I've only seen two. One is that there was a company Ryan McFarland that made a COBOL compiler for Unix named "rmcobol" - everybody thought that sounded like a good idea.
The other came out on Usenet in the 80s and went something like
Hey Rocky, watch me pull a computer program out of my hat! Oh, Bullwinkle, that trick never works! 100 PROCEDURE DIVISION.... [couple more lines like that] Guess I oughta get another hat!
Some of my coworkers got it, but a couple of them didn't. The disturbing part was that they recognized the Cobol program, but were too young to recognize Rocky and Bullwinkle...
When Karl says he's going to spend time with his family, it's mainly his half-brother Cthulhu and his cousin Hastur. Just be careful not to say his name three times or he'll be back....
In some cases there's half an excuse for government to track ethnicity along with other physical characteristics, e.g. if the picture on your ID card shows your white face, blond hair, and blue eyes, and the data fields in the card say you're black with brown eyes and black hair, that's a hint that the card's been tampered with. And sometimes there are other very specialized reasons for tracking it, such as (in the US) if you're a registered member of an Indian tribe or Native Hawaiian, then there's some recently-stolen land that the government keeps track of your claim to.
But for the most part, tracking ethnicity is a spectacularly bad idea, and tracking religion is not only worse but also much less reliable because it's not a constant. Ignore the questions of whether your mom thinks you're still a good Catholic and can look up whether you've shown up at Mass lately or see that you visited the Zendo down the street or guess about that party on Solstice (or, ahem, May Day) where most of your friends are pagans... (and at least the database won't let your mom see whether you're gay, probably because in China that's Just Not Talked About.)
Ethnic cleansing is a lot easier if you've got a database saying where all the members of Ethnic Group X live in your town*. And if you don't want to hire Jews, you've got a database that says who they are. And it's much easier to get the no-fly list right when you can tell if somebody named Malik Muhammed is African-American (ok, he's just one of Those People, and make sure to give him the kosher airline meal) or Arab (he goes on the list.) I'm sure China has their equivalent issues about which ethnicities get privileges and which don't, even outside of special provinces like Tibet.
And at least it's just China - in most of the Moslem countries, if you weren't born a Moslem, that's usually ok, though you might have to pay a tax, but if you were born a Moslem and you've converted to other religions, Sharia says you have to either convert back or die, and even in the more moderate countries like Egypt, they'll throw ex-Moslems in jail for preaching the wrong religion, as happened to friends-of-friends of mine back in the 80s. In some countries they've got Sunni-vs-Shia issues that are better off not having database support, and I don't know what happens if you convert to another branch of Islam such as Sufiism. And the Baha'i also seem to be a special always-infidel case.
* I get to mention the ethnic-cleansing-rounding-up-Ethnic-Group-X example without triggering Godwin's Law this time - one of the people at the party I was at last night was talking about how he didn't know two of his grandparents because they didn't make it back from the Japanese-American internment camps during the war (he wasn't doing a political rant - he was dealing with his aging mother's house, where there's still stuff of his other grandmother's as well.) One of the joys of living in California is the wide variety of people you get to be with - most of the examples I gave above are for people I've seen recently, though a few were people I hadn't seen in a while and one or two were from recent news.
* And the one person I've known who *was* a potential suicide bomber is presumably not on the no-fly list; he was a college student in Japan during the war, and was considering volunteering to be a kamikaze pilot, but one of his professors talked him out of it. I worked for him about 30 years ago - his war history didn't stop him from having a US security clearance. (Do I still get didn't-trigger-Godwin's-Law credit if I mention the tattoo his boss had on his arm? Probably not. )
Yes, it's possible that a dinosaur killer could hit us today. But the most likely things to kill us all are global nuclear war (less likely than 20 years ago, but it could still happen) or wrecking the ecology of the planet. But we're not even vaguely close to having sustainable closed ecosystems that we could run in the asteroid belts or on Mars, which are the most likely nearby backup sites, and we're not close to having the physics to build a starship. We're not even close to the point of being able to put a bunch of backup humans into a salt-mine to survive a nuclear war, even if we don't get into a salt-mine race with the Russians or Chinese.
And yes, the rate we're using up resources and growing population is making it harder to sustain the energy needed for space travel - that's one of the things we need to figure out how to fix.
We've got a *lot* of time to get off the planet before the sun flames out, and even the average amount of time between dinosaur-killer-sized asteroid hits is millions of years. On the other hand, we're a long long way from being able to move any significant fraction of the population into space, and we won't succeed at that if we all die from a messed-up planet first.
