The carriers and the manufacturers interests aren't aligned here. The carriers benifit if you don't replace your phone (since as you say you pay for a new phone whether you take it or not). The manufactuers only make money if you take a new phone.
NAT will scale up to ~32k times what we have today. Then you're done IPv4 has about 3 billion usable addresses (given that class E is multicast, F is unusable and some space in A B and C will be taken up by subnetting overhead). According to wikipedia the "UN HIGH" estimate of world population by 2100 is 14 billion. That is about one IP for every 5 people. Even given that things won't be allocated evenly I think that is still a reasonable level of NAT.
Plus there is nothing stopping a NAT using the same natted local port for multiple connections provided they are on different servers (or maybe even different ports on the same server but that is riskier). With such a system in combination with forced proxying (with IPV6 wan sides on the proxy) for the big sites one could scale natted IPV4 service practically without bound.
Of course this will break protocols that try clever tricks for NAT traversal (though many NATs do that anyway).
their connection times out and the site is unreachable. What actually happens is the connection times out and then the browser requests the page over IPV4.
Unfortunately the browsers (at least firefox last time I tried it) seem to rely on the OS timeouts (which are very long) and don't seem to remember that V6 failed before. The result is the site works but is EXCRUCIATINGLY slow (IIRC well over a minute per page).
Well that was really my point, to these sorts of users those ISPs aren't shitty. I'd personally have said the shitty ISPs are the ones who perform badly even for light users- those that really are heavily oversubscribed for the bandwidth they have available. I'm probablly a bit biased because my last experiance with such an ISP was REALLY bad. The ISP division of pipex got bought out by tiscali and we belive (don't know for sure but reports from others at the time say it was almost certainly the case) my parents were migrated without being asked or even informed of the fact from 2MBps BT wholesale ADSL (not fast by modern standards but solid and reliable) to tiscali LLU. The router we were using started refusing to connect at all. A really old router we had haging arround (and had previously retired due to reliability problem) connected but at ridiculously low speeds. Mum called support repeatedly and each time was told an engineer would call back. Noone ever called back.
Eventually mum got a migration authentication code (which to their credit they gave us without any fuss), migrated to IDNET (we wanted a small friendly ISP in case there were further problems with the line and in any case BE hadn't arrived in our area at the time) and the problems went away.
Well that was really my point, to these sorts of users those ISPs aren't shitty. I'd personally have said the shitty ISPs are the ones who perform badly even for light users- those that really are heavily oversubscribed for the bandwidth they have available. Isn't that most of the cheap ones in the UK (how many cheap ones are left anyway? the only one I can think of are talktalk)? SKY are apparently both good and cheap but only if you already have SKY TV.
Which wouldn't really be a problem for download services.
The real question is whether the revenue gained from immediate release (hopefully) reducing piracy would compensate for the revenue lost from no longer being able to sell to a TV network as "new and exclusive". Personally I doubt it would.
This is true, but light users, which will be the target demographic here aren't ever going to notice the difference True, I only mentioned that bit because you said decent ISP, if you had said shitty ISP i'd have agreed with you.
The bigger issue is the one I mentioned in my second paragraph, if you already have a phone line and are prepared to sign a 12 month contract you can get (shitty) internet on top for less than this goverment package will cost. If you don't have a phone line or don't feel you can commit for that long then this governemnt subsidised mobile based deal will probabblly be the cheapest way to get online.
Yes, you can get upto 8mbps in the UK for £5 on a decent ISP If you are getting ADSL at that price in the UK and it's not a time limited offer you are on a really shitty ISP, heavilly traffic limited or most likely both.
Further pretty much all internet prices in the UK assume you either already have a fixed BT phone line* or will be taking (and paying for) phone service/line rental (some phone providers quote for line rental and call packages together, some quote for them seperately) from the provider as well. Afaict dry ADSL exists in theory but most ISPs either don't offer it or at least don't post prices for it online. The voice line rental effectively pays for the upkeep of the line.
Our cable company do offer internet on it's own but even in urban areas coverage is fairly spotty (unlike ADSL which afaict is availiable pretty much anywhere that is close enough to an exchange) can get that and it's relatively expensive.
* by which I mean a line physically provided by BT openreach, the phone service may or may not be provided by BT.
Afaict most gift cards in shops at least in the UK are only assigned their value when the purchaser checks out, the ones sitting on the shelves are worthless. I wonder if they do the same for game cards.
Apparently most consumers can't figure this out though. And even if they can what is their recourse?
I was under the impression that with most providers in the US you effectivly paid for a new phone every couple of years whether you wanted one or not (it used to be like that here in the UK too though it seems to have changed recently I don't know whether this was because of regulatory pressure of because of a provider breaking lockstep and others feeling forced to follow).
