If you pirate it, you are open for prosecution. It doesn't matter if you also have a licensed copy, you still committed copyright violation. If you're OK with that risk, and the BSA don't seem too trigger happy against consumers right now, then go ahead and pirate it, but be aware that risk always exists.
If you buy it, you are supporting UbiSoft. You are supporting their game development team, which may be good, but also the boneheads who selected this DRM technology. They will only be reporting on sales to their managers, and if they can spin a story that their decisions, including the DRM, resulted in higher sales, they'll get a pat on the back and a "jolly good, carry on." If you want to do all of that, go ahead and buy it, but be aware that you can't support just the game devs and not the boneheads.
The ISP industry of Australia have already taken action to block the law. There's been technical deconstruction of it. There's been alternatives proposed that could be more effective at protecting children from online materials. There's even one member of Stephen Conroy's own political party advocating at least changing the legislation to an optional filter.
When asked, "do you think the gubbimunts should make the intartubs safe for kiddiewinks?" most of the Australian population agree. When asked more specifically about the details of the proposed plan, that support sharply declines: less than half think it's a good idea for the government to maintain a list of Naughty Pages. Only a tiny handful of lunatics think that list should be a secret.
So, we have a situation here in which a single Senator is pushing a legislation that wouldn't pass referendum as it stands, isn't supported by the industry, isn't supported by the public, and isn't even well supported by his own political party. It's a poor approach to solving a problem that, for a start, won't actually solve the problem, but worse, will create new problems of its own, increasing the cost of Internet access in Australia while simultaneously decreasing the quality of service.
And yet it's almost certain to become law.
I think I will now go and opiate myself with sugar and caffeine.
Clouds cool the planet by the exact same amount as it's being warmed?
How extraordinary. Almost... implausibly extraordinary, one might say.
Also, quibble time, warming produces water vapour in the air, water vapour absorbs and re-emits infra-red radiation thus further contributing to warming, and water vapour also forms clouds that reflect infra-red radiation, thus mitigating some warming effects. The uncertainty in how much water vapour contributes to warming, and how much to the prevention of warming, is why the IPCC report has such a large range of possible temperature changes. I'm pretty sure that *is* science.
Hover-based JS menus are a usability nightmare on the iPhone. On my veterinarian's website, to check their opening hours I have to click "About" and then very quickly click the "Hours" link before the site loads the basic "About" page and re-collapses the menu.
On the flip side, though, hover-based JS menus are a usability nightmare on a desktop with a pointing device. The fuckers disappear if you get a pixel too far over to one side, or appear when you didn't want them to because you shoved the cursor off the text you were trying to read.
So... I guess I have no point? Oh! I have one now: we need to accept that hovering is specific to one form of input only, and that form of input is no longer the only kid on the block.
An insecure CPE device is never the end user's fault.
Device manufacturers have no reason to be shipping devices that, by default, expose any services to the external network. Bind to the internal network, and the scope of attack is reduced to people in wi-fi range; that sort of attack doesn't scale up to be significantly damaging.
Network providers have no reason to be installing custom images that allow them remote access but also use the OEM's default password. Use a single nonce for all your devices if you don't have the capability to store a nonce for each customer, and at least the scope of attack is reduced to the service providers that the malware author can get default passwords for, instead of the much smaller set of default passwords each device provider uses.
Device and service providers who are so close to going bust that the minute cost of these steps would push them over the edge are, in two words, already fucked.
If you consider the motivation for an SSL root to issue a certificate, I don't think you could really consider any of them trustable.
I used to remove CAs from my trust chain when a news story broke about them giving certificates to fraudsters, but two things stopped me: (1) the CAs have us by the balls in that removing any one CA "breaks" a significant part of the web; and (2) browsers and OSes re-install the CAs with the next update.
I remain of the opinion that x.509 has failed on the Web; https provides relatively weak end-to-end encryption, but even weaker authentication.
Most voters are unaware of this issue. For most of that majority, awareness of this issue takes a very distant back seat stacked against working conditions, taxation, health care, infrastructure, education, and a whole long list of other matters. For most of that majority of that majority, all issues take a very distant back seat to which party your father voted for.