The two activities overlap significantly - a critical skill we need to learn for surviving in space is how to run a viable ecosystem, whether it's on a closed-system spaceship or a terraformed planet. So far we've only run a few small closed-system terrarium experiments like the Biosphere (which had to cheat and bring in extra oxygen, something that's only easy to do when you're on a working planet) - even non-closed-system spacecraft like the Space Station have been getting weird mold problems we don't know how to manage well. And we've got one experiment running on terraforming a planet (Earth) which is going pretty badly at the present time - we don't even know how the thermostat works yet. So we're going to need to learn to fix planets before we can get off this one, and the best way to learn that is by trying to fix this planet.
Also, the energy requirements for getting lots of people off the planet are amazingly high; we're decades away from building even space elevators, much less mass-production rockets, and since we don't know how to run portable ecosystems yet, it doesn't make sense to give high priority to the transport parts; we can let Moore's Law crank for another century or two just fine.
There are one or maybe two exceptions to that - satellites studying and observing the Earth are really useful in learning how to fix the planet, and we can launch those with our current low technology. Unlike other parts of the space program, which have given us powdered orange drink and better military missiles by diverting scientists and engineers from making better commercial aircraft or more efficient automobiles, the satellite part of the space program may have been a big win. Also, power satellites *might* be useful as an alternative to carbon-fuel or nuclear energy, and it might make sense to work on them early, but that'll take a lot of earth-based design to show whether it might be feasible.
If they've got warrants, it's legal*, just dirty. Or if they've got FISA court permission. If they were doing a strictly partisan political attack against Democrats, it might be dirty enough to actually be illegal, but investigating a former Administration official for possibly leaking military secrets is ostensibly the responsible thing for them to do.
If the information that was allegedly possibly leaked had been enough for somebody to actually prosecute some Executive Branch people (whether FBI or Pentagon or NSA or whatever) and they'd gotten convicted already, *then* this kind of raid might count as "obstruction of justice", but they're acting sufficiently proactively that they'll at least get away with it until the Bush Administration is out of office. And probably after that as well.
------
* If they don't have warrants or permission, well, the Bush Administration thinks it's legal anyway, and they've got a Justice Department who wouldn't prosecute them for doing it and they've stacked the Federal courts as well.
The Bush/Cheney Administration has spent the last 6+ years building an organizational, legal, and technical infrastructure for Executive Branch power, including anything from wiretap infrastructures to the Patriot Act to stuffing the courts and Justice Department with pro-executive-power people, and getting states, banks, credit companies, airlines, etc. to do massive data collection. And it's not like it started with them - the FBI wiretap enthusiasts like Louis Freeh, the NSA anti-public-crypto people, the Echelon project, etc. all date to the Clinton or GHWBush/Reagan administrations or earlier.
It's going to take a *long* time to tear down that stuff and turn this back into America again, and most of that won't happen unless we replace the current Executive Branch with one that's actually committed to doing it. Most of the major candidates aren't talking like that - certainly Hillary and Rudy and John Edwards and McCain and Romney don't have a history of wanting to do that, and you're pretty much down to Dennis Kucinich and Ron Paul before you'd get to anybody who'd talk about that kind of concept as a campaign strategy. Perhaps if the Democrats not only win the White House but also increase their control of the Senate and House they'll have some willingness to do that after a couple of years.
For now, though, Homeland Security Anonymous Spokescritters report that Enhanced Terrorist Surveillance Program has been reporting increased frequency of terrorist chatter saying "Booga Booga", so if you're even suggesting that we decrease wiretapping then you're a threat to national security and our precious bodily fluids.
> > The other popular reason for getting PI space is to make it easier to renumber if you change ISPs
> I think you mean "unnecessary" not "easier".
"Unnecessary" is *such* an optimistic term (though for that matter, so is "easier":-). If everything goes really well, sometimes it even applies, but I wouldn't call it 100%.
> > IPv4 PI space is seriously non-scalable, and you can't simply duplicate it in IPv6.
Why not? IPv6 has bigger addresses, that's the whole point of it.
The problem isn't just raw addresses - it's easy to duplicate that. The problem is the scalability of the routing system itself - how many routes does an internet backbone router have to keep track of? It's gone from 100K to 200K in the last couple of years, and demand is growing.
A single-homed site that uses PA space doesn't add any global routes, just fills up it's provider's space a bit more.
