Um, yes, science is about predictive power. If you understand how a system works, then you can predict how it will react given different inputs. Some of evolution's proof comes from predictions of intermediate species that were unknown and then later discovered. IANABiologist, but I imagine that it's quite difficult to predict future evolution because it takes so long and the inputs can vary widely (how will the environment change over the next million years, what natural disasters will occur, human actions, etc). Though I wouldn't be surprised if there are predictions on various species evolution -- they will just take a long time to verify!
Fact: Senator Blowhard doesn't have to stand up and read from the phonebook anymore to infinitely prolong debate on a bill
My point is that he should have to do that. Anything less is really just a threat. The fact that by convention, the majority party lets the threat count as a real filibuster is the problem. A filibuster should be something you do when you really believe in something -- not a tactic to require 60 votes.
The Republicans didn't filibuster anything. They said they'd filibuster if they had to, and that was enough to keep Democrats from even bringing the bill up. Just the mere threat of a filibuster and the Democrats backed down. Why not call the Republicans bluff? Force them to get up and actually filibuster the 9/11 responder's bill. Hold a cloture vote every hour (or as often as rules allow), and throw up a website with each senator's name and how they voted. Widely publicize it. How long do you think the Republicans could really keep the filibuster up?
On the contrary, I got a note from the IRS this past year saying they were no longer going to send me paper forms and instructions (and more importantly, the envelope to mail the forms back in). The note suggested that instead I should use some expensive Windows only software to "e-file" my taxes. This saves the IRS money, but pushes the costs on to me!
I have always been anti-Microsoft, in this regard I was always pulling for Apple.
This is what I don't understand. Apple has always been this way, since their very beginning. There's even a term for it -- the Reality Distortion Field. It can take many forms, and denying users access to information about competitors is just one of them. Apple has always at heart been a much more about control than Microsoft ever was. It's the company culture.
If you want freedom, cheer on Linux (I almost said root for Linux -- ha!) There is where you'll find true user freedom, because freedom is at the heart of the OSS culture.
I suppose I'll be modded troll by the Apple fanboys. That's not my intention, but so be it.
Normally, I'd agree, but the OP specifically talked about a user supplied password to be able to add a plugin. That password could control access to a private key that is used to sign a hash of the valid list of plugins. On startup, Firefox could use the public key to validate the list of plugins, and throw up a big error if the list is invalid (because someone snuck one in).
Of course, recovering from this state would be difficult -- maybe Firefox could provide a way to disable plugins until the new list matched it's hash? But it would at least alert the user that something fishy was going on. Think of it as a tripwire for plugins.
I also don't buy the logic that no-DRM gets more customers
I currently only have computers running Linux at home. Because of the DRM, Netflix doesn't work on Linux. It doesn't "just work." Despite this, I have considered several times becoming a Netflix customer and just not being able to use the streaming option, but for just DVDs, Redbox seems to be a better option.
I'm probably in the (vast) minority, but the idea that DRM is stopping people from being customers isn't true.
Clever code? The Google guidelines forbid things like exceptions and dynamic_cast (and other RTTI). Are exceptions really that uncommon as to be labeled "clever"? They also discourage the use of inheritance (guess that's why you don't need RTTI) and even simple operator overloading, like operator==(). Really, operator==() is just the C++ name for equals(). They also discourage the use of operator=(), operator(), and streams in general.
I suppose their restrictions are internally consistent (no need for operator if you're not using streams, no need for RTTI if you have minimal inheritance, etc), but they turn C++ into "C with classes". It makes the code more understandable if you're a C programmer, but for a good C++ programmer, who understands things like templates, exceptions, and inheritance, a lot of those restrictions are very limiting. A good C++ programmer means that banning the use of exceptions just means that your code is probably not going to be exception safe, and will likely have problems when an exception is thrown through it (like a bad_alloc exception if you run out of memory). Unlike the GP, I think the things Google leaves out are incredibly important to writing good C++ code instead of just "C with classes."
Because C and C++ are enabling, not disabling, languages. And while 99% of the time it's a bad idea to do something (like write a custom allocator for a class), occasionally, it is the right thing to do, and you're glad to have that option.
Update to previous post: I downloaded the Google example images and compressed a few of them in JPEG2000. I limited it to 1 bpp (by default, my compressor uses as much as it wants, so without limiting it makes images bigger than the JPEGs). The results were comparable to the WebP images. About the same quality, about the same size. I don't know what settings Google use to generate their images. If their compressor automatically chooses a good bitrate/compression, then there's some advantage there (rather than with me fiddling JPEG2000 knobs to get a nice compression vs. quality mix).