For the few people who care about this issue enough for it to affect their vote, they have two choices: Liberal or Labour. Neither party is showing any interest in listening to industry or community, so it doesn't seem to make a difference.
The Australian election system is set up such that voting for a third party really is throwing your vote away. If your preferred candidate is not successful on the first round, that vote is discarded and your second preference is instilled. So in effect, you vote for either Liberal or Labour, and any other marks you put on the paper are just chicken scratches.
Sometimes I draw on an extra box and vote for Optimus Prime. No luck so far.
I nearly agree with you, but pituitary and hypothalamic disorders are not common and have not trebled in the past thirty years like obesity has, and they make weight loss more difficult, not impossible.
It seems pretty clear the author has just put together a few hot-topic items (all of which have known answers) spread out over a six-page article to generate impressions. EEesh. Checking...yep, kdawson article. Ugh, my bad.
Fair enough, it seems to me that you're saying it can't perform many (most?) of the same functions as well as a desktop or laptop computer can (aka, a modern personal computer). I can't disagree with that:-) I was just being an anal pedant about the terminology, anyway.
Producing into a dead-end is not a boon. Consumers should be using the consumed product to provide energy, knowledge or materials for more refined production. Entertainment can fall into either category: entertainment to refresh and relax is good, entertainment to pass the time before death is not. (Though frankly, I'm OK with people choosing to reap the benefits of living in an advanced society and going down the hedonism route!)
But in terms of society, the hacker bent over his keyboard is wasting his time if he is only working to feed the desires of the couch potato. At best, that's a means to sustain some other form of productivity that feeds back into a loop of production. At worst, it's a means of funding his own hedonism:-)
One. As a student, it's light, portable, relatively inexpensive, and has deals with textbook publishers. Silent virtual keyboard for lecture note-taking. Games for boring as hell lectures. If I were still at University, I'd be salivating.
Two. As a frequent traveller, weight and size reign supreme. A slim, lightweight device with good battery life and enough power to stay on top of email, touch up presentations, take notes, and manage a schedule could easily replace my laptop -- which is already a Macbook Air, having sacrificed performance and an optical drive for weight. Add a good eBook facility (judgement reserved, the iBookstore sounds overpriced, but Stanza has won my heart on the iPhone) and capacity for movies, and the flights suck a little less.
Three. With low technical literacy, relatively poor eyesight, and painful but not crippling arthritis, my mother is unsuited to the tiny screen, tinier keyboards, and very poorly adapted operating systems that we know and love with netbooks. A 3G enabled device which can manage photos, handle email, and browse enough of the web to make a booking at the next destination seems almost purpose built for her when she starts her obligatory post-retirement around-the-country travelling.
Whether Apple aimed deliberately at any of those situations is another question -- I suspect they aimed at a new market, the couch-based web browsing experience. In my household, the iPhone has overtaken the laptop as most frequently used browsing device, not because it's better at browsing the web (which obviously it isn't) but because it's a hell of a lot more convenient. I think that's what all the fuss about "full page" browsing is from. But I don't see that as a saleable use case for me -- I'm too cheap to buy a home browsing device. But I'd use it for that, if I got one for travel.
I do not think you understand what a general purpose computer is.
It is not a personal computer, though a personal computer is also a general purpose computer.
It is not a minimum set of performance characteristics, eg, able to play Crysis.
It is not a particular combination of peripherals, eg, keyboard, mouse, or display.
A general purpose computer is a device which can perform general computation, ie, one that is not tied down to a specific task (or set of tasks). It has input and output, memory, arithmetic/logic and control units. What is connected to the input and output is, more or less, irrelevant. A device using punch-cards and line-printers, but able to perform any computation that you specify, is a general purpose computer. That's what computers were, before some smart cookies invented displays and keyboards and mice and touch interfaces and voice control and the F-u-F-me.
The iPad is, in fact, a general purpose computer. It's got all the necessary bits: it has a CPU (ALU & Control Unit combined), it has a multitude of I/O channels (touch screen, 30-pin port, WiFi, optional 3G, and Bluetooth), and it has memory. It can be used for general purpose computation; in fact, if Numbers isn't deliberately crippled, it will even come with a $9.99 Turing-complete language. Not one you'd want to use for anything significant, mind you. Or, for a low $99 a year, you can use any number of other languages to program the device to suit your purposes.
