Having read a lot about this flaw it's actually amazing that AMD's quality control found the problem in the first place.
The actions needed to cause the problem to arise are so extreme that they'd never happen in the field. i.e. Loop through tight floating-point only instructions without any comparisons for maybe hours before the error occurs.
This would *NEVER* happen in the field. Firstly, in any modern OS the process would have been pre-empted long before any problem could occur (causing other instructions to run and hence stopping the overheating). Secondly, no real-world program would ever do this sort of thing as there would always be a comparison in the loop within the timeframe.
This is a theoretical problem only in the real world, especially as it only affects about 3000 processors in total (it has been quoted). This is why AMD gave it such a low priority. We should just forget about it and move on.
I saw those BMWs in the media. In the UK it was classed as a motorcycle and hence the law required the wearing of a crash helmet.
As for the CLEVER, it looks like a bit of a rip off of a Carver One which has just gone on sale. Though the Carver One is definitely designed as more of a fun/performance vehicle.
Well, I think the original poster was from the Confederated Nations of North America, oh, sorry, United States of America. They're all strange over there.. they can't help it you see, it's the sugar rich diet.;-)
Actually, the two items you mention aren't linked at all.
The X25/X29 PAD addressing thing was very much akin to using the Internet without a DNS, that's all. A PAD was merely a terminal server which gave you a command line access. I've used TCP/IP terminal servers which were very similar.
The naming convention used in the UK for e-mail (which was supported long after the transition to TCP/IP) was purely that, an e-mail address convention. At the time it was decided upon the ARPAnet were making their own decision and opted for the opposite to the UK (and New Zealand). C'est la vie.
Before JANET transitioned to TCP/IP it was "interesting" keeping the mail system up to date. You had a special version of the Sendmail config called UK-Sendmail which had a list of every JANET mail server address. What fun!
Anyway, I always thought that ARPAnet got it the wrong way around for the domainnames as it's easier to parse a big-endian address. e.g. uk.ac.ucl.ts means, it's in the UK.. ok.. look that up and pass the message on.. then once in the UK, it's an "ac" academic address, pass it to the mail server which deals with that, and so on. Just like routing packets.:-)
Actually, the JSF is a joint project between the US and the UK. The "joint" part relates to more than the use of the aircraft in the joint services of the USA. That's why this is such a big thing.
From what I understand a lot of the avionics (as well as the engines etc.) are being partly developed by BAE Systems, a UK company, so the technology isn't even all US based.
Now, I can understand why the UK government wouldn't want to rely upon a closed source flight control system. The US government has stated in the past that it wants to be able to remotely shut down weapons systems sold to other nations if that nation is behaving against the interest of the USA.
Now, the term "behaving against the interest of the USA" doesn't necessarily mean attacking the USA or any of its forces. It can mean attacking a 3rd party which, for various reasons, the USA has a vested interest in or the use of the weapon in an area could be embarassing for the USA.
Why would any soverign country want to rely upon a weapon system that another country can shut down at will for defence, however friendly that country seems?
In the 1960's pressure from the US caused the cancellation of the british TSR-2 programme. The government cancelled the TSR-2 and ordered F-111's.. which were then cancelled a few years down the line. A total fiasco.
Similarly, all the plans and prototypes for the TSR-2 were destroyed.
Hmm.. I'd like to see how long overhead electric cabling would last on the sea-cliff section around Dawlish. I'm guessing one winter storm.
I've heard they had to take off the new trains on that route due to them getting swamped by sea water than the electrics failing. Looks like the old HSTs are going to have to run that route for some time.
The APT-E was the experimental gas turbine powered test train built in 1973/4. At the same time the prototype HST (Intercity 125 as it became) with the prototype (253001) running by 1975, the production versions (class 253 (great western) and the slightly more powerful class 254 (east coast mainline) going into service in 1977 as a stop-gap as the APT would take a while to come into production.
The HST vehicles (and the Mark 3 coaches) used technology developed in the APT project, such as high speed bogies, wheel design and brakes, the designs for which were licensed throughout the world.
In 1979/80 the APT-P vehicles were produced. These were the prototype technology test/demonstration machines and were electricly powered.
Due to the new Conservative government wanting to see a return on the money already spent on the APT project (which was in total less than 1Km of french TGV track) a political decision was made to force the prototypes into regular service before they were ready.
The inaugural journey was a comedy of errors. Firstly, it was known that the tilt system was not fully debugged and test had shown that some people became "air sick." So, the PR office plied a load of Fleet Street journalists with alcohol, piled them onto the train along with some minor celebrities and then gave them more drink.