A dual-homed site using PI space adds two global routes, one through each ISP.
And a dual-homed site using PA space still adds at least one route (If the PA belongs to ISP#1, then it's adding a ISP#2 route to its subnet of ISP#1 space, and in practice it often adds two routes, including the more-specific route for its subnet of ISP#1 space on ISP#1.)
IPv6 was supposed to fix this problem by getting everybody to use address space in a hierarchical fashion, though I'm afraid it was a combination of wishful thinking and vigorous hand-waving rather than reflecting reality. IPv6 does have enough space to hack around it a bit, e.g. ISP#1 and ISP#2 could get a/32 block that they use to assign all of the customers who are homed to both of them, but that takes N**2 of those blocks to support N ISPs, which isn't too bad if you just support (say) a cabal of US Tier1 ISPs (about 25 of them => 625 blocks), but there are 5-10000 smaller ISPs in the US (you could force the small ISPs to get dual-home space from a Tier 1 or something, but that would take really entertaining politics to make that happen.)
There have also been proposals to manage IPv6 address space geographically, which would have similar scaling wins (e.g. split the US into 1000 pieces by area code or 50 states or whatever), but it confuses the routing and peering structures because not all ISPs connect in all regions (probably works better in Europe where exchanges handle more of the traffic.)
> The big problem with renumbering for me is the idiots with their idiot firewalls
Yup. That matches my experience, though for me its usually the VPNs (separate or integrated into the firewall) rather than the blocking rules themselves.
It's really not that bad - I just helped do that for a customer last week, which required updating routers, their servers, our servers, a VPN tunnel server, and some management equipment. That was changing from a/27 to a/25. In another couple of weeks, we'll need to do it again, because once we got done they found there were some issues with how VMware got along with their applications, so instead of having a few dozen addresses for the host machines, they'll need a few thousand addresses for the virtual machines, but that basically forces them to use another router which they should have done anyway:-)
DNS complicates things a bit, but if you give yourself more than a week for planning you can set your DNS timers appropriately. Most hosts and routers can do just fine with secondary IP addresses, so you can add the secondary addresses to the machines a while before telling the users to use them, and most other addresses can be managed by DHCP.
ECC memory is available, for prices not significantly higher that non-ECC memory (perhaps slightly annoyingly higher for consumer PCs, but it's more like 50% higher, not 5000% higher.) And the consumer stuff does have lifetime warranties, if you buy if from the better brands, and in practice DRAM reliability is *far* better than five-nines uptime unless you've got serious heat or power-quality problems. Router vendors aren't reimbursing you major cash if your router crashes, unless you consider the price of the RAM they're replacing to be "major cash" (in which case it's only major because they charge way too much) - they're not going to reimburse you for network downtime costs, just repair costs.
The price differences are even more egregious for Flash (though the prices are a lot lower, e.g. $700 for a 512MB upgrade vs. $10 commercial), because the stuff is basically only used for storing the OS and booting the machine - good camera memory is fine, and you're not constantly overwriting it so you're not going to burn through write cycles.
We've got a Tivo with a DVD burner. My wife watches more TV than I do, and records more TV than she gets around to watching, so she's mainly in charge. She leaves my shows on the recorder (with occasional snide remarks about Grey's Anatomy being a soap opera (which I can't actually refute:-)), saves the stuff we both like for when we both want to watch TV, and if the drive gets full, I'll occasionally burn a DVD of stuff she hasn't watched yet (or she'll burn DVDs of movies.)
I mostly use my work laptop, and I set up the home desktop; my wife's laptops from her old consulting business are obsolete, so she uses the desktop these days. In theory I could boot it in Linux, but in practice it runs XP with three users, me, her, and root, in theory with only root having admin privileges. In practice, she runs the important Windows application (Turbotax) and installs interesting software and has admin privileges, even though she's mostly browsing, while I run it occasionally for iTunes (which unfortunately had to be installed by root, and gets confused about what's mine and what's root's.)
Dude! You've got an amazingly secure briefcase, with 8 digits! Mine only has 4 digits, and nobody'd ever guess whose birthday the password is (oops...)
Back to reality, though, it's amazing how many Unix passwords were "abc123", back when our systems required at least six characters including some non-letters:-)
Depends on whether the vertical hold was adjusted correctly....
