Okay, you'd done your duty to the Microsoft propaganda dept.
Of course, one of the major reasons that the "Tomato OS" is so popular is Microsoft lockin. If you're using a MIcrosoft web server (like Bank of China does... did?) you really have to use the Microsoft browser for it to work properly but that's far too expensive for the working Chinese. Add to that the fact of the political decision to force the browser into the OS 'shell', ie explorer.exe. and Microsoft have effectively bazookered their feet.
So they have a "super popular" OS bonded to a "super unpopular" web browser. Well Microsoft standard practice is to use the upgrade treadmill and force some upgrades, make the new version incompatible with the old version to force the network effect to work for you. Problem is that doesn't work very well with pirated copies.
Still it gives the propaganda dept. a nice "poor Microsoft" sound-bite.
They have to show they're a worthwhile tenant for the main debian ftp archive or they will be forced to find a new home.... Sounds like they're being threatened with "eviction" to me!
Their new landlord would probably be the debian-ports archive, which is under a distinct administration, but this by no means certain. However, wherever they go, being removed from the main archive would be a severe blow to the port/project because it would normally only happen to a port for which hardware is almost impossible to find. An example would be the m68k port which has been moribund since Apple switched to the PowerPC processor.
The TODO list is getting better... but we shall see.
In a discussion with the Debian Hurd porters it was decided that the Hurd port stays on FTPMaster until Wheezy is released. Should they have managed to get the port into a state that it is released together with all the others (probably as a technology preview), it is kept in the archive. Should they not manage this the port will be removed from the main archive and move fully to debian-ports.org. It may then reenter the main archive whenever it is ready to get released with the next release. (Obviously when we say "move to debian-ports" this does not mean we expect the debian-ports people to "just eat it". They are running their archive and may have their own needs and pre-conditions prior to accepting a port, like getting help with the work that needs to be done or with the hardware for it, so any port who has to look for new place should ensure to coordinate with the involved people.) In case it does not work out with debian-ports.org, the removal from main will still be done, but we are confident that the teams can work out something acceptable.
A single WiFi 802.11g transmission occupies 16.25Mhz officially, with interference out to 20Mhz width, That's the amount of frequency space available here. But unlike WiFi that's limited to a couple of hundred metres an FM signal can go for miles with potentially thousands of transmitters in range, sharing the bandwidth.
With landline broadband starting to use 'fibre to the box' and getting headline speeds at or above the 54Mbps official 802.11g bandwidth and practical speeds exceeding a quarter of that a wireless broadband implementation trying to serve more than THREE users on this frequency is going to be 'slower'. Even for 'officially broadband' speeds you'd be lucky to be able to serve more than ten or twenty.
If I'm going to 'pull a number out of my airs' I wouldn't go above 100kbps for sustained bandwidth in a 'high use' area, and that high only because people would be leaving because of the 'crap speed'.
If it's done right this can be the best of both worlds.
You have to TURN OFF the traffic lights when there isn't much traffic.
When there's a lot of traffic the lights actually increase the total amount that can get through at the cost of it taking longer (on average) for a car to cross the roundabout. The one you're thinking of is probably configured for maximum capacity.
As for timing, they do sometimes try to make some routes round have green lights all the way, but usually this forces other routes have longer delays. Basically, you can't win them all without turning off the lights.
What are you stupid, roundabouts are great, especially if you don't know the area. Go round it once to check out all the exits then pick one on the second time around.
If you have a navigator who can't make up her mind, no problem, just go round again. Nobody needs to get pissed.
Magicroundabouts are even better, you can go round the outside one way, round the inside the other way and even change your mind half way through.
Plus notice all the little roundabouts on a magic roundabout are "mini-roundabouts" that means you can, quite legally go straight over. You still have to give way, but there's no turning needed.
I don't think that's quite right, what you're after is the sort of facilities provided by a "Continuous backup" system.
The "low-level kernel" can't be changed without the key and doesn't need to be changed unless you open the computer and change the hardware. It only provides one major service and so can reasonably be created without bugs.