Of course, as others have noted, simply re-compressing the source images with a higher JPEG compression gives a smaller size than either WebP or JPEG2000, without a noticeable quality loss. For example, using 'convert 1_original.jpg -quality 50 1_im.jpg' gives a 60% reduction in size and no visible quality loss (it was better than the JPEG2000 I did, and just a little bigger). Google's WebP version of that image only had a 10% reduction in size.
Also, it would be good to use an uncompressed image as the original source. Doing compression tests on an image that's already been compressed (JPEG) is somewhat bogus, since the JPEG has already introduced some artifacts.
What you're talking about is often called an "image pyramid" or "reduced resolution data sets". Basically, if you have a 2000x1500 full res image, you make an image that is half the size (1000x750). Then you take that image and make an image that is half that size (500x375), and so on until you're down to a very small size.
This process is actually inherent in the JPEG200 format. You get the reduced resolution images for free there. Based on how the J2K packets are laid out in the file, you can display a lower res version of the image after downloading only a small fraction of the file. JPEG2000 also has the concept of quality layers, which can further reduce how much data has to be read.
Unfortunately, both of those only really work well when you can seek to any position in the file to read the data you need to reconstruct the image. Many JPEG2000 images are tiled (both the EPJE and NPJE standards require 1024x1024 tiles) and each tile is essentially an independent JPEG2000 image. So readers end up reading little parts from all over the file for really big images. This is essentially why the JPIP streaming protocol was created -- the reader needs more ability to seek in the file than traditional HTTP allows. But for comparatively small images (like 16 MP) that could be encoded as one big tile, you could end up being able to save quite a bit of downloading, at least with an intelligent client.
So why isn't JPEG2000 more widely used? Well, for starters, it's insanely complicated. Lots and lots of features, many of them very nice (like a lossless compression option). But that means implementing a full codec that works efficiently isn't easy. There's no libj2k library that everyone can just use like there is for JPEG. There are libraries (Jasper, Kakadu, and others) but they aren't as widespread and some are costly (Kakadu). And plus, JPEG works well enough that most people don't bother.
It would be interesting to compare Google's new WebP format to JPEG2000. Maybe I'll download their source file zip and see what results I get.
You're thinking of what police would be like in an anarchist world. In a libertarian world, the government still exists and still provides essential services(*). There's no reason to assume that the police force has to be corrupt, or at least not any more corrupt in any larger government system. But really, the case in this article is just plain corruption, and could occur no matter how much or how little government there is.
(*) Of course, defining what an "essential service" for the government to provide is always tricky, and is complicated when most of the people deciding benefit from bigger government.
I was curious, so I compared searching for "open office" on Bing and Google. Google returns a list of sites that would be pretty much what anyone would expect, lots of links to openoffice.org sites, and some other related sites that all seem reputable. Bing is somewhat similar, except for one glaring exception. The third link (and for some reason, my eye was drawn to it, probably because the title was simply "OpenOffice") is to http://openoffice.org-suite.com/. Note that this is NOT a site associated with the real OOo, but a site that tries to make money off of OOo. Now, this is mostly legitimate (I'm sure the OOo license allows it), but the style of the site makes it look like they created OOo. Their download link is to a site called "preview.licenseaquisition.org". I didn't try it (only Windows binaries), but I can only imagine that it asks for money at some point. I know I've recommended OpenOffice.org to people in the past and they've come back complaining about how it costs money. I wonder if this was the site they stumbled upon.
Google's results do not include this site, at least not within the first few pages. I don't know why Bing would feature it so prominently.
Like the GP, I also use tree style tabs and TMP. The thing I like best about TMP is that it can change the text style on unread tabs. This is nice for a site like/. where I will read the summaries and open the ones I'm interested in in new tabs. Then I'll read the comments over time. I can easily see which tabs I've looked at and which I haven't.
There are lots of other nice things in TMP, but that's the big one for me.
Why do media sites always complain about Apple, then proceed to give them so much free advertising? Every time Apple releases a new product, the media falls over themselves to tell everyone about it. If they wanted to teach Apple a lesson, they'd impose their own ban on Apple news. No free product placement on the front page. No glowing "reviews" about how great the new iWhatever is and how you just have to have one.