PAYG was/is an option but afaict it's one that only makes sense for very light users.
Which given that afaict most other digital download systems just sell at the price you select would seem to give you a few options
1: let amazon screw you 2: use amazon exclusively 3: avoid amazon completely 4: try to skirt the edge of the agreement by selling a slightly different product through amazon than through other vendors.
Whether that would work would depend on the exact details. It would be quite easy to setup the sale rules to defend against such tactics.
For example such a system could be limited to other racers and have a rule where in the event of multiple requests a random buyer was selected. Limiting the buyers to other racers would make it hard to find large numbers of people to plant and random selection would limit the effectiveness of a single plant.
Noone really cares about image formats anymore. JPEG, PNG and even GIF are good enough for their purposes, supported everywhere and are old enough that future patent issues are unlikely.
The same can't be said for video:( the MPEG LA members are pushing towards a world where users have to pay royalties on video encoding and decoding software and where free software has to chose between not playing and infringing patents.
Google is pushing against that and I really hope they will be sucessful.
Do we even know how to keep a mammoth captive in a way that does not harm the mammoth? I guess it depends on what exactly you mean by "does not harm". From a physical point of view we keep elephants in zoos all the time and afaict we have learnt to build exibits that contain them without causing them to hurt themselves in escape attempts. I'd expect mammoths to be mostly a matter of scaling up the techniques we have already developed for elephants.
Noone will ever really know the physiological side for any zoo animal. We can see outward signs of happiness but we will never know if they really know or care that they are being kept captive.
I'm sure giving birth to a mammoth will have no negative consequences.:-| Maybe some for the mother/keepers and maybe even those near the zoo if a fence fails but I very much doubt there will be any large scale consequences from breeding mammoths. We would be able to deal with them long before they got a chance to establish a breeding colony.
It's the small fast breeding animals you really have to watch as once released into an ecosystem they can be nearly impossible to eliminate.
Another thing to consider is that crude oil is a mixture which we seperate out. Then we crack some heavier fractions to make lighter fractions and in the process produce alkenes (hydrocarbons with a double bond) that are used to make plastics.
There is some flexibility. If you do less cracking then you get more heavy products and less light products but you also get less plastics.
and consider what happens for an average of about 1000 values. All depends on the TYPE of error. IIRC if it's an unweighted average and the errors are random unocorrelated and normally distributed you reduce your error by a factor of 1000.
But if those conditions don't hold it could be much higher! Afaict many people learn basic error analysis but forget about or ignore the underlying assumptions.
Well yes that's in fact what they did. The problem is deciding how much of the less reliable data to throw out. The real problem is decisions on what is an isn't anomalous make it very easy to cherry pick data to show the result you want by only going after problems if they work against your hypothesis and very difficult for an outsider to tell if that was indeed the motivation for selecting what problems to go after.
And afaict trying to estimate an average temperature of the earth for some particular time in the past is full of descisions on what data to reject as anomolous and of the remaining data how to apply weighting.
heh I think I glossed over the article and then came back sometime later to read the comments without remembering what the article was.
However reading the article 1.5FPS does not seem to make for a usable video player, sure you could crank down the resoloution and maybe optimise the code a bit but I'd still be surprised if this ever performs acceptablly.
Why not serve WebM over Flash did adobe ever implement that? They said they would implement VP8 but i've found no evidence of them actually doing so and afaict they only talked about implmenting VP8 not the rest of webm.
old magnetic ballasts run the light at 60Hz If things are working properly (e.g. nothing in the system is having a diode like affect) they should flicker at 120Hz* which is out of most but not all peoples limit of perception.
If the test showed statistical significance, it won't matter if they increase the sample size, the results will be the same.That's what statistical significance is FOR. Its a while since I did any stats but IIRC statistical significance means the chance of getting the results you got or worse results by random chance is lower than some threshold, usually quite a high one. A 5% significance level means that 1 in 20 tests you will get a positive result by random chance.
1 in 20 tests giving a false positive may not sound like much but think how many tests are being run and that you probablly only see those that show a positive result.
I still think there is a huge difference between a place where anyone with a bank account and money can get on the internet and a place where only people high up in the regime can get on the internet.
I think some of the PAYG mobile internet deals can even be used by people without a bank account as long as they have cash to buy top-up cards.
Demand paging in of executables made sense back when we only had a few megs of ram in a general purpose computer but I wonder if it's an idea that is past it's time and if we would be better off loading smaller executables in a single read operation. Especially on hard drive based systems (flash is better but still once you add all the layers on I bet a single large read is faster than many small ones)
Slashdot Mirror
The carriers and the manufacturers interests aren't aligned here. The carriers benifit if you don't replace your phone (since as you say you pay for a new phone whether you take it or not). The manufactuers only make money if you take a new phone.