The service it provides could be described as "protected storage". As the normal system works it saves changes to the "write once" storage managed by the "low-level kernel". A continuous backup.
If there's an "incident" the system gets rebooted in 'fix it' mode. The main system is "forked" (restored but don't delete any backups) from a time when it was clean and user data can be cherry picked out of later backups.
Windows system restore tries to do this but doesn't protect the backup storage area so while it's fine against an accidental problem it's no protection against malware as the malware just infects the backups too. In addition the restore of the backups doesn't work as it should either.
The closest I've seen is Puppy Linux, a tiny distribution that runs from a CD-R. The main system is loaded into memory and changes are burnt to the CD-R when you save the system. If you decide you don't like the last update you can ignore it when you reboot. Of course, CD's are slow and so it's not really a continuous backup.
The weird thing is I don't think I've seen a "flashing 12:00" after a power cut for many years. It either gets an an automatic clock reset PDQ or the clock keeps running (at least a few hours) through the power cut.
Then there's my watch... never needs new batteries, always gets DST right and even if I physically set it to the wrong time it's corrected itself the next time I look at it.
The generators will still have to synch to the grid so they don't get 'yanked' off their foundations but this change would stop the entire grid having to be 'yanked' to match the atomic clock network.
So the frequency on the grid would 'free run' at "almost exactly" 60Hz rather than being governed by the time correction equipment to be "exactly" 60Hz over the long term.
Bye bye time correctors and bye bye their power wastage,
There would be LESS variation than now. Now you get the variation that put the old clocks out of synch then more variation to put them back in synch. With this change the second lot of variations wouldn't happen.
All this is actually rather simple; for people there are two main methods of navigating, call them the cartesian and graph.
The mechanics is easy, you have a rather effective giro-compass in your inner ear most of the time it works very well, but it's not attached to any particular absolute directions. To provide the base for this you have what could be described as "local knowledge" -- at some point/times you know which way is north because there's a sign that says 'to the north' or the sun is in the south or main other similar clues. Though some people never base it in this way, instead they give themselves a 'journey base' and their "north" is the direct route to the destination.
If you navigate by "graph" you're not really using the giro part; you're going to landmarks, you're taking left and right turns you're following the route. If it's used your giro is just giving you a feeling that you're headed in the right general direction.
If you're navigating by cartesian you're not really following the map until you get near the end of the journey, you're just heading off "thataway" for "this long" near the end of the journey you have to "recognise the area" and switch to the graph method so you don't "overshoot" or "miss the turning".
Both methods have problems but normally it all works very well with the known capabilities; there's no need for magnets.
If Bill hadn't ripped off IBM and his supplier IBM would probably have bought a better OS from DEC. Even right at the beginning "Micro soft" wasn't a nice company.
As for IBM, plenty of companies had run-ins with the "Nazgul" when they tried to clone the PC, Compaq only succeeded because they managed to put together some very complete "clean room" documentation to prove their innocence. And even after that it was believed, with strong evidence from the licenses, that the IBM PS/2 range appeared in part to try and kill the clone market.
As for patents, most large companies aren't patent trolls because many of their patents are crap. For a large tech companies patents are defensive, their only real use is to make it really unwise for another large company to sue them for patent infringement because frankly nobody can work out who would end up closest to the top after the reply.
Oddly enough IBM seem to be rather nicer nowadays, probably 'cause they're going back to their "service company" ways. I hope it isn't just in comparison to Microsoft, Apple and Sony.
PS: Nazgul; IBM's lawyers always wore black and always appeared in packs.
Naa, "Delay line memory" is far too open to interference.
I would suggest using Iron crystals; a small one would be a thousand miles across. They take a little while to manufacture but once they've cooled off they make a very stable storage medium. Just be sure to remove the outer layers of slag, especially any organic residue that may be infecting it.
One of these will give you a few zettabytes of storage. BTW: You can put a lot more in if you reduce the ECC requirements but we're talking long term storage here.
In addition for long term storage you'll need several of these to account for losses to stellar incidents and they should really be moved into intergalactic space.
