Since the ISP that I currently pay to resolve my DNS does redirection (Comcast), I'm going to go with "Google is less evil" here, since they're willing to provide DNS service for free without redirection. I'll worry about potential evil after I escape the active, palpable, real evil I'm dealing with now.
If Virgin Media wants to continue to monitor your DNS requests they can just inspect the requests and log them as they flow through their network. You're not gaining any additional privacy by using Google's DNS. The way to do this properly is to have an encrypted DNS lookup like this. As far as I know this technology isn't working yet.
You are wrong. If you take the corrections in the code and graph them, they match the NOAA adjustments very closely. Read what I wrote again more carefully. What we have here is a consensus story, as in "the felons agreed on an alibi".
One would presume that our ability to measure temperatures improves with time, so the obvious best temperature to use for a base for adjustments would be the temperature at the final moment reported, or at least somewhat close if an average is needed. Which point you choose does not alter the shape of the graph. This would prevent misleading people from thinking that current measurements made with NIST calibrated instruments logged every few minutes during the day are being adjusted to compensate for our foggy understanding of the growth of Bristlecone Pines 2000 years ago. Adjusting temperatures by +.5F in 1998 when measured anomaly was far less than.5F is an indication that you're fudging your data. If your adjustment needs to be more than 10% of your data, your data doesn't correlate to the observed phenomena and your reported results will be subject to observer bias. Also, if you're reporting anomalies as deltas from some temperature it's good form to specify exactly how many degrees celsius (or Farenheit if that's your unit) that base point is.
Also, a prudent scientist who creates a model which interprets challenging data making difficult choices will run some basic tests against it. For example as a sanity check he will substitute for his raw data random noise, to see if his model creates the same impression as the real data. If random noise as a temperature input creates the same graph that's a clue that your model is the thing creating the graph, and not a changing climate.
You do know that that page is the origin of the graph image that I posted that drove you to post me a link to that page. Did you think I had not read it? Are you tired or something? You can see on the page the link to the graph I posted.
It adds up to blah, blah, we found a way to find a curve that fits the corrections the other guys were using, and now we can adjust our raw data to fit their conclusions, even though our raw data says quite the opposite. We hope nobody notices that even random data could be tortured in this way.
The data is corrected for time of day, urban heat island effects, and changes in measuring technique.
The corrections in temperature are more than the observed changes in temperature in many cases by more than five times. I'm sorry, but a field where analysis trumps observation by a factor of 5:1 is not a science.
ESR is a conspiracy nut. Has been for a while, actually. His comments on this are about as accurate as would be expected - i.e. a bunch of sensationalist BS:
ESR is a bright guy. When you've done as much to improve my life, maybe I'll respect your opinion as much. That's a tall hill to climb.
* The code in question is fudging temperature measurements from tree cores at high northern latitudes post-1960 in order to match actual temperatures. These aren't used in the reported global temperature figures, because they're known to be wrong.
The tree samples are from Bristlecone Pines. I know this tree, I grew up around them. They're an endangered species. It's a strip bark tree. They were told to not use strip bark trees. The bark of the tree happens to wander around the perimeter over time, sometime covering areas that haven't seen bark in centuries or a millenium. It's the oldest living individual organism on the planet. One study has different samples from the same tree differing on individual rings by six or seven sigmas. You might as well roll dice. You have to see the environment these trees are in to understand that a passing elk can influence their growth for a decade or more.
* The fudged numbers aren't actually used anywhere - the code that would use them is commented out.
I direct your attention to this graph from NOAA that mirrors these corrections. That's a polite way of saying you're full of shit.
* If the code in question was uncommented, it would plot the fudged and uncorrected data against each other, complete with an appropriate title and different colours for each line - hardly something you'd do if hiding the fact you were fudging the figures
The models themselves when substituting red noise for the temperature data give the hockey stick graph 99% of the time.
* In actual fact, as far as anyone can tell, none of this code was final. It appears to have been a temporary hack that was superseded by later code that calculated its own correction.
It was superceded by code that operated on data that was pre-corrected, the original source data for which is now "lost" but which can be reconstructed by removing the "correction".
