I do know what you are getting at and I don't want to see them die, but I don't think you quite grasp how *big* this crater is. It is 13 miles in diameter. Opportunity's total odometer is under 13 miles. It still has another 8 miles to go to reach the crater. Then, to even get *around* would require a doubling of its current lifetime. The crater is just *that* big.
No, sadly, this will be the ultimate destination, but it will be an AMAZING one!
I couldn't wait for my university to upgrade so that I could use Firefox to access my professor's calendar (it was the only way to make appointments). I have no access to any Windows machines, except for the severely crippled machines in the library. Then came rollout, only to discover that OWA uses the browser identification string to determine which modes your browser can do. So, if you are running a rebranded version of Firefox, forget it. If you are using a beta version of Firefox, forget it. You have to go and change your Browser Identity string so that it conforms with the string of an official released version.
I should point out that GMail allows you to log in using eLinks. OWA doesn't even let you attempt to log in.
Bingo, that was the issue I had to deal with. I had to upgrade some code to deal with a larger input dataset. Untangling the code to reduce the number of copy-on-write actions was not fun.
Matlab is a pure scripting language, therefore, the parser would have to re-evaluate the body of a loop at each iteration (although, I think the latest version of Matlab might be using some sort of JIT-like parser to improve things).
Exactly, for accessing purposes, this isn't an issue. However, I find that NumPy's memory management approach is much smarter than Matlab's, and so that, along with other reasons, pushed me away from Matlab.
Both Matlab and SciPy/NumPy use the same BLAS/ATLAS backend. So, a well-written Matlab program will be comparable to a well-written Python program in speed. However, Python has Matlab beat on memory usage. Matlab does have an extensively documented language with a very good help system.
In my personal opinion, I feel that Matlab was too awkward of a language when it came to do anything else besides math. For example, it made it difficult to use when one needed to do string manipulations to figure out what files to load. Meanwhile, Python is a full-fledged language that would allow me to do many supporting tasks for my math related work.
What I find odd about R/S/S+ was that the ones who tended to learn it better were the begineers who never programed in another language or had limited exposure. People like me who have been exposed to many mainstream languages find it frustrating.
The professor who taught the course where I had to use S+ told me something along the lines: "It is a language designed by statisticians, of course it will behave randomly!"
That is one of the things that pushed me away from Matlab. I kept on running into memory bounds whenever I tried to do things "the Matlab way" (i.e., no loops, vectorize, etc.). It seems that Matlab liked to copy my arrays all the time when I was going in and out of functions. It seems ridiculous to me that I would have to do things "the bad way" in order to get my work done.
I might be confusing what happened with CRU, but you are right... some of the data was from third-party sources and they wouldn't allow the release for everyone else to use. This was not desirable to the scientists involved, but their hands were bound in this case.
This is just another reason for why copyright laws need to change. Scientific data should be open for all to access. I should be able to create a webpage with links to my data and publish that along with any research papers.
And you are right... they didn't pull it out of their asses. There were *some* scientific literature out there saying that an ice age was pending. It was just a lot more interesting than the many more scientific literatures that were saying that the Earth was warming.
Yes, sorry for not being clear. PA replaces ALSA as the interface for applications. This will allow for easier (read less painful) changes in the future because there will be an abstraction layer. Userspace stuff is in userspace, kernel stuff is in the kernel (which is why I claim it is a superior design over just using ALSA).
Yes, OSS does have many of the features that PA promises, and so it would seem to be a more logical platform to move towards. But, wasn't there licensing issues surrounding it at one point?
Seems like Pulseaudio was not working properly with your sound card and you had to regress to ALSA. Did you file a bug report? I know several developers have been hard at work rewriting hardware drivers to be compatible with Pulseaudio. The more bug reports, the better as they seem to be churning out the changes based on the reports. If you don't want to commit to an install, you can easily try booting from the LiveCD first to test out the sound.
Fedora uses Pulseaudio. RedHat will too... Linux is migrating to Pulseaudio because it is a superior design over ALSA and OSS. The move has been painful, I admit, but that will happen whenever you mess around with fundamental OS components.