IPv4 actually 'ran out' a while back, we passed the 5 billion devices connected to the (4 billion address) internet back in August '10. Massive NAT and restrictions on public IP address allocations means that IANA ran out quite a bit later. The restrictions are set to get even more severe but most of the NICs won't actually allow their reserves to completely run out for years. I'll just be nearly impossible to be allocated any addresses.
So Cisco are seeing that the current product line will continue to work as long as the ISP will provide any sort of super NAT'd IPv4 address. Only during the end game a few (perhaps five perhaps twenty) years from now will the end user IPv4 devices stop working and will 'mom and pop' have to do something.
Companies are different; Cisco's VPN software, used by many companies, doesn't work with multiple users behind a NAT. Any server software; including Cisco's needs a public address for the clients to connect to. IPv4 exhaustion is already hurting Cisco and their customers for the E4200 router.
PS: I personally have 14 devices with "Local Internet addresses" (talk about an oxymoron!) behind a single IP so I think that 5 billion is an underestimate.... I think I may have miscounted; a laptop with WiFi has two IP addresses.... except this one doesn't... except when it runs Windows... virtual machines too !!!!
APNIC will be first this table is current data and a current prediction.
Note the ARIN actually has the highest stock of IPs at the moment at 6.26/8s as compared to APNIC's 5.78/8s. Also ARIN are only setting aside 8 million addresses for IPv6 to IPv4 connectivity whereas the other NICs (except the much smaller LANIC) are setting aside a full/8.
This may mean that America's interconnects between IPv4 and IPv6 end up going via Africa!
Then you'll have to encapsulate it (UDP, use a tcp redirector, proxy etc. ) because to do it's job properly the program has to know where the connection was supposed to be going. If you've NAT'd the connection to some random machine you've destroyed this information.
The only time it doesn't matter is if you don't care where it was going and you just want to override everything you intercept to the same place.
The reason web proxies will still work is that the protocol has been altered to include the information in stream as the 'Host:' header. But if the client lies or omits the Host: header, like every other protocol, you're fucked by the NAT.
With NAT the packet has to be rewritten and regenerated. The entire communication has to be altered to convince both ends that they're talking to who the believe they are. With simple protocols this may be considered easy. With some protocols it's impossible to trick both ends into believing they're talking to someone else.
With interception no rewriting needs to be done. Once the packet's identified all you need to do is hand it to the program that's going to deal with it. The packet is unmodified and the program that receives the packet knows who it's pretending to be. You only have to convince one end that they're talking to who they asked for. If the program is a transparent http proxy it will start a completely independent connection to the web server where it explicitly states that it's doing this for someone else.
This asymmetry is the key difference because it means that an interception can be completely convincing to that end, basically any protocol that doesn't involve encryption can be intercepted flawlessly and even encryption can be done with collusion from the end that knows this is a fake. (By letting the interceptor 'borrow' the encryption secret)
A NAT box can do them as an edge case it's true and if you've already got a NAT capability it's an easy way to implement the task.
But a program to just intercept a connection doesn't need NAT. It just needs to capture the packets and pass them to a stack here; no packet mangling needed. It could currently be done by using ip6tables to capture the packets to a cut down IP stack in user space, though I expect it'll be added to ip6tables as a specific target because someone will get paid to write it, unlike NAT.
Come on! Windows' IPv6 stack was a 'developer release' in 2001, by 2003 it's "fully supported" ie: as bug free as Microsoft can make it. I don't know Mac OS well, but it's mostly BSD one of the core OSs for IPv6 support.
You're gonna have to do more that just assert that they are broken. Oh and please don't point at MS Vista, you should know that it's IP stack was a rewrite.
Do you think they didn't try to do an 'easy fix'? Five years of searching and nobody found one. Twenty years later (now) one of the quick alternatives to IPv6 (ie NAT) is causing serious problems. There were a lot of possibilities that didn't get to the short lists.
