Sure, the regular users of the forum won't see your name and address, but it's on file somewhere. Probably on some computer connected to the Internet, and probably the same server as the forum. What are the odds that some of the moderators can view it? What are the odds that an automated attack for common forum types will be published?
And what are the odds that some law enforcement officers will abuse this? Would you honestly be surprised by a US president from any party having the NSA or FBI tap a political opponent's phone? Would this be any different?
It isn't impossible, but it would probably slow you down a fair bit. The easiest way would be to keep the heap separated into two halves most of the time. Then one gets a bunch of holes in it; you copy data back and forth until one's mostly filled and the other is nearly empty; then you shrink the second to a minimal size.
Or you could simply create a new heap and compact everything onto that, then get rid of the old heap entirely. Or, if you don't want any overhead associated with a second heap, just copy it raw, empty the heap, and push everything back onto the heap.
It's not difficult per se, but it takes time, time accessing memory, and that's relatively slow. Especially when things have been swapped out of RAM. Moreover, this allows for more potential buffer overflows and such.
And the math department's decent, surprisingly. Good in geometry; for instance, Stony Brook is responsible for FIST (Fast Industrial Strength Triangulation), which was commissioned by Sun for the standard Java library. (Triangulation is basically separating a polygon into a set of triangles.)
The only way to tell if a creature had feathers or not is if you find it in a Lagerstatte, or an area with fossils preserving soft parts. Lagerstatten are relatively rare, the most famous being the Burgess Shale; our half dozen Archaeopteryx specimens are all Lagerstatten fossils, though, so they're not all confined to the Cambrian explosion. (In fact, there exists a pre-Cambrian Lagerstatte, if memory serves.)
Anyway, considering the rarity of Lagerstatten, there would be more excitement over this find if it had soft tissues preserved, even if there were no new species found at the site.
If you made a black satellite, it'd probably stay rather warm when it's behind the earth and heat up lots when exposed to the sun. So it would at least melt, though probably the combustion would be limited.
Why should I have to replace IE6 on my testing box and then return it afterwards just to test a beta web browser? I sure as hell can't waste an entire box to test IE7, even if it's only for a day; and a single day of tests would be insufficient.
Still, testing beta web browsers isn't something you should be doing at work anyway.
It would certainly eliminate repeat offenders; nobody would be convicted twice.
However, this would have the greatest effect when the police corps is well funded and pervasive enough to locate all criminals. If crime is a certain method of suicide, only suicidal people will be criminals.
"Maksim Orlovich discovered an incorrect bounds check in kjs, the JavaScript interpreter engine used by Konqueror and other parts of KDE, that allows a heap based buffer overflow when decoding specially crafted UTF-8 encoded URI sequences."
Modifying a bounds check is a small change, though important; it doesn't break anyone's JavaScript (except the malicious stuff). If it were a problem in a specific function and the fix could modify legitimate return values or place additional constraints on the arguments, then there would be trouble. In that case, the function in question should add an additional error check, and the full fix would be rolled out in the next release.
As for WMF stuff, that's just a failed graphics format with no reason to be included in Windows at all.
No, if the package has dependencies that *are* GPL, you still don't have to release your package under the GPL.
Take Crossover Office, for example. It depends on the Linux kernel, X, libstdc/libstdc++, and almost certainly glibc. Each of those are GPL. But Crossover Office is released under a commercial license.
Dependencies are irrelevant to the GPL; it's a matter of being a derivative work. If a GPL library offers certain functions, my code doesn't incorporate the body of those functions if I use that library; so I can distribute that library under the GPL and distribute my code under whatever license I choose. However, if I copy a portion of that library into my code, I have to GPL my code, provided I distribute it.
Now, if I have a non-GPL dependency for my code, I can still GPL my code. For instance, there are GPL applications for Windows, which is of course closed-source and proprietary.
The logic is this: My code doesn't include any part of the library, only an interface. If the library's released under the GPL, it doesn't matter unless I distribute it, in which case I need to distribute it under the GPL. If I have an application that uses it, it doesn't include it; the OS does the linking, but the user has to get it under the GPL's terms. All I have is an interface to that library.
On the other hand, if it's a commercial library or dependency, then it's still the user's responsibility to get it and follow the license. If the license forbids using it as a dependency, then it will violate the license to install and use my program. But that's not my responsibility.
It's much easier to segregate particular packages to use the GPL than most people think. And most software corporations try to propagate that myth. This isn't the place to do that.
Still, people would complain about the dependencies not being open sourced if the complete version history was released. I say keep it to yourself.