Strangely enough the journos go sick and wrote about it. One car had a tilt failure half way through the journey and properly rotated upright and locked itself there. Strangely, the guard on the train agreed to a certain minor celebrity to stop the train at Carlisle to get off. Because of this the train lost its high-speed slot on the track and arrived late, which pleased the journos even more.. Fleet Street loves stamping on anything new and painting it in the blackest terms.
So, a PR disaster.
After being withdrawn from front-line service (for which the protoypes were never designed) they were used on and off on the West Coast Mainline until 1985, by which time all the bugs had been sorted out, they were reliable and it had been determined that the reason for the "air sickness" was due to the tilting being too good and not giving the brain enough hints that the person was going around the corner.
One set of the APT-Ps has been bought by a private buyer and the last I heard was sitting at Crewe.
The Pendolino trains are actually a decendant of a separate tilting train projetc in Italy, which initially used passive tilting. The technology and information gathered during the APT project was used by the italians after the APT project closed.
It is an interesting point that the West Coast Mainline had been given the green light for 155mph running for the APT in the early 80's using the existing lines and signalling. Yet in the late 90's it was stated by the railway authority that the new pendolino trains could only run on that line at 125mph until new signalling was designed, built and installed.
From my experience, I'd say the first "UNIX standard" was AT&T Version 7, which was the common ancestor. I seem to remember this being used as the base standard at least when it came to basic functionality and API.
// well, actually not so much the systems themselves, but the assinine UNIX mentality of "harder is better" and "more documentation eliminates the need for good design.", which set back Computer Science departments and academia 15 years behind industry. /// fortunately, one of the unintended side-effects of Linux is that the mentality, at least amongst Linux users, is slowly, ever so slowly, fading away.
Hmm.. yes, in/// you say that Linux programmers are going away from//. They are, they're just not doing the documentation.;-)
The main problem (at the time) of the Unix versions of Photoshop was the licensing cost. I remember at the time being stunned by the single-seat license being £1000 when the Macintosh version was less than a quarter of that. And they were the academic prices.
Basically, Adobe (and to be fair a great deal of other companies) thought that if you ran UNIX then you should pay a heavy premium. Maybe they thought UNIX == BigCorporation(tm) and hence they could get away with the huge price.
The problem with Skype is when it becomes a supernode. There are two reasons, bandwidth and the content of calls being relayed.
On the bandwidth side, because the University is on a high bandwidth connection, the host would become a prefered relay and get huge amounts of data directed through it.
On the content side, the JANET (Joint Academic NETwork) regulations state that no commercial use can be made of the WAN connection. It the host becomes a supernode there is no guarantee that the calls passing through it are all non-commercial.
So, as you can see, it's a thorny problem for such organisations.
I can see the reason why the hardware industry want DVD to die as it's got to the point where it's a commodity product. There are so many cheap machines on the market that it's impossible to make a reasonable profit on the devices. There is a very, very small market for high-end players for those with lots of money, but the mid-range has disappeared, where most of the profit could be made. The only solution is to generate a new market with a new product, for which they can get a large margin, at least initially.
Now, the extertainment industry are probably not that bothered either way. OK, if there's a new format for which they can sell yet another version of their product to the same customers then they're for that, but even if there isn't, they can still sell the new products very profitably, thank-you very much.
Nah, everyone knows that Star Trek is real and that all space ships and space stations have gravity plating under the desk plating. It must be true, it's on the telly!
Of course this will mean that to survive the contestants will have to use the ParticleOfTheWeek to thwart the danger. eg. The quantum bozo-on emissions will need to be deflected using the primary communication array, or somesuch.
I imagine that part of the selection process included an IQ test with only those without GCSE's and an IQ below 50 allowed to take part. Otherwise the contestants might be able to work out that it was a hoax.
The term "ocean" when used by an Earth Scientist has a specific meaning which isn't quite the same as the common, everyday usage meaning a large expanse of water.
Geologically, an ocean is a region of the Earth's crust composed of basaltic rocks (of MORB composition) which is generated at a spreading ridge. Because the thickness of this type of crust is pretty uniformly 15Km thick, it's somewhat lower than sea level, hence the expanse of water. This is very different from continental crust which is granitic in composition, far less dense and usualy greater than 30Km thick.
