It's not the RIAA whose business model is dead, they are in their heyday right now, with all these lawsuits and legal fees.
It's the recording industry whose business model is dead. They're different entities, and even different industries, the RIAA is hired by members of the recording industry, and the RIAA is having the most profitable time of their existence.
Seriously, if it's original to Wikipedia, they could have just slapped a "No Original Content" sticker on the article, and it would have been deleted within a week. Or just complain that it's not notable. Nothing is considered notable on Wikipedia any more, so that would have gotten it deleted.
In a criminal case, they would have to show that the defendant was never in any danger of getting a guilty verdict. I'm not sure any such case law actually exists. The concept is that of protection from double jeopardy. If the defendant is in any jeopardy of being found guilty, then jeopardy has "attached," and they cannot be tried a second time. It springs up from hundreds of years ago in England, where they would repeatedly try a person for the same crime until they got the outcome they wanted, either through an exhausted legal system, or through an exhausted defense budget.
Prosecution will request a mistrial before they'd allow this to happen since the burden of proof that there was no jeopardy is so very high. I don't think any such case law exists in the US. You might be able to manage it if the judge nullified the jury's guilty verdict and entered a guilty one, and it was later discovered the judge had been bribed or otherwise coerced into doing so.
If the computer isn't connected to the net (and they aren't able to load inappropriate stuff their friend gave them on a thumb drive), then I don't need access to it. Likewise with a journal. No one ever got kidnapped, raped, and murdered by someone they met by writing in a private journal, and material which the child isn't emotionally and developmentally ready for never spontaneously appeared in it.
Plug it in to the net, or notice little Bobby or Susy loading up stuff on it that you don't recognize from friends, then you bet it's time to want to know what's going on. Kids aren't adults, they don't get the same level of privacy from their parents that adults do, nor should they.
Parents need involvement in their kids lives, it's the way that they shape and mold their kids into functional balanced adults, as well as protect them from dangers the kid doesn't realize exist or doesn't believe in. It's the mark of a good parent, and it's something that's lacking in too many parents.
There are digital watermarks for images already on the market (and have been for at least 10 years). You can apply the watermark to differing strengths, the stronger the watermark, the harder you have to work to remove it, but the more the image is affected. In this case they're looking for image verification, so it would only have to be strong enough for the image to show obvious artifacting from the watermark removal process.
Yes, watermarks reduce image quality, but in general you have to double the quality loss to strip the watermark. You can even crop out a tiny portion of the middle of the image and it will still have its watermark.
I'm guessing they're smart enough to hash the ident data - it can be verified but not reversed.
Re:MOD PARENT UP -- golden rule vulnerability patc
on
Ethics In IT
·
· Score: 1
So your rule is allowed caveats but caveats on the traditional one makes it worse?
I think it's fair to say that it's not possible to deterministically solve ethics in a single sentence. That doesn't make sanity checks like the Golden Rule worthless.
Re:MOD PARENT UP -- golden rule vulnerability patc
on
Ethics In IT
·
· Score: 1
I want others to treat me such that they give me thousands of dollars.
The problem with both systems is that it does not allow for balancing the needs and wants of both parties.
The Christian-based "Do unto others as you'd have done to yourself" rule is closer, and people mis-interpret it by attaching conditionals. For example the original example, "I'd want him to kill me if I were gay." The rule does not say, "As you'd have done to yourself if you were in their shoes." The rule says that if you want to receive that treatment, then so should you persist that treatment, and doesn't mention circumstance at all.
However, it's meant as a general guide, sort of a litmus test of how to behave toward others. It attempts to put behavior into personal perspective. The Golden Rule also fails for people who are suicidal or otherwise self-destructive (just because you cut yourself doesn't mean you should take a knife to strangers). It's not meant to be hard and fast.
Re:You need to clarify your question
on
Ethics In IT
·
· Score: 1
The purpose of a company isn't to provide a steady living for people, but to maximize profits.
This is true for many or most companies. It is not true for all companies.
Anyone with a trusted SSH server they can get access to can do this to a certain extent with nothing more than an SSH client on the computer they're using at the public access point.
ssh -D 1080 [<someuser>@]<somehost>
Now set your programs to use a SOCKS proxy of localhost:1080, and you're done. My macbook has Little Snitch, and I have it configured to deny all outbound traffic which isn't going to one of my trusted SSH servers, and only on port 22. This way I don't even have to worry about accidentally forgetting to set a SOCKS proxy, I just have to always establish my SSH connection whenever I want to get online (explicit connectivity is also a good thing when you regularly work from untrusted networks).
Modern smallpox vaccination never to my knowledge had any outrage. In fact, you are not even vaccinated with smallpox or any derivative thereof. You are vaccinated with cowpox, which is similar to smallpox, but much less severe - many milkmaids would get it as sores on their hands when they first started milking, or first encountered an infected cow. Cowpox and smallpox are similar enough that smallpox triggers the body's immunoresponse once it has adapted to cowpox.