NAT will scale up to ~32k times what we have today. Then you're done
IPv4 has about 3 billion usable addresses (given that class E is multicast, F is unusable and some space in A B and C will be taken up by subnetting overhead). According to wikipedia the "UN HIGH" estimate of world population by 2100 is 14 billion. That is about one IP for every 5 people. Even given that things won't be allocated evenly I think that is still a reasonable level of NAT.
Plus there is nothing stopping a NAT using the same natted local port for multiple connections provided they are on different servers (or maybe even different ports on the same server but that is riskier). With such a system in combination with forced proxying (with IPV6 wan sides on the proxy) for the big sites one could scale natted IPV4 service practically without bound.
Of course this will break protocols that try clever tricks for NAT traversal (though many NATs do that anyway).
their connection times out and the site is unreachable.
What actually happens is the connection times out and then the browser requests the page over IPV4.
Unfortunately the browsers (at least firefox last time I tried it) seem to rely on the OS timeouts (which are very long) and don't seem to remember that V6 failed before. The result is the site works but is EXCRUCIATINGLY slow (IIRC well over a minute per page).
Well that was really my point, to these sorts of users those ISPs aren't shitty. I'd personally have said the shitty ISPs are the ones who perform badly even for light users- those that really are heavily oversubscribed for the bandwidth they have available.
I'm probablly a bit biased because my last experiance with such an ISP was REALLY bad. The ISP division of pipex got bought out by tiscali and we belive (don't know for sure but reports from others at the time say it was almost certainly the case) my parents were migrated without being asked or even informed of the fact from 2MBps BT wholesale ADSL (not fast by modern standards but solid and reliable) to tiscali LLU. The router we were using started refusing to connect at all. A really old router we had haging arround (and had previously retired due to reliability problem) connected but at ridiculously low speeds. Mum called support repeatedly and each time was told an engineer would call back. Noone ever called back.
Eventually mum got a migration authentication code (which to their credit they gave us without any fuss), migrated to IDNET (we wanted a small friendly ISP in case there were further problems with the line and in any case BE hadn't arrived in our area at the time) and the problems went away.
Well that was really my point, to these sorts of users those ISPs aren't shitty. I'd personally have said the shitty ISPs are the ones who perform badly even for light users- those that really are heavily oversubscribed for the bandwidth they have available.
Isn't that most of the cheap ones in the UK (how many cheap ones are left anyway? the only one I can think of are talktalk)? SKY are apparently both good and cheap but only if you already have SKY TV.
Which wouldn't really be a problem for download services.
The real question is whether the revenue gained from immediate release (hopefully) reducing piracy would compensate for the revenue lost from no longer being able to sell to a TV network as "new and exclusive". Personally I doubt it would.
Afaict at least in the UK downloads of individual tracks from legal digital download services are counted towards the singles chart.
This is true, but light users, which will be the target demographic here aren't ever going to notice the difference
True, I only mentioned that bit because you said decent ISP, if you had said shitty ISP i'd have agreed with you.
The bigger issue is the one I mentioned in my second paragraph, if you already have a phone line and are prepared to sign a 12 month contract you can get (shitty) internet on top for less than this goverment package will cost. If you don't have a phone line or don't feel you can commit for that long then this governemnt subsidised mobile based deal will probabblly be the cheapest way to get online.
Yes, you can get upto 8mbps in the UK for £5 on a decent ISP
If you are getting ADSL at that price in the UK and it's not a time limited offer you are on a really shitty ISP, heavilly traffic limited or most likely both.
Further pretty much all internet prices in the UK assume you either already have a fixed BT phone line* or will be taking (and paying for) phone service/line rental (some phone providers quote for line rental and call packages together, some quote for them seperately) from the provider as well. Afaict dry ADSL exists in theory but most ISPs either don't offer it or at least don't post prices for it online. The voice line rental effectively pays for the upkeep of the line.
Our cable company do offer internet on it's own but even in urban areas coverage is fairly spotty (unlike ADSL which afaict is availiable pretty much anywhere that is close enough to an exchange) can get that and it's relatively expensive.
* by which I mean a line physically provided by BT openreach, the phone service may or may not be provided by BT.
Afaict most gift cards in shops at least in the UK are only assigned their value when the purchaser checks out, the ones sitting on the shelves are worthless. I wonder if they do the same for game cards.
Apparently most consumers can't figure this out though.
And even if they can what is their recourse?
I was under the impression that with most providers in the US you effectivly paid for a new phone every couple of years whether you wanted one or not (it used to be like that here in the UK too though it seems to have changed recently I don't know whether this was because of regulatory pressure of because of a provider breaking lockstep and others feeling forced to follow).