You can get ocean-like spreading ridges elsewhere, one type of these being certain back-arc basins. Although they are look very similar, because of the different composition of the basalt, they are not considered to be oceans.
Now, the term "lake" is generally used to describe a body of fresh water which is laying on a continent. A sea (which isn't really a scientific term at all) is generally ocean water washing over a continental basin which is often surrounded by large islands or other land masses.
The term "land" is not really that scientific a term either, merely meaning the common term for the part of the Earth's surface not covered with water.
I hope this has helped to clarify this for you. The (over) simplification of scientific subject by journalists is a constant bain for scientists in all fields of study.
As mentioned in the article, in Europe we've been experimenting with this for some time.
I worked at the Camborne School of Mines Geothermal Energy Project in Cornwall during the late 80's and early 90's. The problem with the technology is that it's difficult to create a reliable reservour and the reservour itself will only last a few years.
During a University of Southampton review of the technology in 1990 it was determined that because you had to drill new holes every few years in a radial pattern, the maximum lifetime of a power station would be around 25 years. The cost per unit would be something in the region of 10 times that of conventional power stations.
Even with advances in drilling recently and the small scale experiments the EU have been doing around Saltz I wouldn't think the cost per unit will have come down by more than half.
The technology itself is basically the following:-
Drill a deep bore hole which becomes sub-horizontal near the end.
Pump a wall paper paste type substance containing sand down the hole at extremely high pressure to open up the joints in the rock. (The sand wedges in the cracks and keeps them open, hopefully.) During this process monitor the micro seismic events so as to locate where the joints have opened.
Drill a second bore hole to intercept the cloud of opened joints.
Pump water down one hole and out the other, hoping that the water doesn't short circuit or wander off. Also hope that the rocks are hot enough to generate super heated steam and that there's enough surface area to do so.
Repeat every 3-4 years as after that time you will have cooled the rocks in the first reservour to the point where they're not usable.
In the market sector where bling is the thing the motherboards are a fashion item, along with the processor and graphics card. Basically, this season's motherboard is out of fashion next season. Hard core gamers (at which these are targeted) will have swapped out the motherboard in a year or so, long before most of the motherboards with faulty capacitors will fail.
Anyway, why should the motherboard manufacturers care? The products will be out of warrenty anyway.
Well, before you can (usefully) use the IPv6 stack your operating system has first the main trunk internet carriers have to use IPv6 natively. This is because otherwise someone will have to translate your IPv6 packets into IPv4 ones somehow for you to access the resources on the net.
After this your ISP would need to support IPv6 natively so that your equipment could talk IPv6 to it.
Finally, any network equipment that you use to connect to your ISP would have to be IPv6 enabled otherwise the packets wouldn't get through. This would be especially the case for SOHO type equipment you can buy from your local retail outlet.
Without all of these the IPv6 stack sitting there in your operating system is useless unless you want to run IPv6 locally for your own systems as an isolated outpost.
Of course, there's very little push from the customers either as long as they see that IPv4 is "good enough" so the manufacturers aren't exactly unhappy to hold back and wait.
It doesn't help that IPv6 has been in gestation for so long and still seems to be being tweeked.
IPv6 will never be adopted by the mainstream unless it's pushed from the centre.
Unless the main core network carriers agree to accept only IPv6 traffic (with IPv4 tunnelled within it) there is no push for the ISPs to change. Indeed, there is a very good business reason NOT to change.. it costs money and the customers generally don't want it either.
It's only at the point where ISPs have to do the work converting IPv4 traffic to IPv6 that the extended transition period could begin. Having forced the ISPs to use IPv6 they would see that it would cost them little or nothing to offer IPv6 directly to their customers.
Over an extended period of many years, first the big customers would slowly convert and as they do and spend money the cost of the IPv6 equipment will fall and the quality of the firmware will increase, finally trickling down to the consumer space.
As I see it, this period would take a long time, probably a decade or more with IPv4 never really dying out until all the systems which used it disappear. IPv4 will probably hang on in legacy systems for another 20 years or so.
Anyone who believes that there will be a "big-bang" where everyone changes over to IPv6 at the same time are living in a fantasy world.
Slashdot Mirror
Having read a lot about this flaw it's actually amazing that AMD's quality control found the problem in the first place.
The actions needed to cause the problem to arise are so extreme that they'd never happen in the field. i.e. Loop through tight floating-point only instructions without any comparisons for maybe hours before the error occurs.