On the flip side, it does have DRM. You can't distribute your binary software, save with Apple's approval (though you can distribute the *code* and others can pay a $99/year fee for the right to install software directly onto their device!) You can only use the approved API (plus a little more, if you don't need Apple's approval of your software), since the pre-installed operating system limits access to various system calls or direct to the hardware. You can't replace the operating system with one that's a little less... authoritarian, either, since the hardware has been taught to trust only Apple.
I really don't like the DRM - but I can compromise on the DRM if the platform offers me enough. I buy DVDs, after all, and those are lumbered with DRM too.
The root servers (and other authoritative servers) use BGP anycast partly to distribute DNS query load, and partly to provide faster response time. Nice, "simple" problem.
Some authoritative servers, typically CDN servers, give different answers to the same DNS question depending on the source IP address. There's a few advantages to doing it in DNS: first, BGP anycast is susceptible to route changes disrupting a TCP session, particularly if there's equal-weight choices to exit a router and least-saturation is used to select a best path. DNS routing "configures" the client at look-up time. At the very least any given single TCP session will always go to the same host. Next up, BGP anycast requires that you use an entire minimum routable block (currently a/24 in v4 and a/48 in v6) in which you only use a single IP address. It's very wasteful. In order for BGP anycast to work, you need to announce a routable block in multiple locations, and you need to be able to withdraw the route announcement if a location becomes unable to serve requests. If you host more than one service in that block, all services at a location must be withdrawn at the same time, which largely defeats the purpose. DNS routing allows you to use single IPs within blocks that other services also inhabit, since withdrawal of a service consists of no longer giving that answer. BGP anycast is good at increasing the availability of a service, since a route can be withdrawn in seconds, but DNS routing depends on cache expiry, and is better suited to load balancing. BGP anycast, on the other hand, is lousy at load balancing: you have to be very careful in your placement of anycast nodes, and constantly monitor and update locations and announcements, to ensure a reasonably even spread of load. DNS routing can do active and automatic load balancing, changing responses to preference less loaded servers.
The Google draft is intended to improve the ability of DNS routing to cope with people doing stupid things like running a centralised DNS resolver whose IP address doesn't have a good correlation with the network location of the originating request. Like, say, Google does.
Google wants Google Public DNS to not suck when doing Akamai requests, that's all. No gains for anyone else, just increased query load and cache entries.
I think that part of figuring out how to do interplanetary travel cheaply is figuring out how to do it expensively.
I also think that the insanely wealthy would, on the whole, prefer to reap the benefits of *other people* taking the incredibly high risks associated with space flight, let alone trying to colonise a location more inimical to human life than even the most barren parts of the Earth.
I don't see rich folk lining up to move to Antarctic research stations...
... IPv4 machines should simply be a.b.c.d.0.0.0.0.0.0 or something equally obvious...
::ffff:1.2.3.4. Not that it helps, since v6 and v4 stacks are different.
IPv6 is still network portion, host portion. You could still specify things in mask notation, if you wanted to, but it's kind of silly. Just use network prefix length notation, it's nicer for both v4 and v6. Gateways are still usually on::1.
Routers and IP stacks could be written to extend the address space a few more bits
Ah yes, the "use more v4 bits" idea that comes up every time. Let's look at what you'd need to do to extend IPv4 addresses by one bit. First, you need somewhere to store the bit. You could use a reserved bit, or you could make a new IP option. Either way you've hit your first roadblock: no existing IPv4 equipment or software will be expecting this, so you need to replace everything with IPv4.1 equipment -- that, or randomly your packets won't go to the right destination, they'll go to the 0-bit destination instead. Oops.
You wave a magic wand and solve that problem (which is the same problem as the IPv6 support problem). Now you turn to DNS. Oops, an A record only contains 32 bits. You'll need some way for a DNS resolver to report the extra bit back, but you can't break compatibility with existing resolvers, so you will probably wind up defining a new record, let's call it the AA record. Now you can map names to IPv4.1 addresses -- but you need to roll out DNS software everywhere to support it. Oops.