This is not a genetically modified organism, it is a naturally occurring one with many known infections and very few known deaths (around 1 in 1 million) and indeed its use was first taken up long enough ago that people would not yet have known to be afraid of it other than to be worried that it would give them smallpox (it was being studied for this purpose as early as the 1770's, and in wide use by the early 1800's).
I for one think that we don't yet have enough information on mutation rate from cloned animals to know whether they pose a risk when eaten (there are no multi-generational studies to this effect because cloning has not been around long enough for this). I don't think it's unreasonable to require food which may contain cloned meat to be labeled as such, and those who don't care that it's cloned can consume it on the cheap, and those who do can pay a little extra for non-cloned food.
I'll agree that modern displays are sufficient. But we were not discussing whether modern displays are fine, we were discussing discernibility from real scenes.
Our eyes have peak sensitivity on red, green, and blue, but those receptors each have sensitivity stretching across most of the range of visible light. We can tell the difference between true cyan and a 50/50 blue/green mix. True cyan is a pleasant color, while blue/green cyan is much less so. Compare a Pantone Cyan color chip against 255 blue 255 green (#00FFFF), or some darker shade thereof, and you will notice a difference. This isn't a matter of color correction on the monitor, it's a matter of our eyes' ability to distinguish these different colors. We can get pretty close by catering to the peak sensitivity of the eyes, but it's not indistinguishable from real.
Dynamic range is likewise sufficient for display purposes; in fact a realistic dynamic range would make a bad display as sunlight in the display would cast too much light into the room (let alone requiring a large amount of power). Displays can simulate the dynamic range adjustments our eyes do on their own via irises and mental alterations (download the HalfLife 2 "The Lost Cost" which is free on Steam for an example of this in action, try it with and without the HDR lighting option), but it again is still very distinguishable from real since only the display dims when looking at a very bright object and wanting to preserve detail in it, not the whole room.
It depends on viewing distance. But resolution is not the most significant contributing factor in the unrealism arena for modern displays. Color accuracy, dynamic range, and two dimensions are the most significant contributors to this.
Modern displays have decent color, but as long as we are limiting ourselves to red / green / blue color sources, this will always be distinguishable from real (real cyan is not a 50/50 mixture of blue and green light, it is a single wavelength between the two).
Dynamic range is how bright the brightest areas are and how dark the darkest areas are. Real dynamic range would show the sun so bright on the screen that it hurt your eyes to look at, with dark areas so dark that your eyes would not be able to see the detail even if the screen had no bright areas (but that detail would still be present).
And of course no matter what, as long as it is projected onto a single flat surface, it is 2D, and it will always look like a movie.
As for pixel count, many people are already unable to distinguish between standard definition and high definition when viewing at a standard viewing distance (of course different screen sizes have a different standard viewing distance). Of course in the TV store you can't get 8-12 feet away from the screen so when you're shopping, the differences seem obvious.
Thanks for the link. I have used that for a while. It tries to bring in some of the features from Web Developer Toolbar and Firebug both. You're right though, it's a poor replacement for either. Sometimes I need to use it, but man when I do, I really miss the extra features from those other tools.
From a user experience, I believe vanilla IE is probably about the same as vanilla Firefox. It's once you start getting into some of the many great extensions for Firefox, which IE has a handful of pitiful cones for that the experience diversifies quickly.
From a developer experience, Firefox is hands down better. It is simply easier to get pages working right the first time when you're trying to break away from 1999 era table layouts and into a modern CSS/CSS2 or XML/XSL page styling. Firefox gives meaningful javascript errors (vs IE's miserable "Object does not support this property or method, line 1234" [which is often not a line number which agrees with the script line which generates the error, even for included scripts]).
Oh so often, we create pages off the bat which look great in Firefox, Opera, Safari, and even Konqueror, fire up IE, and it looks horrible. An hour of fiddling with IE to make it work right later, and now it looks wrong in all the other browsers. Eventually of course we get it to work, but not before we took elegant, short, simple code and bastardized it to pieces. But the user's experience is the same, so the user doesn't know which browser is actually better, because we can't afford to ignore the majority browser even if it is crap.
Also, development in Firefox with these two extensions has made my life immeasurably easier: Web Developer Toolbar by Chris Pedrick, and Firebug (in which I am always managing to discover new features that have been there all along). I shudder and groan when it's time to make it work in IE, because I know that javascript debugging, or CSS debugging, or just "Operation Aborted" errors are going to keep me unproductively busy for hours.
1) electronic security (to add in additional votes), 2) physical security (to add in paper votes, with appropriate timestamps, in the appropriate location within the stack; you can't do 500 within a 2 minute window or they are identifiable as false, and you can't have the time stamps on the ballots be out of sequence or again they are identifiable as false) 3) voter counts (you can't have more votes counted than the count of people who voted)
All three systems have completely different requirements for defeat, voter counts aren't even handled by the electronic systems at all, and the machine can't self-stuff because it doesn't have the ballot paper (this is why you give voters the ballot card which they feed to the machine).