* Even the corrected figures from the newer code don't seem to have been used anywhere. The only version of the MXR tree-core data anyone's been able to find in published papers is the uncorrected one.
See above.
* Oh, and not only was the issue with this data reported in a high-profile paper, it looks like the main author of the paper was the guy behind this code.
Counting as credibility the reputations fooled is a fool's game. We're talking about science. Science is observation, measurement, analysis and prediction. AGW is not science. AGW predicted a spike that didn't come a decade now. It's a show, and the show is over.
The site is a blog of Michael Mann, the man behind Mann-made global warming. For a contrasting viewpoint from real scientists you can read climateaudit.org.
Me, I'm not a scientist and I apparently don't have the reverence for a title that's self-ordained that you have. I am rather particular though that people employ symbols in appropriate ways. I don't think that "scientist" is synonymous with "entertaining storyteller". In my world the generous term for such a person is "bard".
When you modify your raw data to create the desired result, and then delete it, what you're doing isn't science. It's science fiction. Have a nice pdf.
Interesting you should use noise as a temperature input. In 2005, some researchers did just that. It turns out that feeding "red noise" into these bogus climate models instead of real data yields a "hockey stick" graph 99% of the time. link (pdf).
If you look at these correction factors they're virtually identical to the one NOAA uses. These are the only two terrestrial datasets that link us back using proxies to times before reliable measurements. The other datasets calibrate on this adjusted data. In many cases the adjustments are more than five times as large as the observed change present in the raw data. Using these correction factors you can show a 1997 temperature spike using gaussian noise as your temperature input. If you invert the corrections you can almost visualize glaciation returning to Florida by the end of the millenium, which oddly enough was the big climatic change we were worried about when I was a kid - on just as sound evidence.
It's probably rude to point this out, but 2k cores is roughly 43 of these. That's 11 4 socket servers. Less than 1/2 a rack using blades. That's pretty small for a top500 system.
When they tell me that all throughout the history of life on Earth up until 1960, trees grew rings at a specific rate in relation to temperature, and that in 1960 when datalogging was invented the trees suddenly stopped behaving in this way, I have questions about that. When they use that as a reason to add.5c to a calibrated instrument measurement made today, I have a problem with that. When they claim these adjustments that create an AGW signal not present in the raw data are necessary for good justifiable reasons that they can't show us the math for, well, that is not science.
You guys keep pointing back to the same realclimate.org website as if that proves anything. Shall I collect up a few DailyKos, Freeper and HillaryIs44 links to rebut? Realclimate.org is run by the same people who invented global warming. That content may be discussion, but it's not proof.
If you cherry pick the profitable users, with their high-margin high-ticket glitterboxes Apple's doing well. But what really matters is umumble mumble.
All you people are liars. Nobody uses Linux. Nobody! Just read our statistics. There are no real Linux users here. All those other people on all those other sites bashing windows are fictional too! The Free Software Consortium has hired a hundred thousand unemployed Haitian bloggers to type away at these comments sixteen hours a day under the cruelest conditions for pennies a day. They live in cages and actually use Windows 7 and IE 8 and have to pretend to hate all that is right and good. It's inhuman and must be stopped.
What's worse is the FSC funds their immoral astroturfing with the funds from their hacker network, exploiting and pwning our precious Windows machines from their secret hideouts in Eastern Europe and China. It's true! They pore over every orifice of every application desperately seeking to penetrate the purity of Windows and Office in a vile attempt to convert the world to the cruel tyrrany of their "freedoms" and "choice".
People argue left, and right and liberal and conservative. Here we have a group of people who went off the liberal end of anarchy and came out on the other side of unfettered capitalism. Soon there will be moguls, and the moguls will need a stable society to protect their winnings. The imposition of social order will be difficult at first, but they're heavily armed and well motivated.
If you were deliberately trying to falsify data, would you put a comment warning readers about it?
Don't you hate it when people use your own sig against your argument?
'All the world is queer save thee and me, and even thou art a little queer' -- Robert Owen, 1818
If you the data you have is hopelessly munged, and you're trying to get it to fit a predetermined curve so that your output can be deemed "correct", this is exactly what you do. It's wrong. It's not science. It's definitely not a good reason to turn the ship of state. But it's what you do.