Another magic wand later, you come to the application layer. It turns out that a bunch of software has a bunch of struct sockaddr_in variables that it uses to connect to services and to figure out who connected back in turn. You'll need some way to deal with that. Maybe you could define a new structure, sockaddr_in4_1 or something, that has the extra bit of information. Oh, but shit, now you need to rewrite all your application software to be aware of that new structure.
Then you try to figure out DHCP, PPP, reverse DNS, ICMP, BGP, spanning-tree, accounting systems, DOCSIS, and every other IP network protocol known to man, because every single one of them is built on the basis that there's only 32 bits in a network address.
And eventually, it turns out that the people who came up with IPv6 didn't all somehow miss the blindingly obvious solution, because there is no blindingly obvious solution.
IP-over-DNS, or IP-over-ICMP. Your router shouldn't be blocking port 53 *to your ISP's resolver*. Your router probably won't block ICMP. If you run a VLAN, you'll still want to do DNS, and at that point IP-over-DNS can provide a tunnel for an attacker into your "protected" network.
Virus kits mean that complex tricks become commonplace. It's common for a virus to go stealth now, because there's library code to root a machine. It's common for viruses to use a suite of anti-anti-virus tools, because there's library code for that. I doubt anyone's doing IP-over-DNS or other advanced external firewall circumvention techniques, but if anti-virus vendors ever get it into their heads that they could sell black boxes to sit between PC and intarwebs, offering SPI firewall scanning, it will happen.
Come to think of it, a black box virus firewall would be a product worth buying, if done even slightly right. It could alert on high outgoing mail volume, it could fetch a daily list of evil IPs to block and alert on, it could run a self-checking firmware OS to protect itself against tampering, and it could inspect packets for known virus message signatures. And it wouldn't bring your PC to its knees!
Then again, the half-assed work that anti-virus vendors do would probably drop packets all over the floor if your network speed exceeded 56.6k/sec.
Slashdot Mirror
I guess it was a bone-head move, because I meant "made an unauthorized copy". :-)
That's the worst of both worlds.
If you pirate it, you are open for prosecution. It doesn't matter if you also have a licensed copy, you still committed copyright violation. If you're OK with that risk, and the BSA don't seem too trigger happy against consumers right now, then go ahead and pirate it, but be aware that risk always exists.
If you buy it, you are supporting UbiSoft. You are supporting their game development team, which may be good, but also the boneheads who selected this DRM technology. They will only be reporting on sales to their managers, and if they can spin a story that their decisions, including the DRM, resulted in higher sales, they'll get a pat on the back and a "jolly good, carry on." If you want to do all of that, go ahead and buy it, but be aware that you can't support just the game devs and not the boneheads.
Sorry. I think the general message doesn't change though, even if I'm clearly NOT a climatologist :-)
The ISP industry of Australia have already taken action to block the law. There's been technical deconstruction of it. There's been alternatives proposed that could be more effective at protecting children from online materials. There's even one member of Stephen Conroy's own political party advocating at least changing the legislation to an optional filter.
When asked, "do you think the gubbimunts should make the intartubs safe for kiddiewinks?" most of the Australian population agree. When asked more specifically about the details of the proposed plan, that support sharply declines: less than half think it's a good idea for the government to maintain a list of Naughty Pages. Only a tiny handful of lunatics think that list should be a secret.
So, we have a situation here in which a single Senator is pushing a legislation that wouldn't pass referendum as it stands, isn't supported by the industry, isn't supported by the public, and isn't even well supported by his own political party. It's a poor approach to solving a problem that, for a start, won't actually solve the problem, but worse, will create new problems of its own, increasing the cost of Internet access in Australia while simultaneously decreasing the quality of service.
And yet it's almost certain to become law.
I think I will now go and opiate myself with sugar and caffeine.
Yes.
But the black hats who did the testing are still waiting on a higher bid before they do a limited release of that information.
Clouds cool the planet by the exact same amount as it's being warmed?
How extraordinary. Almost... implausibly extraordinary, one might say.
Also, quibble time, warming produces water vapour in the air, water vapour absorbs and re-emits infra-red radiation thus further contributing to warming, and water vapour also forms clouds that reflect infra-red radiation, thus mitigating some warming effects. The uncertainty in how much water vapour contributes to warming, and how much to the prevention of warming, is why the IPCC report has such a large range of possible temperature changes. I'm pretty sure that *is* science.
