it's DOS.
The kind of code you're working on is not going to be run on a consumer-class desktop motherboard or even an "enthusiast PC" unless someone's incredibly masochistic.
Servers won't be going away anytime soon and my comments are related to consumer-class systems.
The ONLY time I ever upgrade a CPU ot ram in a board is when it's well beyond end of life and I need to eke out some extra life from the thing.
In some cases upgrading old boards costs more than a new machine, especially old workstations with FB-RAM. Why bump an old box to 16Gb if you can buy an entirely new machine which is 20 times faster, uses far less power and has 32Gb onboard for 80% of that cost? (This is a real-life comparison FWIW)
Why? I can change out a complete board in less than 10 minutes(*) - and more to the point, most businesses buy systems with support which means that if something fails, it get replaced,
(*) It's easier to changeout a board than a cpu, Seriously.
"enthusiasts" are about 0.0001% of the market.
Homebuilders and whiteboxers are somewhat more - but they won't be complaining too loudly about fixed CPUs as long as everything works as expected.
As for speed, it's a matter of appropriate tools. Windows 7 won't run on ARM, but win8 will and one can easily trade raw speed for core numbers - and to be honest, internal clock speeds are far less important than access speed of the memory and PCIe bus.
(I run Linux exclusively. What CPU the platform has isn't that important as long as it works for what I need it to do)
There are already enough laws on the books about cost-shifted advertising in most countries that any outfit who tried to get ad blockers taken off the market could force regulatory hands enough to implode the entire business.
When a website owner pays all my costs, then he has the option of forcing ads on me.
It's worth noting that ad-supported business models don't usually do too well on the internet. They're tolerated up to a point but pushing beyond that results in the targets kicking back - hard.
"Not that I particularly oppose legalising and regulating drugs, but how does having commercial drugs (hence, not free) prevent a crack head from stealing stuff to pay for the drugs?"
Mainly because the reason crack is sold at all is because there's huge profit to be made from it (tens of thousands of percent at street pricing vs actual production costs). Legalised, regulated and taxed drug supplies would cut off the supply of tax-free money to gangs and enable addiction treatment to be handled as a health issue.
Sensible handling of such issues doesn't feature highly on any politician's resume, anywhere in the world.
most hot, longlived leftovers can be put back in the reactor and reused . There are a bunch of reason why they're not, but they're mainly political.
Curent nuke tech is very wasteful. Don't bury it too deep or our great ^ N grandchildren won't be happy when they've worked out a good way of using it.
Getting back to fences - it's a nice idea but havng sen what rural copper thieves do when confronted with a zapper, I don't think it will work as well as intended.
The actual cost is something less than 1c. SE Asian telcos are still able to make a profit at those levels. (1 second of voice transmission is about the same network data as 5 SMS messages and they're not time-critical. Do the math)
Slashdot Mirror
Forensic scientitsts have been saying for a while that DNA evidence isn't absolute proof - and nor are fingerprints.
The reason is that only a few key DNA and fingerprint points are taken, so there's a small (but finite) probablity of matching on an innocent party.
This hasn't mattered in the past beause the amount of DNA on file has been small (which means that someone innocent could be prosecuted because he has a "bad" history - aka "round up the usual suspects") but as the DNA record pool grows the probability of a false match grows.
Ditto fingerprints. There have been a number of known cases where a match has been found to someone who died years ago. What's "good enough" for court evidence, isn't actually good enough anymore - much like your old DES crypts in a moder environment.
Speaking of fingerprints - identical twins have differing fingerprints, so if they have this dilemma then there are clearly none collected - which smacks of sloppy work in the first place.
Of course your average USA prosecutor doesn't want to know this stuff even if it results in false convictions and massive payouts later on. Hopefully the french beaks are somewhat more clued up.
There have already been at least 2 documented cases of internally carried explosives on a suicide bomber - whilst they didn't detonate on aircraft, at least one of them flew to his destination and the explosive weren't detected enroute.
The security theatre arguably puts more people at risk thanks to the chokepoints it generates.
All it takes is one suicide bomber to walk into the waiting crowd - and this already happened once in Moscow.
You don't even need a bomber. Merely spreading panic in a crowded environment is a potent weapon all by itself.
Civil cases don't need reasonable doubt, just balance of evidence. These kinds of cases are about who's got deeper pockets.
Indicating high risk areas can already be done with roadside signs and in extreme cases roads have been known to be fitted with heating circuits. having the indicator on the road surface itself is a nice idea but I suspect it's not likely to have wide uptake.
For all the USA military is sabre rattling about cyber-attacks. any such attack on a foreign nation which has ample resources to retaliate is a non-starter. It's much easier to simply arrange for top tier carriers to blackhole the ASNs involved.
This is already done routinely, although the usual target is hijacked netblocks and the ASNs advertising their routes. Moving to deal with attacks across the network, vs attacks on the integrityof the network will take a lot more pushing and shoving at high levels.
The choices are:
Conventional precision strikes on major transformers (the same ones everyone's worrying which would be knocked out in a major solar flare)
A suitable yield stratospheric nuclear airburst - why create masses of fallout and civilian deaths when you can simply wipe out half a continent's worth of electrical grid in one go with a good-size EMP?
USA employers solve this by requiring workers to find their own damned accomodation.
Any real incentive to "buy american" went away when the social contract inherent in the New Deal was torn up.
FWIW, (NOT in the USA) the differebce between Huawei 10Gb/s TRILL kit and the equivalently specified Cisco stuff is a factor of 6-8 after discounts are offered (and significantly higher if list prices are looekd at ) Huawei's performance levels are significantly higher. I am a CCNA by necessity but IMO Cisco trade on their name, not on their acual performance. It's a bit like the old "Nobody got fired for buying IBM" mantra.