The fudge factor array, when subtracted out of the result set, gives the same level noise data that all of the unbiased sources I've seen do. The only reasonable conclusion to draw is that it was removed as a filter when the base data was modified to fit, and the programmer who did the modification was so lazy he didn't even bother to edit out the calculation which creates a discarded result (which would amplify the modification doubly and cause the result not to fit) - thereby consuming processing resources for NO REASON.
It's a bad hack. Even if my reasoning is unsound, there's enough doubt here to prevent bold action on this suspect "model".
You're only allowed three or four seconds to make the first post on a slashdot article. It takes several hours to digest several terabytes of data. I didn't care about this issue at all until you were such a dick about it. Now that I care you're going to have to bring something better than "Nyah, Nyah" when I'm directly linking to primary sources like NOAA. And you should probably engage your spell chekker too.
Oh, and check your belt. We wore the BIG onion back in my day.;)
Let me just say that I'm not a purist. I agree with some who sum the measures of their instruments and find in those sums trends which are too fine to be measured by those instruments individually. I've argued this point with those admittedly more educated masters who taught me, and they shook their heads in disgust.
So perhaps I am equally unworthy to discern the truth, but I still know bullshit when I smell it.
Have you ever tried to have a rational discussion about climate change with someone who's either unaware of willfully ignorant of the science?
Have you ever tried to have a rational discussion about science with someone who's unaware of statistical analysis or the importance of reproducibility? It's like talking to a wall.
Take for example the raw climate data. It's level noise. Unless you add in adjustments like this and this it's completely boring annual measurements that vary but don't trend.
Adjust them, and they're sexy. They are alarming. They're a cause for action that makes the science interesting and important. We all like to be important, don't we?
Slashdot Mirror
No, Intel is very good but sometimes the best laid plans of mice and men aft gang agley and all that.
It's cool that they're not afraid to hang themselves out there like that. If you want to see something new you gotta scratch your feet on a new road.
An Itanium class part, then.
Since the ISP that I currently pay to resolve my DNS does redirection (Comcast), I'm going to go with "Google is less evil" here, since they're willing to provide DNS service for free without redirection. I'll worry about potential evil after I escape the active, palpable, real evil I'm dealing with now.
If Virgin Media wants to continue to monitor your DNS requests they can just inspect the requests and log them as they flow through their network. You're not gaining any additional privacy by using Google's DNS. The way to do this properly is to have an encrypted DNS lookup like this. As far as I know this technology isn't working yet.
You are wrong. If you take the corrections in the code and graph them, they match the NOAA adjustments very closely. Read what I wrote again more carefully. What we have here is a consensus story, as in "the felons agreed on an alibi".
One would presume that our ability to measure temperatures improves with time, so the obvious best temperature to use for a base for adjustments would be the temperature at the final moment reported, or at least somewhat close if an average is needed. Which point you choose does not alter the shape of the graph. This would prevent misleading people from thinking that current measurements made with NIST calibrated instruments logged every few minutes during the day are being adjusted to compensate for our foggy understanding of the growth of Bristlecone Pines 2000 years ago. Adjusting temperatures by +.5F in 1998 when measured anomaly was far less than .5F is an indication that you're fudging your data. If your adjustment needs to be more than 10% of your data, your data doesn't correlate to the observed phenomena and your reported results will be subject to observer bias. Also, if you're reporting anomalies as deltas from some temperature it's good form to specify exactly how many degrees celsius (or Farenheit if that's your unit) that base point is.
Also, a prudent scientist who creates a model which interprets challenging data making difficult choices will run some basic tests against it. For example as a sanity check he will substitute for his raw data random noise, to see if his model creates the same impression as the real data. If random noise as a temperature input creates the same graph that's a clue that your model is the thing creating the graph, and not a changing climate.
So I'll be reconfiguring the home router tonight. It's that simple. Ad services on the search pages aren't just annoying - they're a security risk.
Replying for a second time. Bad form, I know.
You do know that that page is the origin of the graph image that I posted that drove you to post me a link to that page. Did you think I had not read it? Are you tired or something? You can see on the page the link to the graph I posted.