I think I understand what you're suggesting: we should pave roads with the corpses of someone else's children, right?
Hover-based JS menus are a usability nightmare on the iPhone. On my veterinarian's website, to check their opening hours I have to click "About" and then very quickly click the "Hours" link before the site loads the basic "About" page and re-collapses the menu.
On the flip side, though, hover-based JS menus are a usability nightmare on a desktop with a pointing device. The fuckers disappear if you get a pixel too far over to one side, or appear when you didn't want them to because you shoved the cursor off the text you were trying to read.
So... I guess I have no point? Oh! I have one now: we need to accept that hovering is specific to one form of input only, and that form of input is no longer the only kid on the block.
The bot owns the router. If you rely on a setting in the router to prevent the owned router from causing harm, you've already failed.
Heh, software firewalls.
An insecure CPE device is never the end user's fault.
Device manufacturers have no reason to be shipping devices that, by default, expose any services to the external network. Bind to the internal network, and the scope of attack is reduced to people in wi-fi range; that sort of attack doesn't scale up to be significantly damaging.
Network providers have no reason to be installing custom images that allow them remote access but also use the OEM's default password. Use a single nonce for all your devices if you don't have the capability to store a nonce for each customer, and at least the scope of attack is reduced to the service providers that the malware author can get default passwords for, instead of the much smaller set of default passwords each device provider uses.
Device and service providers who are so close to going bust that the minute cost of these steps would push them over the edge are, in two words, already fucked.
If you consider the motivation for an SSL root to issue a certificate, I don't think you could really consider any of them trustable.
I used to remove CAs from my trust chain when a news story broke about them giving certificates to fraudsters, but two things stopped me: (1) the CAs have us by the balls in that removing any one CA "breaks" a significant part of the web; and (2) browsers and OSes re-install the CAs with the next update.
I remain of the opinion that x.509 has failed on the Web; https provides relatively weak end-to-end encryption, but even weaker authentication.
Most voters are unaware of this issue. For most of that majority, awareness of this issue takes a very distant back seat stacked against working conditions, taxation, health care, infrastructure, education, and a whole long list of other matters. For most of that majority of that majority, all issues take a very distant back seat to which party your father voted for.
For the few people who care about this issue enough for it to affect their vote, they have two choices: Liberal or Labour. Neither party is showing any interest in listening to industry or community, so it doesn't seem to make a difference.
The Australian election system is set up such that voting for a third party really is throwing your vote away. If your preferred candidate is not successful on the first round, that vote is discarded and your second preference is instilled. So in effect, you vote for either Liberal or Labour, and any other marks you put on the paper are just chicken scratches.
Sometimes I draw on an extra box and vote for Optimus Prime. No luck so far.
Every generation needs something they don't understand to fear and attempt to destroy. Games are the new rock and roll!
Nerd Rage is always funny.
I nearly agree with you, but pituitary and hypothalamic disorders are not common and have not trebled in the past thirty years like obesity has, and they make weight loss more difficult, not impossible.
It seems pretty clear the author has just put together a few hot-topic items (all of which have known answers) spread out over a six-page article to generate impressions. EEesh. Checking...yep, kdawson article. Ugh, my bad.
Fair enough, it seems to me that you're saying it can't perform many (most?) of the same functions as well as a desktop or laptop computer can (aka, a modern personal computer). I can't disagree with that :-) I was just being an anal pedant about the terminology, anyway.
Producing into a dead-end is not a boon. Consumers should be using the consumed product to provide energy, knowledge or materials for more refined production. Entertainment can fall into either category: entertainment to refresh and relax is good, entertainment to pass the time before death is not. (Though frankly, I'm OK with people choosing to reap the benefits of living in an advanced society and going down the hedonism route!)
But in terms of society, the hacker bent over his keyboard is wasting his time if he is only working to feed the desires of the couch potato. At best, that's a means to sustain some other form of productivity that feeds back into a loop of production. At worst, it's a means of funding his own hedonism :-)
Three use cases for it spring to my mind:
One. As a student, it's light, portable, relatively inexpensive, and has deals with textbook publishers. Silent virtual keyboard for lecture note-taking. Games for boring as hell lectures. If I were still at University, I'd be salivating.