You can't be intimidated to hand over your number, because you have plausible deniability that you even wrote down your number (this is why voters have to choose to write down their own number instead of everyone being given a slip or stub). If you were being especially coerced before hand (You better make sure you get your vote number!), you could even choose to accidentally transpose some of the digits, etc. Vote coercion is illegal, so if this is happening, you have further recourse available to you.
Vote selling happens now even without voter-verifiable audit trails. Further after several rounds of appeals, it was deemed legal. Besides, voter self-verification is such an important security tool which eliminates the ability to both modify or delete votes, that the fraction of people who would bother to sell their votes is insignificant compared to the damage which can be done if this type of control is not in place. Further, if you promised to vote a certain way, you could look at the vote outcomes, find a different voter who did vote the way you said you did, and give over that number instead, while still voting the way you wanted to (of course this depends on timely publishing of the vote results, but since it's all electronic, there's no reason these votes couldn't be online minutes after the polls close, or even updated throughout the day).
Here is an ideal setup: + The voting machine has no internal audit trail paper. + The voter is given a ballot paper which they feed to the voting machine. + The voter makes their selections on a screen or other input device. + Upon confirming their choice, the voter feeds the paper to the machine. + The machine prints the user's choices in a format simultaneously readable by both humans and computers (eg, the numbers on checks), a vote serial number (not necessarily incremental, but unique), and a time and date on the paper, then displays it to the voter through a piece of glass. + The user verifies the details are correct. + If not, they may choose to cancel the vote with a big red rejection button (this feeds the ballot back through the printer which prints VOID on it and returns it to the user, as well as removing that vote from the totals). + If the vote on the paper is correct, the user may accept the vote with a big green acceptance button, which feeds the ballot into a locked ballot box.
Recounts can be accomplished quickly and accurately by machine-reading the ballots, and spot checking the machine accuracy by hand (feed it a stack of ballots, then hand count that stack to make sure they agree).
The machine keeps a complete log of the voting outcome for each serial number. This log can be put on public websites (in randomized order so you don't know how the guy in front of you voted), and users who chose to write down their vote's serial number can access the website to verify their vote was counted as they recalled.
You can't silently discard votes from the electronic record because there's both a physical ballot you have to discard, and a serial number which the voter may choose to verify later. You can't alter votes because you have the user-verified physical confirmation, and the ability of random users to be able to verify that their vote was recorded as expected. You can't stuff ballots because you would have to spread the timestamps on the physical ballot out, and insert them into the ballot stack at the right locations. You'd also have to alter the electronic record - which means you'd have to compromise both the electronic and the physical security, and take the time to add entries in the correct locations in the physical ballot stack. You can have entirely different organizations or groups of people handle the electronic and the physical side of things, with no crossover, so that any vote tampering requires a conspiracy and a lot of luck (where the conspirators have to be given access to modify the same vote stack as each other). Finally ballot stuffing is impossible without the ability to discard ballots because the number of votes cast and the number of votes counted will be different.
You can also further mitigate ballot stuffing by printing different ballot sheets for each district, and not deciding which district gets which sheet type until very very late in the game (eg, the morning of the vote).
There are basically three forms of vote fraud. Vote stuffing, vote altering, and vote discarding (add, change, and delete). This adds checks and balances to each of these forms of fraud.
Even if not every voter bothers to verify his or her own paper trail, the fact is that some will, so rejecting or altering votes is too likely to be discovered. We now have both anonymity (to protect against coercion), verifiability (to protect against change and delete), and separate physical and electronic access, vote ordering, and vote totals (to protect against add).
As a worst case scenario, should all else fail, this would break down to machine-assisted paper ballots with some additional checks and balances over what we have historically had for paper ballots.
Mac OS 9 and under used to do this for both writes and reads. But sometimes you'd end up with two programs wanting different disks in the same drive. You ended up with the OS automatically ejecting disk A and asking for disk B. When you put in disk B, it'd automatically eject and ask for disk A. Because OS 9 was shared and not truly multi-process, when this happened you basically had to cut your losses and reboot.
Of course in a multi-processing OS, when you put in the requested disk A, it could do the work it needs to do for disk A while the dialog for disk B is still on the screen.
I'm certainly not a Linux filesystem guru, but I believe the way that ext3 and other journaling filesystems do their writes (even updates in the middle of a file) is that if you change 1 byte in the middle of an inode (or the entire inode) (file data is stored in chunks of a fixed size called inodes; same idea as Windows block size option; and this size is predetermined at time of formatting and stored as part of the filesystem headers... w/ 4096 byte inode size, a 1 byte file consumes as much disk space as a 4096 byte file... exactly 1 inode), it copies the entire chunk of data in that inode, plus your modifications, to a free inode, then updates the node map for that file, and finally marks the old inode as unused (or more accurately decrements its use counter, which accounts for times when a single inode is used by multiple files (hard links)).