Not that I agree with the vast majority of what they're used for, but reputation.com is one the outfits that springs to mind. Definietely go on the trademarking of one's own name then yank the domain on a uDRP complaint.
If you want to bury HVDC lines, be my guest. It's not a trivial undertaking.
The issue isn''t burying urban cabling. That's a well-known, fairly straightforward job and deals nicely with localised issues.
Burying cross-country distribution cables and switchyards is a whole 'nother ballgame and not even the europeans do it, except in extreme cases. (I live in europe).
I can already buy SSD DOM (disk on module) only slightly larger than the SATA connector in sizes up to 128Gb. They aren't particularly cheap but they're a great solution for diskless setups (fewer cables) where the main requirement is a read-mostly device and most storage is on the LAN.
Try using native zfs instead of zfs-fuse - just make sure you have enough ram if you want to futz around with dedupication.
Which is why the first thing you must do after a sucessful revolution is kill the leader if he won't step down. History has plenty of examples of what happens when you don't.
Krakatoa wasn't even close to being "super" - it only just qualifies as "major". Taupo was 20 times bigger when it blew ~1900 years ago and that was tiny compared to the eruption there ~20k years ago which created the caldera visible today. (The current lake in Taupo's caldera could easily accomodate Singapore Island and it only occupies about 1/3 of the caldera's area).
Rule of thumb: Cone-shaped volcanoes are relatively harmless as long as you're not in the immediate vicinity when they erupt.
Moving back to (A)GW, the reason that there are so many arguments is that the calculated warming from the known inputs don't match observations (what's observed is higher than it should be). Removing all recording stations where heat island effects have happened in the last 70 years (paving runways/aprons near airfield thermometers and rural airfields being swallowed up by suburbia) helps a bit, but the observations still don't match any of the theories closely enough.
It doesn't help that fossil evidence has shown the poles to have gone ice-free several times during the current ice-age cycles (last 12 million years), and the arctic has shown signs of previously going ice-free during cooling trends (Ice extent was believed to be a lot lower 1000 years ago, f'instance). We haven't been observing long enough to know if things are due to AGW or various long-term natural cycles (Atlantic hurricanes were unnaturally low in occurance and intensity for the 60 years up to 2005, f'instance. Is a return to "normal" because of AGW or just business as usual?)
Precautionary moves aren't a bad idea but there are extremists at both ends who end up discrediting the entire field.
Out of the hobbyist arena, money becomes available for hardware.
This leads to the problem of people who think IDL is a programming language (I have to work with them), especially in the astrophysics arena. There would be a lot of milage in forcing anyone who has to write code to actually take a few programming courses.
This discussion thread is about consumer PC motherboards, not servers. There are a whole universe of other issues in that arena (It's my dayjob) which most "enthusiasts" (more-or-less an industry euphemism for "gamers") wouldn't even think of.
For single processes you're right, however Operating Systems are seldom a single process - unless it's DOS. The kind of code you're working on is not going to be run on a consumer-class desktop motherboard or even an "enthusiast PC" unless someone's incredibly masochistic. Servers won't be going away anytime soon and my comments are related to consumer-class systems.
If you want serious number crunching, you use GPGPU solutions. Everything else is just gaming.
The ONLY time I ever upgrade a CPU ot ram in a board is when it's well beyond end of life and I need to eke out some extra life from the thing.
In some cases upgrading old boards costs more than a new machine, especially old workstations with FB-RAM. Why bump an old box to 16Gb if you can buy an entirely new machine which is 20 times faster, uses far less power and has 32Gb onboard for 80% of that cost? (This is a real-life comparison FWIW)
Why? I can change out a complete board in less than 10 minutes(*) - and more to the point, most businesses buy systems with support which means that if something fails, it get replaced, (*) It's easier to changeout a board than a cpu, Seriously.
"enthusiasts" are about 0.0001% of the market. Homebuilders and whiteboxers are somewhat more - but they won't be complaining too loudly about fixed CPUs as long as everything works as expected. As for speed, it's a matter of appropriate tools. Windows 7 won't run on ARM, but win8 will and one can easily trade raw speed for core numbers - and to be honest, internal clock speeds are far less important than access speed of the memory and PCIe bus. (I run Linux exclusively. What CPU the platform has isn't that important as long as it works for what I need it to do)
There are already enough laws on the books about cost-shifted advertising in most countries that any outfit who tried to get ad blockers taken off the market could force regulatory hands enough to implode the entire business. When a website owner pays all my costs, then he has the option of forcing ads on me. It's worth noting that ad-supported business models don't usually do too well on the internet. They're tolerated up to a point but pushing beyond that results in the targets kicking back - hard.
Mainly because the reason crack is sold at all is because there's huge profit to be made from it (tens of thousands of percent at street pricing vs actual production costs). Legalised, regulated and taxed drug supplies would cut off the supply of tax-free money to gangs and enable addiction treatment to be handled as a health issue.
Sensible handling of such issues doesn't feature highly on any politician's resume, anywhere in the world.
most hot, longlived leftovers can be put back in the reactor and reused . There are a bunch of reason why they're not, but they're mainly political. Curent nuke tech is very wasteful. Don't bury it too deep or our great ^ N grandchildren won't be happy when they've worked out a good way of using it. Getting back to fences - it's a nice idea but havng sen what rural copper thieves do when confronted with a zapper, I don't think it will work as well as intended.
The actual cost is something less than 1c. SE Asian telcos are still able to make a profit at those levels. (1 second of voice transmission is about the same network data as 5 SMS messages and they're not time-critical. Do the math)