For TCP there's even a "dirty hack" that would have worked on day 0. You put up a NAT device (AKA 56bit router) and for your outgoing packets you bury the real (behind the NAT) address in a Timestamp option. For a 32bit TCP connection this is reflected by the remote and you have a stateless NAT ie: it's a router. For a 56bit TCP remote it can put the right timestamp value in the SYN packet and "hole punch" without a problem. You don't even need a distinct public ports on the NAT; the stamp tells you it's not for this router so there's no port mapping needed.
But this only works with TCP, not even UDP let alone all the other protocols (Ping, ipsec, gre...) and even with TCP it's really nasty if both ends are NAT'd. To get rid of the NAT you'd have to upgrade the ARP protocol to understand 56bit addresses... okay...
The problem with these schemes is that everyone would still need an IPv4 address to be a "smart host" for them and nobody would want to have a second class IPv4.5 address. Even if someone doesn't know what an IP address is they'd known there are two types and you don't want the wrong sort 'cause things will break.
In the end every node (and protocol) would have to be upgraded to understand the full IPv4.5 address so that it wouldn't break anything... and you have IPv6 [dirty version] again.
It appears that this happens every time you break the 'end to end principal'. It's simple really, if a node's only job is to send a packet to the right address that's all that can go wrong. As soon as you give a node another job it's another thing that can go wrong. Which brings us to a guy named Murphy.
Okay, firstly it goes against the principals of the organisations that distribute the IP addresses were setup with.
But that's not the main problem. The main problem is the routing table, this is a huge table that lists he physical location of every single IP address and it sits in the core routers of the internet. The table is compressed tremendously because long runs of addresses are located at the same place as seen from these routers so that the table is only a few hundred thousand entries long. At that size it's a significant problem, a problem that IPv6 doesn't have because the allocation is a lot less random. To get a viable market you would need to allow small groups of IP addresses to migrate to anywhere on the planet. Every time you do this you will increase the size of the table and slow down the routers even more.
The estimate is that the routers may be able to cope with around a million routes, if you replace most of the border routers with top of the range ones. Put it another way the smallest saleable unit would be 4096 addresses.
One million units total of 4000 IPs each doesn't sound like a good market to me.
Slashdot Mirror
Okay, you'd done your duty to the Microsoft propaganda dept.
Of course, one of the major reasons that the "Tomato OS" is so popular is Microsoft lockin. If you're using a MIcrosoft web server (like Bank of China does ... did?) you really have to use the Microsoft browser for it to work properly but that's far too expensive for the working Chinese. Add to that the fact of the political decision to force the browser into the OS 'shell', ie explorer.exe. and Microsoft have effectively bazookered their feet.
So they have a "super popular" OS bonded to a "super unpopular" web browser. Well Microsoft standard practice is to use the upgrade treadmill and force some upgrades, make the new version incompatible with the old version to force the network effect to work for you.
Problem is that doesn't work very well with pirated copies.
Still it gives the propaganda dept. a nice "poor Microsoft" sound-bite.
It's only IE8+ that doesn't run on XP.
All the others run on XP and several of them will still run on 2000 and even 95/98.
The problem is that activeX seems to be very popular in Chinese websites.
They have to show they're a worthwhile tenant for the main debian ftp archive or they will be forced to find a new home. ... Sounds like they're being threatened with "eviction" to me!
Their new landlord would probably be the debian-ports archive, which is under a distinct administration, but this by no means certain. However, wherever they go, being removed from the main archive would be a severe blow to the port/project because it would normally only happen to a port for which hardware is almost impossible to find. An example would be the m68k port which has been moribund since Apple switched to the PowerPC processor.
From the minutes of the March 2011 FTPMaster meeting if it's not ready for some sort of release it will be evicted from the main archives.
The TODO list is getting better ... but we shall see.
20Mhz! Don't make me cry!
A single WiFi 802.11g transmission occupies 16.25Mhz officially, with interference out to 20Mhz width, That's the amount of frequency space available here. But unlike WiFi that's limited to a couple of hundred metres an FM signal can go for miles with potentially thousands of transmitters in range, sharing the bandwidth.