Actually that's what ext2 does. Ext3, being journaled, creates the new inode, then creates a journal of what it is about to do (a pointer to the file in question, and the number of the old and the new inode, and the expected use counts for each), before updating the inode and usage counts and finally removing the journal. If the process gets interrupted during the writing of the new inode, the whole write fails (with the old file unmodified), and the new inode is still marked as unused. If the process gets interrupted during the journal write, the whole write fails (again with the old file unmodified). Once the journal has successfully been written, it can be guaranteed to succeed, any remaining work can be recovered. On startup of an uncleanly unmounted volume, the OS looks for complete (not completed) journal entries (if they exist, they are not yet complete), and does the work of the journal write again (even if the journal entry had completely succeeded, but had been interrupted before it was cleaned up, re-writing that data doesn't hurt anything).
So any interruption up to the point that the journal was completely written results in the entire write failing, and no loss of original data. Any interruption at that point or past results in the entire write succeeding (being finished once the OS mounts the partition).
From what I understand, ext3 actually is ext2, with the journal being handled as hidden files by the ext2 subsystem.
Now like I said, take all that with a grain of salt, it's mostly cobbled together understanding from incidental exposure to these things over time.
The success of his project is dependent on getting enough orders to give them economy of scale savings. He is courting the wealthier countries right now, and as he gets more orders, he can reduce cost further, and reach a more and more impoverished market. If Intel uses this as a way to have OLPC do all the front work of identifying markets, getting in to talk with the people who can make the buying decisions, and then end-gaming them on the sale, they are not only hurting the success of the OLPC project and shortening its eventual reach, they are also directly stealing value from the project which did the initial effort investment. Some of these countries might not have talked to Intel directly, but had heard good things about this OLPC thing and so first entertained the idea only because of the merits brought by that project. Intel is therefore operating under the guise of friendship and help, while cannibalizing the underlying foundation of the project.
OLPC laptops are more open, more free, better designed, and less expensive. Intel ClassMate PC's are proprietary, less rugged, and require more power to operate. Worst of all they are for-profit, and those profits are sent to Intel stock holders, making wealthy business men wealthier at the expense of money which would better be used satisfying an educational need in the exact same arena as the laptop was advertised as intending to assist. They unnecessarily drain valuable resources from the very market they are pretending to aid.
Basically this is about as disgustingly slimy as I think they are able to be.
The zany thing is that you don't need an actual bomb. All you have to do is communicate that you're holding a walkie-talkie whose other end is wired to a bomb in the cargo hold, push the button and BOOM.
If you're intending to hijack the plane, then you don't need an actual bomb, only the plausible and imminent threat of a bomb. Box cutters, people may charge you and expect to get cut a few times; maybe one guy dies. A bomb which can be activated with the push of a button, may easily go off before the button is pushed, taking out everyone in the plane.
Of course if they call your bluff, you have no recourse.
And X, Y, and Z turn around and sue B for misrepresenting the image to begin with. But like you said, this problem is not a Creative Commons problem, it's problem endemic to copyright which Creative Commons inherits because it is a license under copyright law.
I agree with you on that. The only time I ever use a WYSIWYG editor for HTML is when I have a large amount of copy to do a lot of formatting on (eg, lots of lists, bullets, bolds, italics, paragraphs, etc). Then I'm just using it to create extremely simple output (H's, P's, ol, ul, li, b, i, u, etc).
Likewise in Flex, I stick to hand-coding the MXML. I'm convinced this is actually faster than using the GUI, especially with the tag insight and completion. Plus I know the structure exactly. However, the GUI does a pretty decent job of generating clean code anyway. I still prefer hand coding.
Bluetooth. Works great without any external receiver if you already have bluetooth in your case.
Also you're considering the keyboard/mouse and receiver as separate products. It's a single product, wireless device and wireless receiver.
There is no standard other than bluetooth on wireless mouse/keyboard communications. They make a non-bluetooth version because not everyone has bluetooth or wants to use a bluetooth mouse (there's a small but noticeable latency which can be frustrating when doing fine mouse manipulations or gaming).
Canon & Nikon lenses don't support standard mounts because standard mounts are incompatible with their existing lenses. Modern SLR Nikons will still work just great with all Nikon lenses going back to 1985 (and a few before) (with the exception of the D40 & D40x which have a small body and some of those lenses rears extend too far into the body to allow the mirror to function). They don't support a standard mount because it means either locking out existing lenses (which means THOUSANDS of dollars for many people, it'd be about $3,000 for me) or getting an adapter (which will cost you 1-2 stops of light [yikes!] and increasing the focal length [not always a bad thing, but they make devices called extension tubes for when you want to do this]). There is still modern demand for Nikon lenses produced in the 80's and 90's, because some of those lenses were just really well engineered and manufactured. This is one of the reasons some people shoot exclusively Nikon; they want to know their lenses will not get obviated when they spend $2-3,000 on a single lens (I don't have one like that, but I have used one, and it is a wonderful creation).
You can get adapters to put just about any manufacturer's lens on just about any other manufacturer's body. Some times you lose some of the features such as auto focus (for example, if the lens depends on there being a drive in the camera, and the camera lacks a drive, there's nothing you can do!)
If you mean Flex (the product with which Silverlight is competing), you can use a standard text editor to create MXML files also. Of course the SDK is free and comes with an Eclipse plugin which supports both code insight and a GUI layout creator, so I'm not sure why you'd use a text editor when for no more cost you can have a full IDE.