Last time I displayed computer output on a TV was around 1995. It wasn't intentional - I was visiting my parents, and the signal from my laptop showed up on their TV. It wasn't quite in sync - there were about three copies of part of the text, scrolling slowly vertically. But it was semi-readable, and was definitely enough to answer the discussions about "so can you avoid TEMPEST eavesdropping by using a laptop?" I suspect the leakage was coming from the CRT port on the back of the PC, as opposed to from the LCD monitor electronics.
Recent Orangutan Article - Among the great apes, orangutans are probably the least like humans (as opposed to bonobos, who are even closer than the common chimp.) But they do have some similar communication patterns - some of the recent research talks about them using charades as a way to convey ideas, though they usually don't get quite as far as "third syllable sounds like ____". The researchers commented that if Orangs can do that, probably the more human-like great apes can too.
We and our fellow apes are related to the other primates; Wikipedia says that there's some disagreement over whether primates are descended from Plesiadapiformes or just related do them.
Of course you don't want to connect an insecure computer to the Internet today without some kind of firewalling, and most computers aren't very secure, even if they're not running Windows.
NAT is a cheapass way to build a slightly-stateful firewall. I'm not exactly sure how you fit your NAT routers between your notebook and the nearby wireless pod (:-), but if you're only doing pure-client stuff or are willing to tweak your NAT box you can make it mostly work most of the time for most applications.
That doesn't mean that NAT isn't evil. It breaks the end-to-end paradigm that makes it easy to develop new applications for the Internet, and forces most people to just be clients unless they're running software which does various levels of ugliness to work around NAT. It's easy to make a client that works behind NAT to reach a server that's not, and a bit harder to let non-NATted clients reach NATted servers, but it's a lot harder if both ends run NAT. For instance, do you know why Skype is so popular, in spite of being a closed-source closed-documentation proprietary application that doesn't use either of the common VOIP protocol standards, doesn't interoperate with anything, runs Repeckt-Mah-Obscuritay unverifiable crypto, and turns random well-connected users into supernodes? It's partly because it was well-done and shiny, but it's largely because it does an effective job of NAT traversal, and the supernode business is one of the tricks it uses to do that.
Think about your options for firewalling in an IPv6 environment. You can still build firewalls that let in stuff you want and don't let in stuff you don't want, and even do it statefully so you only let in good stuff when you're interested in listening for it, and with 2**64 IPv6 addresses for your house, you could even just leave stuff unprotected and nobody'll find it by port-scanning (in practice you wouldn't do that, because if you reveal your IP address on one protocol, such as by initiating http to a web site, then a miscreant knows to scan other ports on that address, or he could go scanning MAC addresses for recent Dell PC models or whatever. But you *could*, and you could even set up your machine to have different protocols use different IPv6 addresses.
I agree that witness tampering is a crime - I just don't know enough of the actual facts (as opposed to the reported-in-the-press facts or the allegations by various cops) to know if what he did was close enough to the legal definitions of the crime; if they haven't charged him for it yet, maybe they've decided it wasn't close enough, or they don't have enough evidence to be likely to get a conviction, or they're talking their time.
It does sound like the facts are at least sufficiently close to witness tampering to win him yet another legal-system Darwin Award, regardless of whether he gets separately prosecuted or convicted for it.
As another poster said, what you need for those people is an 80-90% solution; if they want a 100% solution they can get their PC to run IPv6 and tunnel if they can't go native.
You only need to talk to another server on the net if there's some application it's serving that you want. These days, usually that's the web or email or maybe some IM protocol, so a box that proxies a couple of popular services will take care of connecting your lameoid PC to most new and interesting IPv6-only servers.
For now, the more entertaining problems are when there's a server out there with IPv4 AND ipv6, and their DNS advertises both, and your PC decides to connect using IPv6, but you don't actually have IPv6 connectivity from your ISP. Oops. It's probably Bill's fault.
You're counting the addresses from the wrong end - a "Class C" is equivalent to a/24, not a/8. I do know a few businesses that have trouble fitting in a/8, but most of them are really bad at managing address space (:-) and the rest are consumer ISPs. It's actually quite easy to switch from a/24 to a/22 if you actually need to, but you're still getting Provider Allocated IP address space; if you want to get your *own* space that's globally routable, you usually need enough machines for a/20 or so.
In IPv6-land, the typical allocation for an end-user organization is/48 - that's enough for you to have 2**16 buildings each of which has 2**16 LANs with hosts that have 48-bit MAC addresses, but in practice that's not usually how your space needs to be split up.