This would *NEVER* happen in the field. Firstly, in any modern OS the process would have been pre-empted long before any problem could occur (causing other instructions to run and hence stopping the overheating). Secondly, no real-world program would ever do this sort of thing as there would always be a comparison in the loop within the timeframe.
This is a theoretical problem only in the real world, especially as it only affects about 3000 processors in total (it has been quoted). This is why AMD gave it such a low priority. We should just forget about it and move on.
I saw those BMWs in the media. In the UK it was classed as a motorcycle and hence the law required the wearing of a crash helmet.
As for the CLEVER, it looks like a bit of a rip off of a Carver One which has just gone on sale. Though the Carver One is definitely designed as more of a fun/performance vehicle.
Well, I think the original poster was from the Confederated Nations of North America, oh, sorry, United States of America. They're all strange over there.. they can't help it you see, it's the sugar rich diet. ;-)
No, no!
The chocolate eggs are laid by special, genetically modified Easter bunnies.
You just don't want to see the ones which produce the Cadbury's Creame Eggs though... one phrase... "Holy Hand Grenade of Antioch." Nasty creatures!
(nb. Bunnies were called cunnies/connies before the Victorians changed their name 'cos they sounded too much like the slang term for something else.)
Actually, the two items you mention aren't linked at all.
:-)
The X25/X29 PAD addressing thing was very much akin to using the Internet without a DNS, that's all. A PAD was merely a terminal server which gave you a command line access. I've used TCP/IP terminal servers which were very similar.
The naming convention used in the UK for e-mail (which was supported long after the transition to TCP/IP) was purely that, an e-mail address convention. At the time it was decided upon the ARPAnet were making their own decision and opted for the opposite to the UK (and New Zealand). C'est la vie.
Before JANET transitioned to TCP/IP it was "interesting" keeping the mail system up to date. You had a special version of the Sendmail config called UK-Sendmail which had a list of every JANET mail server address. What fun!
Anyway, I always thought that ARPAnet got it the wrong way around for the domainnames as it's easier to parse a big-endian address. e.g. uk.ac.ucl.ts means, it's in the UK.. ok.. look that up and pass the message on.. then once in the UK, it's an "ac" academic address, pass it to the mail server which deals with that, and so on. Just like routing packets.
Actually, the JSF is a joint project between the US and the UK. The "joint" part relates to more than the use of the aircraft in the joint services of the USA. That's why this is such a big thing.
From what I understand a lot of the avionics (as well as the engines etc.) are being partly developed by BAE Systems, a UK company, so the technology isn't even all US based.
Now, I can understand why the UK government wouldn't want to rely upon a closed source flight control system. The US government has stated in the past that it wants to be able to remotely shut down weapons systems sold to other nations if that nation is behaving against the interest of the USA.
Now, the term "behaving against the interest of the USA" doesn't necessarily mean attacking the USA or any of its forces. It can mean attacking a 3rd party which, for various reasons, the USA has a vested interest in or the use of the weapon in an area could be embarassing for the USA.
Why would any soverign country want to rely upon a weapon system that another country can shut down at will for defence, however friendly that country seems?
Well, if they get the planes without an OS, they can always run Linux on it!
</Obliglatory Slashdot Linux Reference>
In the 1960's pressure from the US caused the cancellation of the british TSR-2 programme. The government cancelled the TSR-2 and ordered F-111's.. which were then cancelled a few years down the line. A total fiasco.
Similarly, all the plans and prototypes for the TSR-2 were destroyed.
Of course, the class 91 was build as a modified APT power car, of course.
Hmm.. I'd like to see how long overhead electric cabling would last on the sea-cliff section around Dawlish. I'm guessing one winter storm.
I've heard they had to take off the new trains on that route due to them getting swamped by sea water than the electrics failing. Looks like the old HSTs are going to have to run that route for some time.
Actually...
The APT-E was the experimental gas turbine powered test train built in 1973/4. At the same time the prototype HST (Intercity 125 as it became) with the prototype (253001) running by 1975, the production versions (class 253 (great western) and the slightly more powerful class 254 (east coast mainline) going into service in 1977 as a stop-gap as the APT would take a while to come into production.
The HST vehicles (and the Mark 3 coaches) used technology developed in the APT project, such as high speed bogies, wheel design and brakes, the designs for which were licensed throughout the world.
In 1979/80 the APT-P vehicles were produced. These were the prototype technology test/demonstration machines and were electricly powered.
Due to the new Conservative government wanting to see a return on the money already spent on the APT project (which was in total less than 1Km of french TGV track) a political decision was made to force the prototypes into regular service before they were ready.