Slashdot Mirror
It's not the RIAA whose business model is dead, they are in their heyday right now, with all these lawsuits and legal fees.
It's the recording industry whose business model is dead. They're different entities, and even different industries, the RIAA is hired by members of the recording industry, and the RIAA is having the most profitable time of their existence.
Seriously, if it's original to Wikipedia, they could have just slapped a "No Original Content" sticker on the article, and it would have been deleted within a week. Or just complain that it's not notable. Nothing is considered notable on Wikipedia any more, so that would have gotten it deleted.
In a criminal case, they would have to show that the defendant was never in any danger of getting a guilty verdict. I'm not sure any such case law actually exists. The concept is that of protection from double jeopardy. If the defendant is in any jeopardy of being found guilty, then jeopardy has "attached," and they cannot be tried a second time. It springs up from hundreds of years ago in England, where they would repeatedly try a person for the same crime until they got the outcome they wanted, either through an exhausted legal system, or through an exhausted defense budget.
Prosecution will request a mistrial before they'd allow this to happen since the burden of proof that there was no jeopardy is so very high. I don't think any such case law exists in the US. You might be able to manage it if the judge nullified the jury's guilty verdict and entered a guilty one, and it was later discovered the judge had been bribed or otherwise coerced into doing so.
Civil law is a different matter.
If the computer isn't connected to the net (and they aren't able to load inappropriate stuff their friend gave them on a thumb drive), then I don't need access to it. Likewise with a journal. No one ever got kidnapped, raped, and murdered by someone they met by writing in a private journal, and material which the child isn't emotionally and developmentally ready for never spontaneously appeared in it.
Plug it in to the net, or notice little Bobby or Susy loading up stuff on it that you don't recognize from friends, then you bet it's time to want to know what's going on. Kids aren't adults, they don't get the same level of privacy from their parents that adults do, nor should they.
Parents need involvement in their kids lives, it's the way that they shape and mold their kids into functional balanced adults, as well as protect them from dangers the kid doesn't realize exist or doesn't believe in. It's the mark of a good parent, and it's something that's lacking in too many parents.
There are digital watermarks for images already on the market (and have been for at least 10 years). You can apply the watermark to differing strengths, the stronger the watermark, the harder you have to work to remove it, but the more the image is affected. In this case they're looking for image verification, so it would only have to be strong enough for the image to show obvious artifacting from the watermark removal process.
Yes, watermarks reduce image quality, but in general you have to double the quality loss to strip the watermark. You can even crop out a tiny portion of the middle of the image and it will still have its watermark.
I'm guessing they're smart enough to hash the ident data - it can be verified but not reversed.
So your rule is allowed caveats but caveats on the traditional one makes it worse?
I think it's fair to say that it's not possible to deterministically solve ethics in a single sentence. That doesn't make sanity checks like the Golden Rule worthless.
I want others to treat me such that they give me thousands of dollars.
The problem with both systems is that it does not allow for balancing the needs and wants of both parties.
The Christian-based "Do unto others as you'd have done to yourself" rule is closer, and people mis-interpret it by attaching conditionals. For example the original example, "I'd want him to kill me if I were gay." The rule does not say, "As you'd have done to yourself if you were in their shoes." The rule says that if you want to receive that treatment, then so should you persist that treatment, and doesn't mention circumstance at all.
However, it's meant as a general guide, sort of a litmus test of how to behave toward others. It attempts to put behavior into personal perspective. The Golden Rule also fails for people who are suicidal or otherwise self-destructive (just because you cut yourself doesn't mean you should take a knife to strangers). It's not meant to be hard and fast.
Anyone with a trusted SSH server they can get access to can do this to a certain extent with nothing more than an SSH client on the computer they're using at the public access point.
ssh -D 1080 [<someuser>@]<somehost>
Now set your programs to use a SOCKS proxy of localhost:1080, and you're done. My macbook has Little Snitch, and I have it configured to deny all outbound traffic which isn't going to one of my trusted SSH servers, and only on port 22. This way I don't even have to worry about accidentally forgetting to set a SOCKS proxy, I just have to always establish my SSH connection whenever I want to get online (explicit connectivity is also a good thing when you regularly work from untrusted networks).
Modern smallpox vaccination never to my knowledge had any outrage. In fact, you are not even vaccinated with smallpox or any derivative thereof. You are vaccinated with cowpox, which is similar to smallpox, but much less severe - many milkmaids would get it as sores on their hands when they first started milking, or first encountered an infected cow. Cowpox and smallpox are similar enough that smallpox triggers the body's immunoresponse once it has adapted to cowpox.
This is not a genetically modified organism, it is a naturally occurring one with many known infections and very few known deaths (around 1 in 1 million) and indeed its use was first taken up long enough ago that people would not yet have known to be afraid of it other than to be worried that it would give them smallpox (it was being studied for this purpose as early as the 1770's, and in wide use by the early 1800's).