Most businesses have four kinds of IP applications
Clients at the business's office, which need reliable outbound connectivity,
Roaming clients, which aren't part of the problem here,
Servers that need reliable inbound connectivity, and
Servers that need to be reliable and need to be at the business's office.
The latter group is really small, and for most companies is limited to cranky VPN appliances. You *can* run customer-facing servers at your office, but it's become much more common to run them in colo centers, partly for reliability and partly for cost reasons, and for those cases it's usually just fine to use the colo provider's IP address space instead of PI space.
The economics change around from year to year, but colo usually wins for a lot of applications except for a small number of very large companies or for people who don't need dual-homing levels of uptime. Otherwise the number of global routes would be a lot higher. (Last time I looked it was around 200,000 - a few years ago the Imminent Death of the Net was predicted as the number approached 100K, but bigger routers with more memory have come out since then.)
IPv4 PI space is seriously non-scalable, and you can't simply duplicate it in IPv6. Tried to buy any Class-C swamp space lately? One thing that has slowed the explosive growth of demand for IPv4 PI for multihomed customers is the lack of IPv4 space (and RIR address-conservation policies), and IPv6 will "fix" that.
Another is that fortunately many of the businesses that would want multi-homing for servers are putting them in colo space rather than on their premises, so they're ok with using provider-allocated space, and it's only the colo provider that has to advertise multiple routes. Another is the policy issue that ARIN will normally not sell you PI space smaller than some size (is it/21 these days?), while NAT and firewalls mean that most businesses don't need much more than a/28 per site.
Shim6 is supposed to fix this problem, but IMHO it's an ugly ugly hack that won't succeed.
The other popular reason for getting PI space is to make it easier to renumber if you change ISPs. Unlike multihoming, this is a problem that can be made to go away by fiat. It made more sense back in the 1980s, before DHCP and DNS support became relatively universal. Renumbering servers and VPN tunnel appliances is still a bit annoying, but usually not bad, and you don't really need to renumber client machines any more, you just expire their DHCP leases if they're non-laptops, or unplug their LAN connections if they are. (Yeah, I know, it's not really quite that simple, but it's still fixable, especially because the parts that are hardest to fix are usually behind firewalls or NAT so you don't care.)
Hardware's cheap - staff time is expensive, and consultant time is more expensive.
Over a 3 year timeframe, most of your legacy servers will look pretty old, slow, expensive, and financially-depreciated next to the new shiny servers people are buying to replace them. Most of the exceptions are on private networks anyway, not the public internet, like that mainframe your corporate HR department uses. So even if your public internet connectivity is all IPv6, you can still tunnel 10.0.0.0/8 or 192.168.0.0/24 through it and nobody'll mind.
Also, most of those IPv4 applications can handle 6to4 translation if they're doing all the work on the server and not running fat clients, so you can put up a PC farm running translators.
Routers are more of a problem - handling IPv4 bits is actually the primary job they're doing, as opposed to being a communication mechanism for getting to some database application that's the primary job. They're often implementing stuff in ASICs, and they're harder to replace cost-effectively.
One of the things ICANN did for a while was insist that IPv6 addresses only be handed out in big chunks, and priced them high enough (e.g. $2500 for a/48) that you wouldn't get them unless you actually had some use for them. I'm not convinced that it was a *good* strategy, but it did prevent the growth of an IPv6 swamp resembling the IPv4 portable class-C swamp space.
If you want the space at that price, and can document that you need it, you can get it; otherwise you can get IPv6 service from an ISP or IPv6 tunnel broker and have them assign you space.
Slashdot Mirror
The other came out on Usenet in the 80s and went something like
Some of my coworkers got it, but a couple of them didn't. The disturbing part was that they recognized the Cobol program, but were too young to recognize Rocky and Bullwinkle...
When Karl says he's going to spend time with his family, it's mainly his half-brother Cthulhu and his cousin Hastur. Just be careful not to say his name three times or he'll be back....
But for the most part, tracking ethnicity is a spectacularly bad idea, and tracking religion is not only worse but also much less reliable because it's not a constant. Ignore the questions of whether your mom thinks you're still a good Catholic and can look up whether you've shown up at Mass lately or see that you visited the Zendo down the street or guess about that party on Solstice (or, ahem, May Day) where most of your friends are pagans... (and at least the database won't let your mom see whether you're gay, probably because in China that's Just Not Talked About.)