It adds up to blah, blah, we found a way to find a curve that fits the corrections the other guys were using, and now we can adjust our raw data to fit their conclusions, even though our raw data says quite the opposite. We hope nobody notices that even random data could be tortured in this way.
The data is corrected for time of day, urban heat island effects, and changes in measuring technique.
The corrections in temperature are more than the observed changes in temperature in many cases by more than five times. I'm sorry, but a field where analysis trumps observation by a factor of 5:1 is not a science.
ESR is a conspiracy nut. Has been for a while, actually. His comments on this are about as accurate as would be expected - i.e. a bunch of sensationalist BS:
ESR is a bright guy. When you've done as much to improve my life, maybe I'll respect your opinion as much. That's a tall hill to climb.
* The code in question is fudging temperature measurements from tree cores at high northern latitudes post-1960 in order to match actual temperatures. These aren't used in the reported global temperature figures, because they're known to be wrong.
The tree samples are from Bristlecone Pines. I know this tree, I grew up around them. They're an endangered species. It's a strip bark tree. They were told to not use strip bark trees. The bark of the tree happens to wander around the perimeter over time, sometime covering areas that haven't seen bark in centuries or a millenium. It's the oldest living individual organism on the planet. One study has different samples from the same tree differing on individual rings by six or seven sigmas. You might as well roll dice. You have to see the environment these trees are in to understand that a passing elk can influence their growth for a decade or more.
* The fudged numbers aren't actually used anywhere - the code that would use them is commented out.
I direct your attention to this graph from NOAA that mirrors these corrections. That's a polite way of saying you're full of shit.
* If the code in question was uncommented, it would plot the fudged and uncorrected data against each other, complete with an appropriate title and different colours for each line - hardly something you'd do if hiding the fact you were fudging the figures
The models themselves when substituting red noise for the temperature data give the hockey stick graph 99% of the time.
* In actual fact, as far as anyone can tell, none of this code was final. It appears to have been a temporary hack that was superseded by later code that calculated its own correction.
It was superceded by code that operated on data that was pre-corrected, the original source data for which is now "lost" but which can be reconstructed by removing the "correction".
* Even the corrected figures from the newer code don't seem to have been used anywhere. The only version of the MXR tree-core data anyone's been able to find in published papers is the uncorrected one.
See above.
* Oh, and not only was the issue with this data reported in a high-profile paper, it looks like the main author of the paper was the guy behind this code.
Counting as credibility the reputations fooled is a fool's game. We're talking about science. Science is observation, measurement, analysis and prediction. AGW is not science. AGW predicted a spike that didn't come a decade now. It's a show, and the show is over.
The site is a blog of Michael Mann, the man behind Mann-made global warming. For a contrasting viewpoint from real scientists you can read climateaudit.org.
Me, I'm not a scientist and I apparently don't have the reverence for a title that's self-ordained that you have. I am rather particular though that people employ symbols in appropriate ways. I don't think that "scientist" is synonymous with "entertaining storyteller". In my world the generous term for such a person is "bard".
When you modify your raw data to create the desired result, and then delete it, what you're doing isn't science. It's science fiction. Have a nice pdf.
Interesting you should use noise as a temperature input. In 2005, some researchers did just that. It turns out that feeding "red noise" into these bogus climate models instead of real data yields a "hockey stick" graph 99% of the time. link (pdf).
If you look at these correction factors they're virtually identical to the one NOAA uses. These are the only two terrestrial datasets that link us back using proxies to times before reliable measurements. The other datasets calibrate on this adjusted data. In many cases the adjustments are more than five times as large as the observed change present in the raw data. Using these correction factors you can show a 1997 temperature spike using gaussian noise as your temperature input. If you invert the corrections you can almost visualize glaciation returning to Florida by the end of the millenium, which oddly enough was the big climatic change we were worried about when I was a kid - on just as sound evidence.
Don't worry. All those guys will get test samples and port their stuff. When the chip is available your favorite software will be well on the way.
It's probably rude to point this out, but 2k cores is roughly 43 of these. That's 11 4 socket servers. Less than 1/2 a rack using blades. That's pretty small for a top500 system.