If you want flash and java working OOTB, then you need to head over to an unofficial distro like Linux Mint. The official Ubuntu releases can not come with flash installed by default (I can't remember if Java comes installed or not as I have been doing upgrades). As for pulseaudio, it is *much* better, but we are still encountering a few odds and ends with the "rarer" hardware (or undocumented hardware).
But if you cause me to take legal action you can bet if I win (and believe me the city was going to win) then I'm going to take my legal fees out of your pocket book.
Ok, and each of those links do not support your main assertion of "can't reproduce the same results with their data". If anything, I am quite amazed by the amount of information given in the README file. I frequently find myself bashing my head against a wall trying to figure out why I can't process some piece of data correctly or something or another, but I don't keep a diary like this guy did.
One of the links was a rant about a squared error value going negative and saying that any programmer would have caught that and knew how to deal with it. Well, in the README file, the guy did notice it, recognized that he was grabbing error values instead of obs and fixed the data to properly mask them out as they should have been in the first place. So, what's the problem?
In the end, it shows that he persevered in being able to reproduce much of the original data despite changes in the grid layout, system upgrades and software changes.
Whether it is true or false, posting such statements without attribution is just elitist and smug because you assume that everyone else in the conversation are aware of the same information that you have. It does no good to people in this forum to not provide sources for your statements.
I'm not knowledgeable about all the hardware, but shouldn't it be possible to reset the graphics system somehow without a full reboot?
Yes, it is possible, but it requires accessing the terminal. Often in the frozen graphics mode, just about all of the keypresses aren't getting through, so you can't do Ctrl-Alt-BkSp or do a Ctrl-F1 to get to a terminal. I haven't tried the SysRq (or whatever it is) tricks, though...
Slashdot Mirror
I do know what you are getting at and I don't want to see them die, but I don't think you quite grasp how *big* this crater is. It is 13 miles in diameter. Opportunity's total odometer is under 13 miles. It still has another 8 miles to go to reach the crater. Then, to even get *around* would require a doubling of its current lifetime. The crater is just *that* big.
No, sadly, this will be the ultimate destination, but it will be an AMAZING one!
True, but that wasn't the main issue. The program was taking up too much memory because it kept creating copies of the arrays.
I couldn't wait for my university to upgrade so that I could use Firefox to access my professor's calendar (it was the only way to make appointments). I have no access to any Windows machines, except for the severely crippled machines in the library. Then came rollout, only to discover that OWA uses the browser identification string to determine which modes your browser can do. So, if you are running a rebranded version of Firefox, forget it. If you are using a beta version of Firefox, forget it. You have to go and change your Browser Identity string so that it conforms with the string of an official released version.
I should point out that GMail allows you to log in using eLinks. OWA doesn't even let you attempt to log in.
Bingo, that was the issue I had to deal with. I had to upgrade some code to deal with a larger input dataset. Untangling the code to reduce the number of copy-on-write actions was not fun.
Matlab is a pure scripting language, therefore, the parser would have to re-evaluate the body of a loop at each iteration (although, I think the latest version of Matlab might be using some sort of JIT-like parser to improve things).
Exactly, for accessing purposes, this isn't an issue. However, I find that NumPy's memory management approach is much smarter than Matlab's, and so that, along with other reasons, pushed me away from Matlab.
In my personal opinion, I feel that Matlab was too awkward of a language when it came to do anything else besides math. For example, it made it difficult to use when one needed to do string manipulations to figure out what files to load. Meanwhile, Python is a full-fledged language that would allow me to do many supporting tasks for my math related work.
The professor who taught the course where I had to use S+ told me something along the lines: "It is a language designed by statisticians, of course it will behave randomly!"
That is one of the things that pushed me away from Matlab. I kept on running into memory bounds whenever I tried to do things "the Matlab way" (i.e., no loops, vectorize, etc.). It seems that Matlab liked to copy my arrays all the time when I was going in and out of functions. It seems ridiculous to me that I would have to do things "the bad way" in order to get my work done.