Ethnic cleansing is a lot easier if you've got a database saying where all the members of Ethnic Group X live in your town*. And if you don't want to hire Jews, you've got a database that says who they are. And it's much easier to get the no-fly list right when you can tell if somebody named Malik Muhammed is African-American (ok, he's just one of Those People, and make sure to give him the kosher airline meal) or Arab (he goes on the list.) I'm sure China has their equivalent issues about which ethnicities get privileges and which don't, even outside of special provinces like Tibet.
And at least it's just China - in most of the Moslem countries, if you weren't born a Moslem, that's usually ok, though you might have to pay a tax, but if you were born a Moslem and you've converted to other religions, Sharia says you have to either convert back or die, and even in the more moderate countries like Egypt, they'll throw ex-Moslems in jail for preaching the wrong religion, as happened to friends-of-friends of mine back in the 80s. In some countries they've got Sunni-vs-Shia issues that are better off not having database support, and I don't know what happens if you convert to another branch of Islam such as Sufiism. And the Baha'i also seem to be a special always-infidel case.
* I get to mention the ethnic-cleansing-rounding-up-Ethnic-Group-X example without triggering Godwin's Law this time - one of the people at the party I was at last night was talking about how he didn't know two of his grandparents because they didn't make it back from the Japanese-American internment camps during the war (he wasn't doing a political rant - he was dealing with his aging mother's house, where there's still stuff of his other grandmother's as well.) One of the joys of living in California is the wide variety of people you get to be with - most of the examples I gave above are for people I've seen recently, though a few were people I hadn't seen in a while and one or two were from recent news.
* And the one person I've known who *was* a potential suicide bomber is presumably not on the no-fly list; he was a college student in Japan during the war, and was considering volunteering to be a kamikaze pilot, but one of his professors talked him out of it. I worked for him about 30 years ago - his war history didn't stop him from having a US security clearance. (Do I still get didn't-trigger-Godwin's-Law credit if I mention the tattoo his boss had on his arm? Probably not. )
And yes, the rate we're using up resources and growing population is making it harder to sustain the energy needed for space travel - that's one of the things we need to figure out how to fix.
The two activities overlap significantly - a critical skill we need to learn for surviving in space is how to run a viable ecosystem, whether it's on a closed-system spaceship or a terraformed planet. So far we've only run a few small closed-system terrarium experiments like the Biosphere (which had to cheat and bring in extra oxygen, something that's only easy to do when you're on a working planet) - even non-closed-system spacecraft like the Space Station have been getting weird mold problems we don't know how to manage well. And we've got one experiment running on terraforming a planet (Earth) which is going pretty badly at the present time - we don't even know how the thermostat works yet. So we're going to need to learn to fix planets before we can get off this one, and the best way to learn that is by trying to fix this planet.
Also, the energy requirements for getting lots of people off the planet are amazingly high; we're decades away from building even space elevators, much less mass-production rockets, and since we don't know how to run portable ecosystems yet, it doesn't make sense to give high priority to the transport parts; we can let Moore's Law crank for another century or two just fine.
There are one or maybe two exceptions to that - satellites studying and observing the Earth are really useful in learning how to fix the planet, and we can launch those with our current low technology. Unlike other parts of the space program, which have given us powdered orange drink and better military missiles by diverting scientists and engineers from making better commercial aircraft or more efficient automobiles, the satellite part of the space program may have been a big win. Also, power satellites *might* be useful as an alternative to carbon-fuel or nuclear energy, and it might make sense to work on them early, but that'll take a lot of earth-based design to show whether it might be feasible.
If the information that was allegedly possibly leaked had been enough for somebody to actually prosecute some Executive Branch people (whether FBI or Pentagon or NSA or whatever) and they'd gotten convicted already, *then* this kind of raid might count as "obstruction of justice", but they're acting sufficiently proactively that they'll at least get away with it until the Bush Administration is out of office. And probably after that as well.
------
* If they don't have warrants or permission, well, the Bush Administration thinks it's legal anyway, and they've got a Justice Department who wouldn't prosecute them for doing it and they've stacked the Federal courts as well.
and getting states, banks, credit companies, airlines, etc. to do massive data collection. And it's not like it started with them - the FBI wiretap enthusiasts like Louis Freeh, the NSA anti-public-crypto people, the Echelon project, etc. all date to the Clinton or GHWBush/Reagan administrations or earlier.