With landline broadband starting to use 'fibre to the box' and getting headline speeds at or above the 54Mbps official 802.11g bandwidth and practical speeds exceeding a quarter of that a wireless broadband implementation trying to serve more than THREE users on this frequency is going to be 'slower'. Even for 'officially broadband' speeds you'd be lucky to be able to serve more than ten or twenty.
If I'm going to 'pull a number out of my airs' I wouldn't go above 100kbps for sustained bandwidth in a 'high use' area, and that high only because people would be leaving because of the 'crap speed'.
Fitbit blocked it; they've forced all the profiles to private then requested the search engines to do a fast rescan of their site.
Nothing to see, move along.
If it's done right this can be the best of both worlds.
You have to TURN OFF the traffic lights when there isn't much traffic.
When there's a lot of traffic the lights actually increase the total amount that can get through at the cost of it taking longer (on average) for a car to cross the roundabout. The one you're thinking of is probably configured for maximum capacity.
As for timing, they do sometimes try to make some routes round have green lights all the way, but usually this forces other routes have longer delays. Basically, you can't win them all without turning off the lights.
What are you stupid, roundabouts are great, especially if you don't know the area. Go round it once to check out all the exits then pick one on the second time around.
If you have a navigator who can't make up her mind, no problem, just go round again. Nobody needs to get pissed.
Magic roundabouts are even better, you can go round the outside one way, round the inside the other way and even change your mind half way through.
Plus notice all the little roundabouts on a magic roundabout are "mini-roundabouts" that means you can, quite legally go straight over. You still have to give way, but there's no turning needed.
I don't think that's quite right, what you're after is the sort of facilities provided by a "Continuous backup" system.
The "low-level kernel" can't be changed without the key and doesn't need to be changed unless you open the computer and change the hardware. It only provides one major service and so can reasonably be created without bugs.
The service it provides could be described as "protected storage". As the normal system works it saves changes to the "write once" storage managed by the "low-level kernel". A continuous backup.
If there's an "incident" the system gets rebooted in 'fix it' mode. The main system is "forked" (restored but don't delete any backups) from a time when it was clean and user data can be cherry picked out of later backups.
Windows system restore tries to do this but doesn't protect the backup storage area so while it's fine against an accidental problem it's no protection against malware as the malware just infects the backups too. In addition the restore of the backups doesn't work as it should either.
The closest I've seen is Puppy Linux, a tiny distribution that runs from a CD-R. The main system is loaded into memory and changes are burnt to the CD-R when you save the system. If you decide you don't like the last update you can ignore it when you reboot. Of course, CD's are slow and so it's not really a continuous backup.
The weird thing is I don't think I've seen a "flashing 12:00" after a power cut for many years. It either gets an an automatic clock reset PDQ or the clock keeps running (at least a few hours) through the power cut.
Then there's my watch ... never needs new batteries, always gets DST right and even if I physically set it to the wrong time it's corrected itself the next time I look at it.
The generators will still have to synch to the grid so they don't get 'yanked' off their foundations but this change would stop the entire grid having to be 'yanked' to match the atomic clock network.
So the frequency on the grid would 'free run' at "almost exactly" 60Hz rather than being governed by the time correction equipment to be "exactly" 60Hz over the long term.
Bye bye time correctors and bye bye their power wastage,
There would be LESS variation than now. Now you get the variation that put the old clocks out of synch then more variation to put them back in synch. With this change the second lot of variations wouldn't happen.
All this is actually rather simple; for people there are two main methods of navigating, call them the cartesian and graph.
The mechanics is easy, you have a rather effective giro-compass in your inner ear most of the time it works very well, but it's not attached to any particular absolute directions. To provide the base for this you have what could be described as "local knowledge" -- at some point/times you know which way is north because there's a sign that says 'to the north' or the sun is in the south or main other similar clues. Though some people never base it in this way, instead they give themselves a 'journey base' and their "north" is the direct route to the destination.
If you navigate by "graph" you're not really using the giro part; you're going to landmarks, you're taking left and right turns you're following the route. If it's used your giro is just giving you a feeling that you're headed in the right general direction.