PAYG was/is an option but afaict it's one that only makes sense for very light users.
Which given that afaict most other digital download systems just sell at the price you select would seem to give you a few options
1: let amazon screw you
2: use amazon exclusively
3: avoid amazon completely
4: try to skirt the edge of the agreement by selling a slightly different product through amazon than through other vendors.
Whether that would work would depend on the exact details. It would be quite easy to setup the sale rules to defend against such tactics.
For example such a system could be limited to other racers and have a rule where in the event of multiple requests a random buyer was selected. Limiting the buyers to other racers would make it hard to find large numbers of people to plant and random selection would limit the effectiveness of a single plant.
Noone really cares about image formats anymore. JPEG, PNG and even GIF are good enough for their purposes, supported everywhere and are old enough that future patent issues are unlikely.
The same can't be said for video :( the MPEG LA members are pushing towards a world where users have to pay royalties on video encoding and decoding software and where free software has to chose between not playing and infringing patents.
Google is pushing against that and I really hope they will be sucessful.
Do we even know how to keep a mammoth captive in a way that does not harm the mammoth?
I guess it depends on what exactly you mean by "does not harm". From a physical point of view we keep elephants in zoos all the time and afaict we have learnt to build exibits that contain them without causing them to hurt themselves in escape attempts. I'd expect mammoths to be mostly a matter of scaling up the techniques we have already developed for elephants.
Noone will ever really know the physiological side for any zoo animal. We can see outward signs of happiness but we will never know if they really know or care that they are being kept captive.
I'm sure giving birth to a mammoth will have no negative consequences. :-|
Maybe some for the mother/keepers and maybe even those near the zoo if a fence fails but I very much doubt there will be any large scale consequences from breeding mammoths. We would be able to deal with them long before they got a chance to establish a breeding colony.
It's the small fast breeding animals you really have to watch as once released into an ecosystem they can be nearly impossible to eliminate.
Another thing to consider is that crude oil is a mixture which we seperate out. Then we crack some heavier fractions to make lighter fractions and in the process produce alkenes (hydrocarbons with a double bond) that are used to make plastics.
There is some flexibility. If you do less cracking then you get more heavy products and less light products but you also get less plastics.
and consider what happens for an average of about 1000 values.
All depends on the TYPE of error. IIRC if it's an unweighted average and the errors are random unocorrelated and normally distributed you reduce your error by a factor of 1000.
But if those conditions don't hold it could be much higher! Afaict many people learn basic error analysis but forget about or ignore the underlying assumptions.
Well yes that's in fact what they did. The problem is deciding how much of the less reliable data to throw out.
The real problem is decisions on what is an isn't anomalous make it very easy to cherry pick data to show the result you want by only going after problems if they work against your hypothesis and very difficult for an outsider to tell if that was indeed the motivation for selecting what problems to go after.
And afaict trying to estimate an average temperature of the earth for some particular time in the past is full of descisions on what data to reject as anomolous and of the remaining data how to apply weighting.
heh I think I glossed over the article and then came back sometime later to read the comments without remembering what the article was.
However reading the article 1.5FPS does not seem to make for a usable video player, sure you could crank down the resoloution and maybe optimise the code a bit but I'd still be surprised if this ever performs acceptablly.
Why not serve WebM over Flash
did adobe ever implement that? They said they would implement VP8 but i've found no evidence of them actually doing so and afaict they only talked about implmenting VP8 not the rest of webm.
old magnetic ballasts run the light at 60Hz
If things are working properly (e.g. nothing in the system is having a diode like affect) they should flicker at 120Hz* which is out of most but not all peoples limit of perception.
If the test showed statistical significance, it won't matter if they increase the sample size, the results will be the same.That's what statistical significance is FOR.
Its a while since I did any stats but IIRC statistical significance means the chance of getting the results you got or worse results by random chance is lower than some threshold, usually quite a high one. A 5% significance level means that 1 in 20 tests you will get a positive result by random chance.
1 in 20 tests giving a false positive may not sound like much but think how many tests are being run and that you probablly only see those that show a positive result.
he travelled back in time shot himself, I saw it on the BBC ;)
I still think there is a huge difference between a place where anyone with a bank account and money can get on the internet and a place where only people high up in the regime can get on the internet.
I think some of the PAYG mobile internet deals can even be used by people without a bank account as long as they have cash to buy top-up cards.
Demand paging in of executables made sense back when we only had a few megs of ram in a general purpose computer but I wonder if it's an idea that is past it's time and if we would be better off loading smaller executables in a single read operation. Especially on hard drive based systems (flash is better but still once you add all the layers on I bet a single large read is faster than many small ones)