It's going to take a *long* time to tear down that stuff and turn this back into America again, and most of that won't happen unless we replace the current Executive Branch with one that's actually committed to doing it. Most of the major candidates aren't talking like that - certainly Hillary and Rudy and John Edwards and McCain and Romney don't have a history of wanting to do that, and you're pretty much down to Dennis Kucinich and Ron Paul before you'd get to anybody who'd talk about that kind of concept as a campaign strategy. Perhaps if the Democrats not only win the White House but also increase their control of the Senate and House they'll have some willingness to do that after a couple of years.
For now, though, Homeland Security Anonymous Spokescritters report that Enhanced Terrorist Surveillance Program has been reporting increased frequency of terrorist chatter saying "Booga Booga", so if you're even suggesting that we decrease wiretapping then you're a threat to national security and our precious bodily fluids.
What? With your bare hands? ...
Neither had I, and I'm afraid that's about all I know about them :-)
The problem isn't just raw addresses - it's easy to duplicate that. The problem is the scalability of the routing system itself - how many routes does an internet backbone router have to keep track of? It's gone from 100K to 200K in the last couple of years, and demand is growing.
IPv6 was supposed to fix this problem by getting everybody to use address space in a hierarchical fashion, though I'm afraid it was a combination of wishful thinking and vigorous hand-waving rather than reflecting reality. IPv6 does have enough space to hack around it a bit, e.g. ISP#1 and ISP#2 could get a /32 block that they use to assign all of the customers who are homed to both of them, but that takes N**2 of those blocks to support N ISPs, which isn't too bad if you just support (say) a cabal of US Tier1 ISPs (about 25 of them => 625 blocks), but there are 5-10000 smaller ISPs in the US (you could force the small ISPs to get dual-home space from a Tier 1 or something, but that would take really entertaining politics to make that happen.)
There have also been proposals to manage IPv6 address space geographically, which would have similar scaling wins (e.g. split the US into 1000 pieces by area code or 50 states or whatever), but it confuses the routing and peering structures because not all ISPs connect in all regions (probably works better in Europe where exchanges handle more of the traffic.)
> The big problem with renumbering for me is the idiots with their idiot firewalls
Yup. That matches my experience, though for me its usually the VPNs (separate or integrated into the firewall) rather than the blocking rules themselves.
DNS complicates things a bit, but if you give yourself more than a week for planning you can set your DNS timers appropriately. Most hosts and routers can do just fine with secondary IP addresses, so you can add the secondary addresses to the machines a while before telling the users to use them, and most other addresses can be managed by DHCP.
The price differences are even more egregious for Flash (though the prices are a lot lower, e.g. $700 for a 512MB upgrade vs. $10 commercial), because the stuff is basically only used for storing the OS and booting the machine - good camera memory is fine, and you're not constantly overwriting it so you're not going to burn through write cycles.
I mostly use my work laptop, and I set up the home desktop; my wife's laptops from her old consulting business are obsolete, so she uses the desktop these days. In theory I could boot it in Linux, but in practice it runs XP with three users, me, her, and root, in theory with only root having admin privileges. In practice, she runs the important Windows application (Turbotax) and installs interesting software and has admin privileges, even though she's mostly browsing, while I run it occasionally for iTunes (which unfortunately had to be installed by root, and gets confused about what's mine and what's root's.)
Back to reality, though, it's amazing how many Unix passwords were "abc123", back when our systems required at least six characters including some non-letters
Last time I displayed computer output on a TV was around 1995. It wasn't intentional - I was visiting my parents, and the signal from my laptop showed up on their TV. It wasn't quite in sync - there were about three copies of part of the text, scrolling slowly vertically. But it was semi-readable, and was definitely enough to answer the discussions about "so can you avoid TEMPEST eavesdropping by using a laptop?" I suspect the leakage was coming from the CRT port on the back of the PC, as opposed to from the LCD monitor electronics.
We and our fellow apes are related to the other primates; Wikipedia says that there's some disagreement over whether primates are descended from Plesiadapiformes or just related do them.
NAT is a cheapass way to build a slightly-stateful firewall. I'm not exactly sure how you fit your NAT routers between your notebook and the nearby wireless pod (:-), but if you're only doing pure-client stuff or are willing to tweak your NAT box you can make it mostly work most of the time for most applications.