Two. As a frequent traveller, weight and size reign supreme. A slim, lightweight device with good battery life and enough power to stay on top of email, touch up presentations, take notes, and manage a schedule could easily replace my laptop -- which is already a Macbook Air, having sacrificed performance and an optical drive for weight. Add a good eBook facility (judgement reserved, the iBookstore sounds overpriced, but Stanza has won my heart on the iPhone) and capacity for movies, and the flights suck a little less.
Three. With low technical literacy, relatively poor eyesight, and painful but not crippling arthritis, my mother is unsuited to the tiny screen, tinier keyboards, and very poorly adapted operating systems that we know and love with netbooks. A 3G enabled device which can manage photos, handle email, and browse enough of the web to make a booking at the next destination seems almost purpose built for her when she starts her obligatory post-retirement around-the-country travelling.
Whether Apple aimed deliberately at any of those situations is another question -- I suspect they aimed at a new market, the couch-based web browsing experience. In my household, the iPhone has overtaken the laptop as most frequently used browsing device, not because it's better at browsing the web (which obviously it isn't) but because it's a hell of a lot more convenient. I think that's what all the fuss about "full page" browsing is from. But I don't see that as a saleable use case for me -- I'm too cheap to buy a home browsing device. But I'd use it for that, if I got one for travel.
I do not think you understand what a general purpose computer is.
It is not a personal computer, though a personal computer is also a general purpose computer.
It is not a minimum set of performance characteristics, eg, able to play Crysis.
It is not a particular combination of peripherals, eg, keyboard, mouse, or display.
A general purpose computer is a device which can perform general computation, ie, one that is not tied down to a specific task (or set of tasks). It has input and output, memory, arithmetic/logic and control units. What is connected to the input and output is, more or less, irrelevant. A device using punch-cards and line-printers, but able to perform any computation that you specify, is a general purpose computer. That's what computers were, before some smart cookies invented displays and keyboards and mice and touch interfaces and voice control and the F-u-F-me.
The iPad is, in fact, a general purpose computer. It's got all the necessary bits: it has a CPU (ALU & Control Unit combined), it has a multitude of I/O channels (touch screen, 30-pin port, WiFi, optional 3G, and Bluetooth), and it has memory. It can be used for general purpose computation; in fact, if Numbers isn't deliberately crippled, it will even come with a $9.99 Turing-complete language. Not one you'd want to use for anything significant, mind you. Or, for a low $99 a year, you can use any number of other languages to program the device to suit your purposes.
On the flip side, it does have DRM. You can't distribute your binary software, save with Apple's approval (though you can distribute the *code* and others can pay a $99/year fee for the right to install software directly onto their device!) You can only use the approved API (plus a little more, if you don't need Apple's approval of your software), since the pre-installed operating system limits access to various system calls or direct to the hardware. You can't replace the operating system with one that's a little less ... authoritarian, either, since the hardware has been taught to trust only Apple.
I really don't like the DRM - but I can compromise on the DRM if the platform offers me enough. I buy DVDs, after all, and those are lumbered with DRM too.
Different problems to solve.
The root servers (and other authoritative servers) use BGP anycast partly to distribute DNS query load, and partly to provide faster response time. Nice, "simple" problem.
Some authoritative servers, typically CDN servers, give different answers to the same DNS question depending on the source IP address. There's a few advantages to doing it in DNS: first, BGP anycast is susceptible to route changes disrupting a TCP session, particularly if there's equal-weight choices to exit a router and least-saturation is used to select a best path. DNS routing "configures" the client at look-up time. At the very least any given single TCP session will always go to the same host. Next up, BGP anycast requires that you use an entire minimum routable block (currently a /24 in v4 and a /48 in v6) in which you only use a single IP address. It's very wasteful. In order for BGP anycast to work, you need to announce a routable block in multiple locations, and you need to be able to withdraw the route announcement if a location becomes unable to serve requests. If you host more than one service in that block, all services at a location must be withdrawn at the same time, which largely defeats the purpose. DNS routing allows you to use single IPs within blocks that other services also inhabit, since withdrawal of a service consists of no longer giving that answer. BGP anycast is good at increasing the availability of a service, since a route can be withdrawn in seconds, but DNS routing depends on cache expiry, and is better suited to load balancing. BGP anycast, on the other hand, is lousy at load balancing: you have to be very careful in your placement of anycast nodes, and constantly monitor and update locations and announcements, to ensure a reasonably even spread of load. DNS routing can do active and automatic load balancing, changing responses to preference less loaded servers.