The inaugural journey was a comedy of errors. Firstly, it was known that the tilt system was not fully debugged and test had shown that some people became "air sick." So, the PR office plied a load of Fleet Street journalists with alcohol, piled them onto the train along with some minor celebrities and then gave them more drink.
Strangely enough the journos go sick and wrote about it. One car had a tilt failure half way through the journey and properly rotated upright and locked itself there. Strangely, the guard on the train agreed to a certain minor celebrity to stop the train at Carlisle to get off. Because of this the train lost its high-speed slot on the track and arrived late, which pleased the journos even more.. Fleet Street loves stamping on anything new and painting it in the blackest terms.
So, a PR disaster.
After being withdrawn from front-line service (for which the protoypes were never designed) they were used on and off on the West Coast Mainline until 1985, by which time all the bugs had been sorted out, they were reliable and it had been determined that the reason for the "air sickness" was due to the tilting being too good and not giving the brain enough hints that the person was going around the corner.
One set of the APT-Ps has been bought by a private buyer and the last I heard was sitting at Crewe.
The Pendolino trains are actually a decendant of a separate tilting train projetc in Italy, which initially used passive tilting. The technology and information gathered during the APT project was used by the italians after the APT project closed.
It is an interesting point that the West Coast Mainline had been given the green light for 155mph running for the APT in the early 80's using the existing lines and signalling. Yet in the late 90's it was stated by the railway authority that the new pendolino trains could only run on that line at 125mph until new signalling was designed, built and installed.
From my experience, I'd say the first "UNIX standard" was AT&T Version 7, which was the common ancestor. I seem to remember this being used as the base standard at least when it came to basic functionality and API.
// well, actually not so much the systems themselves, but the assinine UNIX mentality of "harder is better" and "more documentation eliminates the need for good design.", which set back Computer Science departments and academia 15 years behind industry.
/// fortunately, one of the unintended side-effects of Linux is that the mentality, at least amongst Linux users, is slowly, ever so slowly, fading away.
/// you say that Linux programmers are going away from //. They are, they're just not doing the documentation. ;-)
Hmm.. yes, in
The main problem (at the time) of the Unix versions of Photoshop was the licensing cost. I remember at the time being stunned by the single-seat license being £1000 when the Macintosh version was less than a quarter of that. And they were the academic prices.
Basically, Adobe (and to be fair a great deal of other companies) thought that if you ran UNIX then you should pay a heavy premium. Maybe they thought UNIX == BigCorporation(tm) and hence they could get away with the huge price.
Actually, after discussions with Skype, Oxford University has now allowed the use of Skype with specific conditions for its use and configuration.
The problem with Skype is when it becomes a supernode. There are two reasons, bandwidth and the content of calls being relayed.
On the bandwidth side, because the University is on a high bandwidth connection, the host would become a prefered relay and get huge amounts of data directed through it.
On the content side, the JANET (Joint Academic NETwork) regulations state that no commercial use can be made of the WAN connection. It the host becomes a supernode there is no guarantee that the calls passing through it are all non-commercial.
So, as you can see, it's a thorny problem for such organisations.
I can see the reason why the hardware industry want DVD to die as it's got to the point where it's a commodity product. There are so many cheap machines on the market that it's impossible to make a reasonable profit on the devices. There is a very, very small market for high-end players for those with lots of money, but the mid-range has disappeared, where most of the profit could be made. The only solution is to generate a new market with a new product, for which they can get a large margin, at least initially.
Now, the extertainment industry are probably not that bothered either way. OK, if there's a new format for which they can sell yet another version of their product to the same customers then they're for that, but even if there isn't, they can still sell the new products very profitably, thank-you very much.
Nah, everyone knows that Star Trek is real and that all space ships and space stations have gravity plating under the desk plating. It must be true, it's on the telly!
Of course this will mean that to survive the contestants will have to use the ParticleOfTheWeek to thwart the danger. eg. The quantum bozo-on emissions will need to be deflected using the primary communication array, or somesuch.
I imagine that part of the selection process included an IQ test with only those without GCSE's and an IQ below 50 allowed to take part. Otherwise the contestants might be able to work out that it was a hoax.
It's a pity that the name "Spaced" has been already taken for a TV show in that case.
The term "ocean" when used by an Earth Scientist has a specific meaning which isn't quite the same as the common, everyday usage meaning a large expanse of water.