The problem solves itself!
When they tell me that all throughout the history of life on Earth up until 1960, trees grew rings at a specific rate in relation to temperature, and that in 1960 when datalogging was invented the trees suddenly stopped behaving in this way, I have questions about that. When they use that as a reason to add .5c to a calibrated instrument measurement made today, I have a problem with that. When they claim these adjustments that create an AGW signal not present in the raw data are necessary for good justifiable reasons that they can't show us the math for, well, that is not science.
You guys keep pointing back to the same realclimate.org website as if that proves anything. Shall I collect up a few DailyKos, Freeper and HillaryIs44 links to rebut? Realclimate.org is run by the same people who invented global warming. That content may be discussion, but it's not proof.
If you cherry pick the profitable users, with their high-margin high-ticket glitterboxes Apple's doing well. But what really matters is umumble mumble.
All you people are liars. Nobody uses Linux. Nobody! Just read our statistics. There are no real Linux users here. All those other people on all those other sites bashing windows are fictional too! The Free Software Consortium has hired a hundred thousand unemployed Haitian bloggers to type away at these comments sixteen hours a day under the cruelest conditions for pennies a day. They live in cages and actually use Windows 7 and IE 8 and have to pretend to hate all that is right and good. It's inhuman and must be stopped.
What's worse is the FSC funds their immoral astroturfing with the funds from their hacker network, exploiting and pwning our precious Windows machines from their secret hideouts in Eastern Europe and China. It's true! They pore over every orifice of every application desperately seeking to penetrate the purity of Windows and Office in a vile attempt to convert the world to the cruel tyrrany of their "freedoms" and "choice".
People argue left, and right and liberal and conservative. Here we have a group of people who went off the liberal end of anarchy and came out on the other side of unfettered capitalism. Soon there will be moguls, and the moguls will need a stable society to protect their winnings. The imposition of social order will be difficult at first, but they're heavily armed and well motivated.
Ah, there really is nothing new under the sun.
If you were deliberately trying to falsify data, would you put a comment warning readers about it?
Don't you hate it when people use your own sig against your argument?
'All the world is queer save thee and me, and even thou art a little queer' -- Robert Owen, 1818
If you the data you have is hopelessly munged, and you're trying to get it to fit a predetermined curve so that your output can be deemed "correct", this is exactly what you do. It's wrong. It's not science. It's definitely not a good reason to turn the ship of state. But it's what you do.
The fudge factor array, when subtracted out of the result set, gives the same level noise data that all of the unbiased sources I've seen do. The only reasonable conclusion to draw is that it was removed as a filter when the base data was modified to fit, and the programmer who did the modification was so lazy he didn't even bother to edit out the calculation which creates a discarded result (which would amplify the modification doubly and cause the result not to fit) - thereby consuming processing resources for NO REASON.
It's a bad hack. Even if my reasoning is unsound, there's enough doubt here to prevent bold action on this suspect "model".
You're only allowed three or four seconds to make the first post on a slashdot article. It takes several hours to digest several terabytes of data. I didn't care about this issue at all until you were such a dick about it. Now that I care you're going to have to bring something better than "Nyah, Nyah" when I'm directly linking to primary sources like NOAA. And you should probably engage your spell chekker too.
Oh, and check your belt. We wore the BIG onion back in my day. ;)
Let me just say that I'm not a purist. I agree with some who sum the measures of their instruments and find in those sums trends which are too fine to be measured by those instruments individually. I've argued this point with those admittedly more educated masters who taught me, and they shook their heads in disgust.
So perhaps I am equally unworthy to discern the truth, but I still know bullshit when I smell it.
Have you ever tried to have a rational discussion about climate change with someone who's either unaware of willfully ignorant of the science?
Have you ever tried to have a rational discussion about science with someone who's unaware of statistical analysis or the importance of reproducibility? It's like talking to a wall.
Take for example the raw climate data. It's level noise. Unless you add in adjustments like this and this it's completely boring annual measurements that vary but don't trend.
Adjust them, and they're sexy. They are alarming. They're a cause for action that makes the science interesting and important. We all like to be important, don't we?