If you're navigating by cartesian you're not really following the map until you get near the end of the journey, you're just heading off "thataway" for "this long" near the end of the journey you have to "recognise the area" and switch to the graph method so you don't "overshoot" or "miss the turning".
Both methods have problems but normally it all works very well with the known capabilities; there's no need for magnets.
If Bill hadn't ripped off IBM and his supplier IBM would probably have bought a better OS from DEC. Even right at the beginning "Micro soft" wasn't a nice company.
As for IBM, plenty of companies had run-ins with the "Nazgul" when they tried to clone the PC, Compaq only succeeded because they managed to put together some very complete "clean room" documentation to prove their innocence. And even after that it was believed, with strong evidence from the licenses, that the IBM PS/2 range appeared in part to try and kill the clone market.
As for patents, most large companies aren't patent trolls because many of their patents are crap. For a large tech companies patents are defensive, their only real use is to make it really unwise for another large company to sue them for patent infringement because frankly nobody can work out who would end up closest to the top after the reply.
Oddly enough IBM seem to be rather nicer nowadays, probably 'cause they're going back to their "service company" ways. I hope it isn't just in comparison to Microsoft, Apple and Sony.
PS: Nazgul; IBM's lawyers always wore black and always appeared in packs.
Naa, "Delay line memory" is far too open to interference.
I would suggest using Iron crystals; a small one would be a thousand miles across. They take a little while to manufacture but once they've cooled off they make a very stable storage medium. Just be sure to remove the outer layers of slag, especially any organic residue that may be infecting it.
One of these will give you a few zettabytes of storage. BTW: You can put a lot more in if you reduce the ECC requirements but we're talking long term storage here. In addition for long term storage you'll need several of these to account for losses to stellar incidents and they should really be moved into intergalactic space.
IPv4 actually 'ran out' a while back, we passed the 5 billion devices connected to the (4 billion address) internet back in August '10. Massive NAT and restrictions on public IP address allocations means that IANA ran out quite a bit later. The restrictions are set to get even more severe but most of the NICs won't actually allow their reserves to completely run out for years. I'll just be nearly impossible to be allocated any addresses.
So Cisco are seeing that the current product line will continue to work as long as the ISP will provide any sort of super NAT'd IPv4 address. Only during the end game a few (perhaps five perhaps twenty) years from now will the end user IPv4 devices stop working and will 'mom and pop' have to do something.
Companies are different; Cisco's VPN software, used by many companies, doesn't work with multiple users behind a NAT. Any server software; including Cisco's needs a public address for the clients to connect to. IPv4 exhaustion is already hurting Cisco and their customers for the E4200 router.
PS: I personally have 14 devices with "Local Internet addresses" (talk about an oxymoron!) behind a single IP so I think that 5 billion is an underestimate. ... I think I may have miscounted; a laptop with WiFi has two IP addresses. ... except this one doesn't ... except when it runs Windows ... virtual machines too !!!!
APNIC will be first this table is current data and a current prediction.
Note the ARIN actually has the highest stock of IPs at the moment at 6.26 /8s as compared to APNIC's 5.78 /8s. Also ARIN are only setting aside 8 million addresses for IPv6 to IPv4 connectivity whereas the other NICs (except the much smaller LANIC) are setting aside a full /8.
This may mean that America's interconnects between IPv4 and IPv6 end up going via Africa!
Can anyone find the IPv6 prefix that's assigned to this protocol?
And have they upgraded it to include tits yet?
Then you'll have to encapsulate it (UDP, use a tcp redirector, proxy etc. ) because to do it's job properly the program has to know where the connection was supposed to be going. If you've NAT'd the connection to some random machine you've destroyed this information.
The only time it doesn't matter is if you don't care where it was going and you just want to override everything you intercept to the same place.
The reason web proxies will still work is that the protocol has been altered to include the information in stream as the 'Host:' header. But if the client lies or omits the Host: header, like every other protocol, you're fucked by the NAT.