Geologically, an ocean is a region of the Earth's crust composed of basaltic rocks (of MORB composition) which is generated at a spreading ridge. Because the thickness of this type of crust is pretty uniformly 15Km thick, it's somewhat lower than sea level, hence the expanse of water. This is very different from continental crust which is granitic in composition, far less dense and usualy greater than 30Km thick.
You can get ocean-like spreading ridges elsewhere, one type of these being certain back-arc basins. Although they are look very similar, because of the different composition of the basalt, they are not considered to be oceans.
Now, the term "lake" is generally used to describe a body of fresh water which is laying on a continent. A sea (which isn't really a scientific term at all) is generally ocean water washing over a continental basin which is often surrounded by large islands or other land masses.
The term "land" is not really that scientific a term either, merely meaning the common term for the part of the Earth's surface not covered with water.
I hope this has helped to clarify this for you. The (over) simplification of scientific subject by journalists is a constant bain for scientists in all fields of study.
As mentioned in the article, in Europe we've been experimenting with this for some time.
I worked at the Camborne School of Mines Geothermal Energy Project in Cornwall during the late 80's and early 90's. The problem with the technology is that it's difficult to create a reliable reservour and the reservour itself will only last a few years.
During a University of Southampton review of the technology in 1990 it was determined that because you had to drill new holes every few years in a radial pattern, the maximum lifetime of a power station would be around 25 years. The cost per unit would be something in the region of 10 times that of conventional power stations.
Even with advances in drilling recently and the small scale experiments the EU have been doing around Saltz I wouldn't think the cost per unit will have come down by more than half.
The technology itself is basically the following:-
Drill a deep bore hole which becomes sub-horizontal near the end.
Pump a wall paper paste type substance containing sand down the hole at extremely high pressure to open up the joints in the rock. (The sand wedges in the cracks and keeps them open, hopefully.) During this process monitor the micro seismic events so as to locate where the joints have opened.
Drill a second bore hole to intercept the cloud of opened joints.
Pump water down one hole and out the other, hoping that the water doesn't short circuit or wander off. Also hope that the rocks are hot enough to generate super heated steam and that there's enough surface area to do so.
Repeat every 3-4 years as after that time you will have cooled the rocks in the first reservour to the point where they're not usable.
There is another point.
In the market sector where bling is the thing the motherboards are a fashion item, along with the processor and graphics card. Basically, this season's motherboard is out of fashion next season. Hard core gamers (at which these are targeted) will have swapped out the motherboard in a year or so, long before most of the motherboards with faulty capacitors will fail.
Anyway, why should the motherboard manufacturers care? The products will be out of warrenty anyway.
Well, before you can (usefully) use the IPv6 stack your operating system has first the main trunk internet carriers have to use IPv6 natively. This is because otherwise someone will have to translate your IPv6 packets into IPv4 ones somehow for you to access the resources on the net.
After this your ISP would need to support IPv6 natively so that your equipment could talk IPv6 to it.
Finally, any network equipment that you use to connect to your ISP would have to be IPv6 enabled otherwise the packets wouldn't get through. This would be especially the case for SOHO type equipment you can buy from your local retail outlet.
Without all of these the IPv6 stack sitting there in your operating system is useless unless you want to run IPv6 locally for your own systems as an isolated outpost.
Indeed.
Of course, there's very little push from the customers either as long as they see that IPv4 is "good enough" so the manufacturers aren't exactly unhappy to hold back and wait.
It doesn't help that IPv6 has been in gestation for so long and still seems to be being tweeked.
IPv6 will never be adopted by the mainstream unless it's pushed from the centre.
Unless the main core network carriers agree to accept only IPv6 traffic (with IPv4 tunnelled within it) there is no push for the ISPs to change. Indeed, there is a very good business reason NOT to change.. it costs money and the customers generally don't want it either.
It's only at the point where ISPs have to do the work converting IPv4 traffic to IPv6 that the extended transition period could begin. Having forced the ISPs to use IPv6 they would see that it would cost them little or nothing to offer IPv6 directly to their customers.
Over an extended period of many years, first the big customers would slowly convert and as they do and spend money the cost of the IPv6 equipment will fall and the quality of the firmware will increase, finally trickling down to the consumer space.
As I see it, this period would take a long time, probably a decade or more with IPv4 never really dying out until all the systems which used it disappear. IPv4 will probably hang on in legacy systems for another 20 years or so.
Anyone who believes that there will be a "big-bang" where everyone changes over to IPv6 at the same time are living in a fantasy world.