I for one think that we don't yet have enough information on mutation rate from cloned animals to know whether they pose a risk when eaten (there are no multi-generational studies to this effect because cloning has not been around long enough for this). I don't think it's unreasonable to require food which may contain cloned meat to be labeled as such, and those who don't care that it's cloned can consume it on the cheap, and those who do can pay a little extra for non-cloned food.
I'll agree that modern displays are sufficient. But we were not discussing whether modern displays are fine, we were discussing discernibility from real scenes.
Our eyes have peak sensitivity on red, green, and blue, but those receptors each have sensitivity stretching across most of the range of visible light. We can tell the difference between true cyan and a 50/50 blue/green mix. True cyan is a pleasant color, while blue/green cyan is much less so. Compare a Pantone Cyan color chip against 255 blue 255 green (#00FFFF), or some darker shade thereof, and you will notice a difference. This isn't a matter of color correction on the monitor, it's a matter of our eyes' ability to distinguish these different colors. We can get pretty close by catering to the peak sensitivity of the eyes, but it's not indistinguishable from real.
Dynamic range is likewise sufficient for display purposes; in fact a realistic dynamic range would make a bad display as sunlight in the display would cast too much light into the room (let alone requiring a large amount of power). Displays can simulate the dynamic range adjustments our eyes do on their own via irises and mental alterations (download the HalfLife 2 "The Lost Cost" which is free on Steam for an example of this in action, try it with and without the HDR lighting option), but it again is still very distinguishable from real since only the display dims when looking at a very bright object and wanting to preserve detail in it, not the whole room.
It depends on viewing distance. But resolution is not the most significant contributing factor in the unrealism arena for modern displays. Color accuracy, dynamic range, and two dimensions are the most significant contributors to this.
Modern displays have decent color, but as long as we are limiting ourselves to red / green / blue color sources, this will always be distinguishable from real (real cyan is not a 50/50 mixture of blue and green light, it is a single wavelength between the two).
Dynamic range is how bright the brightest areas are and how dark the darkest areas are. Real dynamic range would show the sun so bright on the screen that it hurt your eyes to look at, with dark areas so dark that your eyes would not be able to see the detail even if the screen had no bright areas (but that detail would still be present).
And of course no matter what, as long as it is projected onto a single flat surface, it is 2D, and it will always look like a movie.
As for pixel count, many people are already unable to distinguish between standard definition and high definition when viewing at a standard viewing distance (of course different screen sizes have a different standard viewing distance). Of course in the TV store you can't get 8-12 feet away from the screen so when you're shopping, the differences seem obvious.
Thanks for the link. I have used that for a while. It tries to bring in some of the features from Web Developer Toolbar and Firebug both. You're right though, it's a poor replacement for either. Sometimes I need to use it, but man when I do, I really miss the extra features from those other tools.
From a user experience, I believe vanilla IE is probably about the same as vanilla Firefox. It's once you start getting into some of the many great extensions for Firefox, which IE has a handful of pitiful cones for that the experience diversifies quickly.
From a developer experience, Firefox is hands down better. It is simply easier to get pages working right the first time when you're trying to break away from 1999 era table layouts and into a modern CSS/CSS2 or XML/XSL page styling. Firefox gives meaningful javascript errors (vs IE's miserable "Object does not support this property or method, line 1234" [which is often not a line number which agrees with the script line which generates the error, even for included scripts]).
Oh so often, we create pages off the bat which look great in Firefox, Opera, Safari, and even Konqueror, fire up IE, and it looks horrible. An hour of fiddling with IE to make it work right later, and now it looks wrong in all the other browsers. Eventually of course we get it to work, but not before we took elegant, short, simple code and bastardized it to pieces. But the user's experience is the same, so the user doesn't know which browser is actually better, because we can't afford to ignore the majority browser even if it is crap.
Also, development in Firefox with these two extensions has made my life immeasurably easier: Web Developer Toolbar by Chris Pedrick, and Firebug (in which I am always managing to discover new features that have been there all along). I shudder and groan when it's time to make it work in IE, because I know that javascript debugging, or CSS debugging, or just "Operation Aborted" errors are going to keep me unproductively busy for hours.
Stuffing requires defeating three protections:
1) electronic security (to add in additional votes),
2) physical security (to add in paper votes, with appropriate timestamps, in the appropriate location within the stack; you can't do 500 within a 2 minute window or they are identifiable as false, and you can't have the time stamps on the ballots be out of sequence or again they are identifiable as false)
3) voter counts (you can't have more votes counted than the count of people who voted)
All three systems have completely different requirements for defeat, voter counts aren't even handled by the electronic systems at all, and the machine can't self-stuff because it doesn't have the ballot paper (this is why you give voters the ballot card which they feed to the machine).
You can't be intimidated to hand over your number, because you have plausible deniability that you even wrote down your number (this is why voters have to choose to write down their own number instead of everyone being given a slip or stub). If you were being especially coerced before hand (You better make sure you get your vote number!), you could even choose to accidentally transpose some of the digits, etc. Vote coercion is illegal, so if this is happening, you have further recourse available to you.