Of course, that would imply intelligence in the media, and so far there's been very little evidence of that.
(that, and as others have noted, Gizmodo isn't exactly above reproach in many ways)
What software were you trying to use on Linux? It looks like ZoneMinder supports a number of network cameras, among other ways of connecting a camera to a computer. There are cameras that do stupid things, like require an ActiveX control to get the camera data, that ZM doesn't support, but it looks like most work OK.
I got a D-link DCS-920 recently, and it works fine with ZomeMinder. The biggest problem is that the picture quality is only OK and it doesn't have a very big viewing angle (probably because of the low resolution).
Agreed on a lot of that (I wonder if they ever fixed that auto-run security bug in their tricorders?), but depending on the targeting sensor, it may be possible to tell what the enemy vessel is targeting. For example, even today, our military can "paint" a target with a laser that relays information about the target to whatever weapons system is being used. If the TNG targeting system was a similar active sensor, that might allow the victim ship to be able to tell what is being targeted.
Of course, knowing this, a good weapons system designer would try to switch to a more passive sensor (like reflected light) to do targeting, so that the enemy doesn't know what you're looking at.
Diff the content of two tar.gz files: diff <(tar tvfz foo.tar.gz | sort) <(tar tvfz bar.tar.gz | sort)
Wow! I learned something new and useful on/. today! I had never seen the <() syntax, but I can appreciate how useful it can be, especially for tools like diff that tend to take files. For those wanting to learn more, this is called Process Substitution.
The top half -- the part at the top of the ropes -- seems to have a number of propellant containers. You can see them more clearly as it flies away after letting go of the rover on the ground. Keep in mind that the rover is car sized, and that top half looks bigger than the rover, so the tanks probably have a fair bit of propellant.
The animation makes it seem like it lets go of the parachute pretty early. I would think they could save propellant by drifting down on the parachute longer. But IANARS (rocket scientist) so I'm sure they have a good reason.
Slashdot Mirror
Um, yes, science is about predictive power. If you understand how a system works, then you can predict how it will react given different inputs. Some of evolution's proof comes from predictions of intermediate species that were unknown and then later discovered. IANABiologist, but I imagine that it's quite difficult to predict future evolution because it takes so long and the inputs can vary widely (how will the environment change over the next million years, what natural disasters will occur, human actions, etc). Though I wouldn't be surprised if there are predictions on various species evolution -- they will just take a long time to verify!
Fact: Senator Blowhard doesn't have to stand up and read from the phonebook anymore to infinitely prolong debate on a bill
My point is that he should have to do that. Anything less is really just a threat. The fact that by convention, the majority party lets the threat count as a real filibuster is the problem. A filibuster should be something you do when you really believe in something -- not a tactic to require 60 votes.
The Republicans didn't filibuster anything. They said they'd filibuster if they had to, and that was enough to keep Democrats from even bringing the bill up. Just the mere threat of a filibuster and the Democrats backed down. Why not call the Republicans bluff? Force them to get up and actually filibuster the 9/11 responder's bill. Hold a cloture vote every hour (or as often as rules allow), and throw up a website with each senator's name and how they voted. Widely publicize it. How long do you think the Republicans could really keep the filibuster up?
On the contrary, I got a note from the IRS this past year saying they were no longer going to send me paper forms and instructions (and more importantly, the envelope to mail the forms back in). The note suggested that instead I should use some expensive Windows only software to "e-file" my taxes. This saves the IRS money, but pushes the costs on to me!
I have always been anti-Microsoft, in this regard I was always pulling for Apple.
This is what I don't understand. Apple has always been this way, since their very beginning. There's even a term for it -- the Reality Distortion Field. It can take many forms, and denying users access to information about competitors is just one of them. Apple has always at heart been a much more about control than Microsoft ever was. It's the company culture.
If you want freedom, cheer on Linux (I almost said root for Linux -- ha!) There is where you'll find true user freedom, because freedom is at the heart of the OSS culture.
I suppose I'll be modded troll by the Apple fanboys. That's not my intention, but so be it.
Normally, I'd agree, but the OP specifically talked about a user supplied password to be able to add a plugin. That password could control access to a private key that is used to sign a hash of the valid list of plugins. On startup, Firefox could use the public key to validate the list of plugins, and throw up a big error if the list is invalid (because someone snuck one in).
Of course, recovering from this state would be difficult -- maybe Firefox could provide a way to disable plugins until the new list matched it's hash? But it would at least alert the user that something fishy was going on. Think of it as a tripwire for plugins.