Amen to that... except the part about R. It and its bastard proprietary brother S+ can also go to the bitbucket as far as I am concerned.
I am sorry, did my bitterness just come through there?
Another useful alternative is to use SciPy and NumPy packages for Python. Python doesn't have any issues with dealing with long numericals.
I might be confusing what happened with CRU, but you are right... some of the data was from third-party sources and they wouldn't allow the release for everyone else to use. This was not desirable to the scientists involved, but their hands were bound in this case.
This is just another reason for why copyright laws need to change. Scientific data should be open for all to access. I should be able to create a webpage with links to my data and publish that along with any research papers.
http://journals.ametsoc.org/doi/abs/10.1175/2008BAMS2370.1
I've been asking people that for years. It is ironic if you are a weather forcaster.
No, believe me... it wasn't ironic... it was funny as hell!
Errrr, Pulseaudio doesn't replace ALSA
Yes, sorry for not being clear. PA replaces ALSA as the interface for applications. This will allow for easier (read less painful) changes in the future because there will be an abstraction layer. Userspace stuff is in userspace, kernel stuff is in the kernel (which is why I claim it is a superior design over just using ALSA).
Yes, OSS does have many of the features that PA promises, and so it would seem to be a more logical platform to move towards. But, wasn't there licensing issues surrounding it at one point?
Seems like Pulseaudio was not working properly with your sound card and you had to regress to ALSA. Did you file a bug report? I know several developers have been hard at work rewriting hardware drivers to be compatible with Pulseaudio. The more bug reports, the better as they seem to be churning out the changes based on the reports. If you don't want to commit to an install, you can easily try booting from the LiveCD first to test out the sound.
Fedora uses Pulseaudio. RedHat will too... Linux is migrating to Pulseaudio because it is a superior design over ALSA and OSS. The move has been painful, I admit, but that will happen whenever you mess around with fundamental OS components.
If you want flash and java working OOTB, then you need to head over to an unofficial distro like Linux Mint. The official Ubuntu releases can not come with flash installed by default (I can't remember if Java comes installed or not as I have been doing upgrades). As for pulseaudio, it is *much* better, but we are still encountering a few odds and ends with the "rarer" hardware (or undocumented hardware).
But if you cause me to take legal action you can bet if I win (and believe me the city was going to win) then I'm going to take my legal fees out of your pocket book.
Isn't that only possible in a civil case?
Ok, and each of those links do not support your main assertion of "can't reproduce the same results with their data". If anything, I am quite amazed by the amount of information given in the README file. I frequently find myself bashing my head against a wall trying to figure out why I can't process some piece of data correctly or something or another, but I don't keep a diary like this guy did.
One of the links was a rant about a squared error value going negative and saying that any programmer would have caught that and knew how to deal with it. Well, in the README file, the guy did notice it, recognized that he was grabbing error values instead of obs and fixed the data to properly mask them out as they should have been in the first place. So, what's the problem?
In the end, it shows that he persevered in being able to reproduce much of the original data despite changes in the grid layout, system upgrades and software changes.
I would also like to add to that list the statement from the American Meteorological Society: http://ametsoc.org/policy/climatechangeclarify.html
Whether it is true or false, posting such statements without attribution is just elitist and smug because you assume that everyone else in the conversation are aware of the same information that you have. It does no good to people in this forum to not provide sources for your statements.
http://journals.ametsoc.org/doi/abs/10.1175/2008BAMS2370.1
I'm not knowledgeable about all the hardware, but shouldn't it be possible to reset the graphics system somehow without a full reboot?
Yes, it is possible, but it requires accessing the terminal. Often in the frozen graphics mode, just about all of the keypresses aren't getting through, so you can't do Ctrl-Alt-BkSp or do a Ctrl-F1 to get to a terminal. I haven't tried the SysRq (or whatever it is) tricks, though...
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/461163
There are other reports for different hardware, and we even have a wiki page to help summarize the data. Feel free to help!