Vote selling happens now even without voter-verifiable audit trails. Further after several rounds of appeals, it was deemed legal. Besides, voter self-verification is such an important security tool which eliminates the ability to both modify or delete votes, that the fraction of people who would bother to sell their votes is insignificant compared to the damage which can be done if this type of control is not in place. Further, if you promised to vote a certain way, you could look at the vote outcomes, find a different voter who did vote the way you said you did, and give over that number instead, while still voting the way you wanted to (of course this depends on timely publishing of the vote results, but since it's all electronic, there's no reason these votes couldn't be online minutes after the polls close, or even updated throughout the day).
There are good ways to combat these problems.
Here is an ideal setup:
+ The voting machine has no internal audit trail paper.
+ The voter is given a ballot paper which they feed to the voting machine.
+ The voter makes their selections on a screen or other input device.
+ Upon confirming their choice, the voter feeds the paper to the machine.
+ The machine prints the user's choices in a format simultaneously readable by both humans and computers (eg, the numbers on checks), a vote serial number (not necessarily incremental, but unique), and a time and date on the paper, then displays it to the voter through a piece of glass.
+ The user verifies the details are correct.
+ If not, they may choose to cancel the vote with a big red rejection button (this feeds the ballot back through the printer which prints VOID on it and returns it to the user, as well as removing that vote from the totals).
+ If the vote on the paper is correct, the user may accept the vote with a big green acceptance button, which feeds the ballot into a locked ballot box.
Recounts can be accomplished quickly and accurately by machine-reading the ballots, and spot checking the machine accuracy by hand (feed it a stack of ballots, then hand count that stack to make sure they agree).
The machine keeps a complete log of the voting outcome for each serial number. This log can be put on public websites (in randomized order so you don't know how the guy in front of you voted), and users who chose to write down their vote's serial number can access the website to verify their vote was counted as they recalled.
You can't silently discard votes from the electronic record because there's both a physical ballot you have to discard, and a serial number which the voter may choose to verify later. You can't alter votes because you have the user-verified physical confirmation, and the ability of random users to be able to verify that their vote was recorded as expected. You can't stuff ballots because you would have to spread the timestamps on the physical ballot out, and insert them into the ballot stack at the right locations. You'd also have to alter the electronic record - which means you'd have to compromise both the electronic and the physical security, and take the time to add entries in the correct locations in the physical ballot stack. You can have entirely different organizations or groups of people handle the electronic and the physical side of things, with no crossover, so that any vote tampering requires a conspiracy and a lot of luck (where the conspirators have to be given access to modify the same vote stack as each other). Finally ballot stuffing is impossible without the ability to discard ballots because the number of votes cast and the number of votes counted will be different.
You can also further mitigate ballot stuffing by printing different ballot sheets for each district, and not deciding which district gets which sheet type until very very late in the game (eg, the morning of the vote).
There are basically three forms of vote fraud. Vote stuffing, vote altering, and vote discarding (add, change, and delete). This adds checks and balances to each of these forms of fraud.
Even if not every voter bothers to verify his or her own paper trail, the fact is that some will, so rejecting or altering votes is too likely to be discovered. We now have both anonymity (to protect against coercion), verifiability (to protect against change and delete), and separate physical and electronic access, vote ordering, and vote totals (to protect against add).
As a worst case scenario, should all else fail, this would break down to machine-assisted paper ballots with some additional checks and balances over what we have historically had for paper ballots.
Mac OS 9 and under used to do this for both writes and reads. But sometimes you'd end up with two programs wanting different disks in the same drive. You ended up with the OS automatically ejecting disk A and asking for disk B. When you put in disk B, it'd automatically eject and ask for disk A. Because OS 9 was shared and not truly multi-process, when this happened you basically had to cut your losses and reboot.
Of course in a multi-processing OS, when you put in the requested disk A, it could do the work it needs to do for disk A while the dialog for disk B is still on the screen.
I'm certainly not a Linux filesystem guru, but I believe the way that ext3 and other journaling filesystems do their writes (even updates in the middle of a file) is that if you change 1 byte in the middle of an inode (or the entire inode) (file data is stored in chunks of a fixed size called inodes; same idea as Windows block size option; and this size is predetermined at time of formatting and stored as part of the filesystem headers... w/ 4096 byte inode size, a 1 byte file consumes as much disk space as a 4096 byte file... exactly 1 inode), it copies the entire chunk of data in that inode, plus your modifications, to a free inode, then updates the node map for that file, and finally marks the old inode as unused (or more accurately decrements its use counter, which accounts for times when a single inode is used by multiple files (hard links)).
Actually that's what ext2 does. Ext3, being journaled, creates the new inode, then creates a journal of what it is about to do (a pointer to the file in question, and the number of the old and the new inode, and the expected use counts for each), before updating the inode and usage counts and finally removing the journal. If the process gets interrupted during the writing of the new inode, the whole write fails (with the old file unmodified), and the new inode is still marked as unused. If the process gets interrupted during the journal write, the whole write fails (again with the old file unmodified). Once the journal has successfully been written, it can be guaranteed to succeed, any remaining work can be recovered. On startup of an uncleanly unmounted volume, the OS looks for complete (not completed) journal entries (if they exist, they are not yet complete), and does the work of the journal write again (even if the journal entry had completely succeeded, but had been interrupted before it was cleaned up, re-writing that data doesn't hurt anything).