You do not give up your Constitutional rights simply by choosing a more convenient option.
I also don't buy the logic that no-DRM gets more customers
I currently only have computers running Linux at home. Because of the DRM, Netflix doesn't work on Linux. It doesn't "just work." Despite this, I have considered several times becoming a Netflix customer and just not being able to use the streaming option, but for just DVDs, Redbox seems to be a better option.
I'm probably in the (vast) minority, but the idea that DRM is stopping people from being customers isn't true.
AFAIK, you cannot stream netflix if you run Linux, which is the biggest reason why I personally do not subscribe to their service.
At $1/movie, RedBox has provided me with a good source of recent movies.
Clever code? The Google guidelines forbid things like exceptions and dynamic_cast (and other RTTI). Are exceptions really that uncommon as to be labeled "clever"? They also discourage the use of inheritance (guess that's why you don't need RTTI) and even simple operator overloading, like operator==(). Really, operator==() is just the C++ name for equals(). They also discourage the use of operator=(), operator(), and streams in general.
I suppose their restrictions are internally consistent (no need for operator if you're not using streams, no need for RTTI if you have minimal inheritance, etc), but they turn C++ into "C with classes". It makes the code more understandable if you're a C programmer, but for a good C++ programmer, who understands things like templates, exceptions, and inheritance, a lot of those restrictions are very limiting. A good C++ programmer means that banning the use of exceptions just means that your code is probably not going to be exception safe, and will likely have problems when an exception is thrown through it (like a bad_alloc exception if you run out of memory). Unlike the GP, I think the things Google leaves out are incredibly important to writing good C++ code instead of just "C with classes."
Because C and C++ are enabling, not disabling, languages. And while 99% of the time it's a bad idea to do something (like write a custom allocator for a class), occasionally, it is the right thing to do, and you're glad to have that option.
Update to previous post: I downloaded the Google example images and compressed a few of them in JPEG2000. I limited it to 1 bpp (by default, my compressor uses as much as it wants, so without limiting it makes images bigger than the JPEGs). The results were comparable to the WebP images. About the same quality, about the same size. I don't know what settings Google use to generate their images. If their compressor automatically chooses a good bitrate/compression, then there's some advantage there (rather than with me fiddling JPEG2000 knobs to get a nice compression vs. quality mix).
Of course, as others have noted, simply re-compressing the source images with a higher JPEG compression gives a smaller size than either WebP or JPEG2000, without a noticeable quality loss. For example, using 'convert 1_original.jpg -quality 50 1_im.jpg' gives a 60% reduction in size and no visible quality loss (it was better than the JPEG2000 I did, and just a little bigger). Google's WebP version of that image only had a 10% reduction in size.
Also, it would be good to use an uncompressed image as the original source. Doing compression tests on an image that's already been compressed (JPEG) is somewhat bogus, since the JPEG has already introduced some artifacts.
What you're talking about is often called an "image pyramid" or "reduced resolution data sets". Basically, if you have a 2000x1500 full res image, you make an image that is half the size (1000x750). Then you take that image and make an image that is half that size (500x375), and so on until you're down to a very small size.
This process is actually inherent in the JPEG200 format. You get the reduced resolution images for free there. Based on how the J2K packets are laid out in the file, you can display a lower res version of the image after downloading only a small fraction of the file. JPEG2000 also has the concept of quality layers, which can further reduce how much data has to be read.
Unfortunately, both of those only really work well when you can seek to any position in the file to read the data you need to reconstruct the image. Many JPEG2000 images are tiled (both the EPJE and NPJE standards require 1024x1024 tiles) and each tile is essentially an independent JPEG2000 image. So readers end up reading little parts from all over the file for really big images. This is essentially why the JPIP streaming protocol was created -- the reader needs more ability to seek in the file than traditional HTTP allows. But for comparatively small images (like 16 MP) that could be encoded as one big tile, you could end up being able to save quite a bit of downloading, at least with an intelligent client.
So why isn't JPEG2000 more widely used? Well, for starters, it's insanely complicated. Lots and lots of features, many of them very nice (like a lossless compression option). But that means implementing a full codec that works efficiently isn't easy. There's no libj2k library that everyone can just use like there is for JPEG. There are libraries (Jasper, Kakadu, and others) but they aren't as widespread and some are costly (Kakadu). And plus, JPEG works well enough that most people don't bother.