That doesn't mean that NAT isn't evil. It breaks the end-to-end paradigm that makes it easy to develop new applications for the Internet, and forces most people to just be clients unless they're running software which does various levels of ugliness to work around NAT. It's easy to make a client that works behind NAT to reach a server that's not, and a bit harder to let non-NATted clients reach NATted servers, but it's a lot harder if both ends run NAT. For instance, do you know why Skype is so popular, in spite of being a closed-source closed-documentation proprietary application that doesn't use either of the common VOIP protocol standards, doesn't interoperate with anything, runs Repeckt-Mah-Obscuritay unverifiable crypto, and turns random well-connected users into supernodes? It's partly because it was well-done and shiny, but it's largely because it does an effective job of NAT traversal, and the supernode business is one of the tricks it uses to do that.
Think about your options for firewalling in an IPv6 environment. You can still build firewalls that let in stuff you want and don't let in stuff you don't want, and even do it statefully so you only let in good stuff when you're interested in listening for it, and with 2**64 IPv6 addresses for your house, you could even just leave stuff unprotected and nobody'll find it by port-scanning (in practice you wouldn't do that, because if you reveal your IP address on one protocol, such as by initiating http to a web site, then a miscreant knows to scan other ports on that address, or he could go scanning MAC addresses for recent Dell PC models or whatever. But you *could*, and you could even set up your machine to have different protocols use different IPv6 addresses.
It does sound like the facts are at least sufficiently close to witness tampering to win him yet another legal-system Darwin Award, regardless of whether he gets separately prosecuted or convicted for it.
You only need to talk to another server on the net if there's some application it's serving that you want. These days, usually that's the web or email or maybe some IM protocol, so a box that proxies a couple of popular services will take care of connecting your lameoid PC to most new and interesting IPv6-only servers.
For now, the more entertaining problems are when there's a server out there with IPv4 AND ipv6, and their DNS advertises both, and your PC decides to connect using IPv6, but you don't actually have IPv6 connectivity from your ISP. Oops. It's probably Bill's fault.
In IPv6-land, the typical allocation for an end-user organization is
- Clients at the business's office, which need reliable outbound connectivity,
- Roaming clients, which aren't part of the problem here,
- Servers that need reliable inbound connectivity, and
- Servers that need to be reliable and need to be at the business's office.
The latter group is really small, and for most companies is limited to cranky VPN appliances. You *can* run customer-facing servers at your office, but it's become much more common to run them in colo centers, partly for reliability and partly for cost reasons, and for those cases it's usually just fine to use the colo provider's IP address space instead of PI space.The economics change around from year to year, but colo usually wins for a lot of applications except for a small number of very large companies or for people who don't need dual-homing levels of uptime. Otherwise the number of global routes would be a lot higher. (Last time I looked it was around 200,000 - a few years ago the Imminent Death of the Net was predicted as the number approached 100K, but bigger routers with more memory have come out since then.)
Another is that fortunately many of the businesses that would want multi-homing for servers are putting them in colo space rather than on their premises, so they're ok with using provider-allocated space, and it's only the colo provider that has to advertise multiple routes. Another is the policy issue that ARIN will normally not sell you PI space smaller than some size (is it
Shim6 is supposed to fix this problem, but IMHO it's an ugly ugly hack that won't succeed.
The other popular reason for getting PI space is to make it easier to renumber if you change ISPs. Unlike multihoming, this is a problem that can be made to go away by fiat. It made more sense back in the 1980s, before DHCP and DNS support became relatively universal. Renumbering servers and VPN tunnel appliances is still a bit annoying, but usually not bad, and you don't really need to renumber client machines any more, you just expire their DHCP leases if they're non-laptops, or unplug their LAN connections if they are. (Yeah, I know, it's not really quite that simple, but it's still fixable, especially because the parts that are hardest to fix are usually behind firewalls or NAT so you don't care.)
Over a 3 year timeframe, most of your legacy servers will look pretty old, slow, expensive, and financially-depreciated next to the new shiny servers people are buying to replace them. Most of the exceptions are on private networks anyway, not the public internet, like that mainframe your corporate HR department uses. So even if your public internet connectivity is all IPv6, you can still tunnel 10.0.0.0/8 or 192.168.0.0/24 through it and nobody'll mind.
Also, most of those IPv4 applications can handle 6to4 translation if they're doing all the work on the server and not running fat clients, so you can put up a PC farm running translators.
Routers are more of a problem - handling IPv4 bits is actually the primary job they're doing, as opposed to being a communication mechanism for getting to some database application that's the primary job. They're often implementing stuff in ASICs, and they're harder to replace cost-effectively.
If you want the space at that price, and can document that you need it, you can get it; otherwise you can get IPv6 service from an ISP or IPv6 tunnel broker and have them assign you space.