With NAT the packet has to be rewritten and regenerated. The entire communication has to be altered to convince both ends that they're talking to who the believe they are. With simple protocols this may be considered easy. With some protocols it's impossible to trick both ends into believing they're talking to someone else.
With interception no rewriting needs to be done. Once the packet's identified all you need to do is hand it to the program that's going to deal with it. The packet is unmodified and the program that receives the packet knows who it's pretending to be. You only have to convince one end that they're talking to who they asked for. If the program is a transparent http proxy it will start a completely independent connection to the web server where it explicitly states that it's doing this for someone else.
This asymmetry is the key difference because it means that an interception can be completely convincing to that end, basically any protocol that doesn't involve encryption can be intercepted flawlessly and even encryption can be done with collusion from the end that knows this is a fake. (By letting the interceptor 'borrow' the encryption secret)
Those are not NAT, they are "interception".
A NAT box can do them as an edge case it's true and if you've already got a NAT capability it's an easy way to implement the task.
But a program to just intercept a connection doesn't need NAT. It just needs to capture the packets and pass them to a stack here; no packet mangling needed. It could currently be done by using ip6tables to capture the packets to a cut down IP stack in user space, though I expect it'll be added to ip6tables as a specific target because someone will get paid to write it, unlike NAT.
Come on! Windows' IPv6 stack was a 'developer release' in 2001, by 2003 it's "fully supported" ie: as bug free as Microsoft can make it. I don't know Mac OS well, but it's mostly BSD one of the core OSs for IPv6 support.
You're gonna have to do more that just assert that they are broken. Oh and please don't point at MS Vista, you should know that it's IP stack was a rewrite.
Do you think they didn't try to do an 'easy fix'? Five years of searching and nobody found one. Twenty years later (now) one of the quick alternatives to IPv6 (ie NAT) is causing serious problems. There were a lot of possibilities that didn't get to the short lists.
For TCP there's even a "dirty hack" that would have worked on day 0. You put up a NAT device (AKA 56bit router) and for your outgoing packets you bury the real (behind the NAT) address in a Timestamp option. For a 32bit TCP connection this is reflected by the remote and you have a stateless NAT ie: it's a router. For a 56bit TCP remote it can put the right timestamp value in the SYN packet and "hole punch" without a problem. You don't even need a distinct public ports on the NAT; the stamp tells you it's not for this router so there's no port mapping needed.
But this only works with TCP, not even UDP let alone all the other protocols (Ping, ipsec, gre ...) and even with TCP it's really nasty if both ends are NAT'd. To get rid of the NAT you'd have to upgrade the ARP protocol to understand 56bit addresses... okay...
The problem with these schemes is that everyone would still need an IPv4 address to be a "smart host" for them and nobody would want to have a second class IPv4.5 address. Even if someone doesn't know what an IP address is they'd known there are two types and you don't want the wrong sort 'cause things will break.
In the end every node (and protocol) would have to be upgraded to understand the full IPv4.5 address so that it wouldn't break anything ... and you have IPv6 [dirty version] again.
It appears that this happens every time you break the 'end to end principal'. It's simple really, if a node's only job is to send a packet to the right address that's all that can go wrong. As soon as you give a node another job it's another thing that can go wrong. Which brings us to a guy named Murphy.
The camwhores site does have an IPv6 address too though.
Okay, firstly it goes against the principals of the organisations that distribute the IP addresses were setup with.
But that's not the main problem. The main problem is the routing table, this is a huge table that lists he physical location of every single IP address and it sits in the core routers of the internet. The table is compressed tremendously because long runs of addresses are located at the same place as seen from these routers so that the table is only a few hundred thousand entries long. At that size it's a significant problem, a problem that IPv6 doesn't have because the allocation is a lot less random. To get a viable market you would need to allow small groups of IP addresses to migrate to anywhere on the planet. Every time you do this you will increase the size of the table and slow down the routers even more.
The estimate is that the routers may be able to cope with around a million routes, if you replace most of the border routers with top of the range ones. Put it another way the smallest saleable unit would be 4096 addresses.
One million units total of 4000 IPs each doesn't sound like a good market to me.