It would be interesting to compare Google's new WebP format to JPEG2000. Maybe I'll download their source file zip and see what results I get.
You're thinking of what police would be like in an anarchist world. In a libertarian world, the government still exists and still provides essential services(*). There's no reason to assume that the police force has to be corrupt, or at least not any more corrupt in any larger government system. But really, the case in this article is just plain corruption, and could occur no matter how much or how little government there is.
(*) Of course, defining what an "essential service" for the government to provide is always tricky, and is complicated when most of the people deciding benefit from bigger government.
I was curious, so I compared searching for "open office" on Bing and Google. Google returns a list of sites that would be pretty much what anyone would expect, lots of links to openoffice.org sites, and some other related sites that all seem reputable. Bing is somewhat similar, except for one glaring exception. The third link (and for some reason, my eye was drawn to it, probably because the title was simply "OpenOffice") is to http://openoffice.org-suite.com/. Note that this is NOT a site associated with the real OOo, but a site that tries to make money off of OOo. Now, this is mostly legitimate (I'm sure the OOo license allows it), but the style of the site makes it look like they created OOo. Their download link is to a site called "preview.licenseaquisition.org". I didn't try it (only Windows binaries), but I can only imagine that it asks for money at some point. I know I've recommended OpenOffice.org to people in the past and they've come back complaining about how it costs money. I wonder if this was the site they stumbled upon.
Google's results do not include this site, at least not within the first few pages. I don't know why Bing would feature it so prominently.
Like the GP, I also use tree style tabs and TMP. The thing I like best about TMP is that it can change the text style on unread tabs. This is nice for a site like /. where I will read the summaries and open the ones I'm interested in in new tabs. Then I'll read the comments over time. I can easily see which tabs I've looked at and which I haven't.
There are lots of other nice things in TMP, but that's the big one for me.
Dude, if you're never going to read it, you should just unsubscribe from LKML.
/sys/ of course!
Why do media sites always complain about Apple, then proceed to give them so much free advertising? Every time Apple releases a new product, the media falls over themselves to tell everyone about it. If they wanted to teach Apple a lesson, they'd impose their own ban on Apple news. No free product placement on the front page. No glowing "reviews" about how great the new iWhatever is and how you just have to have one.
Of course, that would imply intelligence in the media, and so far there's been very little evidence of that.
(that, and as others have noted, Gizmodo isn't exactly above reproach in many ways)
What software were you trying to use on Linux? It looks like ZoneMinder supports a number of network cameras, among other ways of connecting a camera to a computer. There are cameras that do stupid things, like require an ActiveX control to get the camera data, that ZM doesn't support, but it looks like most work OK.
I got a D-link DCS-920 recently, and it works fine with ZomeMinder. The biggest problem is that the picture quality is only OK and it doesn't have a very big viewing angle (probably because of the low resolution).
Agreed on a lot of that (I wonder if they ever fixed that auto-run security bug in their tricorders?), but depending on the targeting sensor, it may be possible to tell what the enemy vessel is targeting. For example, even today, our military can "paint" a target with a laser that relays information about the target to whatever weapons system is being used. If the TNG targeting system was a similar active sensor, that might allow the victim ship to be able to tell what is being targeted.
Of course, knowing this, a good weapons system designer would try to switch to a more passive sensor (like reflected light) to do targeting, so that the enemy doesn't know what you're looking at.
Wow! I learned something new and useful on /. today! I had never seen the <() syntax, but I can appreciate how useful it can be, especially for tools like diff that tend to take files. For those wanting to learn more, this is called Process Substitution.
Thanks for this great tip!
You mean because -x tests to see if the file is executable, not if it exists? Or the fact that you're saying the file exists if it is NOT executable?
The easiest way to execute multiple statements if an expression is true is to do something like:
if [ -f "$f" ]; then echo "file $f exists"; continue; fi
It is true that more complicated tasks often require more complicated tools. But for relatively simple tasks, shell scripts are great!
Project Gutenberg also has it: http://www.gutenberg.org/etext/201
The top half -- the part at the top of the ropes -- seems to have a number of propellant containers. You can see them more clearly as it flies away after letting go of the rover on the ground. Keep in mind that the rover is car sized, and that top half looks bigger than the rover, so the tanks probably have a fair bit of propellant.
The animation makes it seem like it lets go of the parachute pretty early. I would think they could save propellant by drifting down on the parachute longer. But IANARS (rocket scientist) so I'm sure they have a good reason.