So any interruption up to the point that the journal was completely written results in the entire write failing, and no loss of original data. Any interruption at that point or past results in the entire write succeeding (being finished once the OS mounts the partition).
From what I understand, ext3 actually is ext2, with the journal being handled as hidden files by the ext2 subsystem.
Now like I said, take all that with a grain of salt, it's mostly cobbled together understanding from incidental exposure to these things over time.
The success of his project is dependent on getting enough orders to give them economy of scale savings. He is courting the wealthier countries right now, and as he gets more orders, he can reduce cost further, and reach a more and more impoverished market. If Intel uses this as a way to have OLPC do all the front work of identifying markets, getting in to talk with the people who can make the buying decisions, and then end-gaming them on the sale, they are not only hurting the success of the OLPC project and shortening its eventual reach, they are also directly stealing value from the project which did the initial effort investment. Some of these countries might not have talked to Intel directly, but had heard good things about this OLPC thing and so first entertained the idea only because of the merits brought by that project. Intel is therefore operating under the guise of friendship and help, while cannibalizing the underlying foundation of the project.
OLPC laptops are more open, more free, better designed, and less expensive. Intel ClassMate PC's are proprietary, less rugged, and require more power to operate. Worst of all they are for-profit, and those profits are sent to Intel stock holders, making wealthy business men wealthier at the expense of money which would better be used satisfying an educational need in the exact same arena as the laptop was advertised as intending to assist. They unnecessarily drain valuable resources from the very market they are pretending to aid.
Basically this is about as disgustingly slimy as I think they are able to be.
The zany thing is that you don't need an actual bomb. All you have to do is communicate that you're holding a walkie-talkie whose other end is wired to a bomb in the cargo hold, push the button and BOOM.
If you're intending to hijack the plane, then you don't need an actual bomb, only the plausible and imminent threat of a bomb. Box cutters, people may charge you and expect to get cut a few times; maybe one guy dies. A bomb which can be activated with the push of a button, may easily go off before the button is pushed, taking out everyone in the plane.
Of course if they call your bluff, you have no recourse.
And X, Y, and Z turn around and sue B for misrepresenting the image to begin with. But like you said, this problem is not a Creative Commons problem, it's problem endemic to copyright which Creative Commons inherits because it is a license under copyright law.
I agree with you on that. The only time I ever use a WYSIWYG editor for HTML is when I have a large amount of copy to do a lot of formatting on (eg, lots of lists, bullets, bolds, italics, paragraphs, etc). Then I'm just using it to create extremely simple output (H's, P's, ol, ul, li, b, i, u, etc).
Likewise in Flex, I stick to hand-coding the MXML. I'm convinced this is actually faster than using the GUI, especially with the tag insight and completion. Plus I know the structure exactly. However, the GUI does a pretty decent job of generating clean code anyway. I still prefer hand coding.
Bluetooth. Works great without any external receiver if you already have bluetooth in your case.
Also you're considering the keyboard/mouse and receiver as separate products. It's a single product, wireless device and wireless receiver.
There is no standard other than bluetooth on wireless mouse/keyboard communications. They make a non-bluetooth version because not everyone has bluetooth or wants to use a bluetooth mouse (there's a small but noticeable latency which can be frustrating when doing fine mouse manipulations or gaming).
Canon & Nikon lenses don't support standard mounts because standard mounts are incompatible with their existing lenses. Modern SLR Nikons will still work just great with all Nikon lenses going back to 1985 (and a few before) (with the exception of the D40 & D40x which have a small body and some of those lenses rears extend too far into the body to allow the mirror to function). They don't support a standard mount because it means either locking out existing lenses (which means THOUSANDS of dollars for many people, it'd be about $3,000 for me) or getting an adapter (which will cost you 1-2 stops of light [yikes!] and increasing the focal length [not always a bad thing, but they make devices called extension tubes for when you want to do this]). There is still modern demand for Nikon lenses produced in the 80's and 90's, because some of those lenses were just really well engineered and manufactured. This is one of the reasons some people shoot exclusively Nikon; they want to know their lenses will not get obviated when they spend $2-3,000 on a single lens (I don't have one like that, but I have used one, and it is a wonderful creation).
You can get adapters to put just about any manufacturer's lens on just about any other manufacturer's body. Some times you lose some of the features such as auto focus (for example, if the lens depends on there being a drive in the camera, and the camera lacks a drive, there's nothing you can do!)
If you mean Flex (the product with which Silverlight is competing), you can use a standard text editor to create MXML files also. Of course the SDK is free and comes with an Eclipse plugin which supports both code insight and a GUI layout creator, so I'm not sure why you'd use a text editor when for no more cost you can have a full IDE.