The Google draft is intended to improve the ability of DNS routing to cope with people doing stupid things like running a centralised DNS resolver whose IP address doesn't have a good correlation with the network location of the originating request. Like, say, Google does.
Google wants Google Public DNS to not suck when doing Akamai requests, that's all. No gains for anyone else, just increased query load and cache entries.
I think that part of figuring out how to do interplanetary travel cheaply is figuring out how to do it expensively.
I also think that the insanely wealthy would, on the whole, prefer to reap the benefits of *other people* taking the incredibly high risks associated with space flight, let alone trying to colonise a location more inimical to human life than even the most barren parts of the Earth.
I don't see rich folk lining up to move to Antarctic research stations...
::ffff:1.2.3.4. Not that it helps, since v6 and v4 stacks are different.
IPv6 is still network portion, host portion. You could still specify things in mask notation, if you wanted to, but it's kind of silly. Just use network prefix length notation, it's nicer for both v4 and v6. Gateways are still usually on ::1.
Ah yes, the "use more v4 bits" idea that comes up every time. Let's look at what you'd need to do to extend IPv4 addresses by one bit. First, you need somewhere to store the bit. You could use a reserved bit, or you could make a new IP option. Either way you've hit your first roadblock: no existing IPv4 equipment or software will be expecting this, so you need to replace everything with IPv4.1 equipment -- that, or randomly your packets won't go to the right destination, they'll go to the 0-bit destination instead. Oops.
You wave a magic wand and solve that problem (which is the same problem as the IPv6 support problem). Now you turn to DNS. Oops, an A record only contains 32 bits. You'll need some way for a DNS resolver to report the extra bit back, but you can't break compatibility with existing resolvers, so you will probably wind up defining a new record, let's call it the AA record. Now you can map names to IPv4.1 addresses -- but you need to roll out DNS software everywhere to support it. Oops.
Another magic wand later, you come to the application layer. It turns out that a bunch of software has a bunch of struct sockaddr_in variables that it uses to connect to services and to figure out who connected back in turn. You'll need some way to deal with that. Maybe you could define a new structure, sockaddr_in4_1 or something, that has the extra bit of information. Oh, but shit, now you need to rewrite all your application software to be aware of that new structure.
Then you try to figure out DHCP, PPP, reverse DNS, ICMP, BGP, spanning-tree, accounting systems, DOCSIS, and every other IP network protocol known to man, because every single one of them is built on the basis that there's only 32 bits in a network address.
And eventually, it turns out that the people who came up with IPv6 didn't all somehow miss the blindingly obvious solution, because there is no blindingly obvious solution.
IP-over-DNS, or IP-over-ICMP. Your router shouldn't be blocking port 53 *to your ISP's resolver*. Your router probably won't block ICMP. If you run a VLAN, you'll still want to do DNS, and at that point IP-over-DNS can provide a tunnel for an attacker into your "protected" network.
Virus kits mean that complex tricks become commonplace. It's common for a virus to go stealth now, because there's library code to root a machine. It's common for viruses to use a suite of anti-anti-virus tools, because there's library code for that. I doubt anyone's doing IP-over-DNS or other advanced external firewall circumvention techniques, but if anti-virus vendors ever get it into their heads that they could sell black boxes to sit between PC and intarwebs, offering SPI firewall scanning, it will happen.
Come to think of it, a black box virus firewall would be a product worth buying, if done even slightly right. It could alert on high outgoing mail volume, it could fetch a daily list of evil IPs to block and alert on, it could run a self-checking firmware OS to protect itself against tampering, and it could inspect packets for known virus message signatures. And it wouldn't bring your PC to its knees!
Then again, the half-assed work that anti-virus vendors do would probably drop packets all over the floor if your network speed exceeded 56.6k/sec.