Let's see. Most desktop Linux distributions take up between 1 and 3GB, I believe; Ubuntu takes about 1.8GB but requires a root partition of 4GB. RHEL takes 1.8GB as well. (Neither of these feature development tools, though--even getting a sane build environment with Ubuntu takes half an hour or so.) WinXP is slightly heavier and gives you less; it's perhaps 2.5GB, but doesn't have such arcana as a Lisp interpreter built into Notepad. Instead, it has a number of poorly understood services which, disabled, may crash your computer or make it run slightly faster.
However, I've seen custom Windows installations that take up less than 500MB. It's a matter of ease--with Linux, someone else has done this and released their distribution, even has a website for it; with Windows, someone else has done this and released the results on BitTorrent without even a forum for support.
The article cited drivers as their main complaint, but also claim that they fix these errors. I hope that they send the relevant patches or snippets to the maintainers, and at least bug reports about everything else.
If Slashdot has at least 100,000 members (well, it's closer to ten times that, but....) and half of them owe $100 due to this patent at $0.25/pound, that's 50,000 people sharing 400 pounds of fat between them. That's 0.008 pounds of fat each.
Given an average body weight of, say, 130 pounds (averaged between male and female), that's less than 0.01% fat each. How many Slashdotters are starving to death as we speak?
If you used only long filenames, would the Windows drivers be able to deal with it? Probably. But you do need a space for the driver to write the short filename if necessary; just bzero it in the factory.
Really, the only issue is with organizations wanting to implement FAT--embedded systems with user-writeable media. If I want to sell flash drives pre-formatted with FAT32, I'm not writing drivers, so I'm not violating patents. And if someone else decides that they want to put files with long filenames on the drive I sold them, then that's their affair--as long as their driver supports it, Bob's their uncle.
Who suffers? Manufacturers of MP3 players, mainly, who either need to license FAT, add a Windows driver, or use a form of FAT that implements *exactly one* filename system (long or short, doesn't matter) but still allows both to be written (for Windows interoperability).
You could also implement only the long filename support and not fall under this patent. Who needs short filename support these days? People using Win95 probably aren't buying MP3 players en masse.
Slashdot Mirror
Sure, the regular users of the forum won't see your name and address, but it's on file somewhere. Probably on some computer connected to the Internet, and probably the same server as the forum. What are the odds that some of the moderators can view it? What are the odds that an automated attack for common forum types will be published?
And what are the odds that some law enforcement officers will abuse this? Would you honestly be surprised by a US president from any party having the NSA or FBI tap a political opponent's phone? Would this be any different?
No, that's not next. This is next; we covered random searches a few years ago.
Yes, but how many are stubs or redirects, and what's the average article size?
MAMEs together.
Couldn't resist.
It isn't impossible, but it would probably slow you down a fair bit. The easiest way would be to keep the heap separated into two halves most of the time. Then one gets a bunch of holes in it; you copy data back and forth until one's mostly filled and the other is nearly empty; then you shrink the second to a minimal size.
Or you could simply create a new heap and compact everything onto that, then get rid of the old heap entirely. Or, if you don't want any overhead associated with a second heap, just copy it raw, empty the heap, and push everything back onto the heap.
It's not difficult per se, but it takes time, time accessing memory, and that's relatively slow. Especially when things have been swapped out of RAM. Moreover, this allows for more potential buffer overflows and such.
And the math department's decent, surprisingly. Good in geometry; for instance, Stony Brook is responsible for FIST (Fast Industrial Strength Triangulation), which was commissioned by Sun for the standard Java library. (Triangulation is basically separating a polygon into a set of triangles.)
In case you were wondering, here's Dr. Sullivan's website: http://www.math.sunysb.edu/~dennis/
http://www.answersingenesis.org/museum/ (though it pains me to link to it)
The only way to tell if a creature had feathers or not is if you find it in a Lagerstatte, or an area with fossils preserving soft parts. Lagerstatten are relatively rare, the most famous being the Burgess Shale; our half dozen Archaeopteryx specimens are all Lagerstatten fossils, though, so they're not all confined to the Cambrian explosion. (In fact, there exists a pre-Cambrian Lagerstatte, if memory serves.)
Anyway, considering the rarity of Lagerstatten, there would be more excitement over this find if it had soft tissues preserved, even if there were no new species found at the site.
It seems that it would be more efficient to guide the roach into its nest, where there are presumably more roaches, than to take the roach away.
That might be dangerous, though.
No, it's sales.
You're free to reverse engineer the DRM and implement a free version, possibly, though oddly enough, nobody does that with DRM. I wonder why.
If you made a black satellite, it'd probably stay rather warm when it's behind the earth and heat up lots when exposed to the sun. So it would at least melt, though probably the combustion would be limited.
Why should I have to replace IE6 on my testing box and then return it afterwards just to test a beta web browser? I sure as hell can't waste an entire box to test IE7, even if it's only for a day; and a single day of tests would be insufficient.
Still, testing beta web browsers isn't something you should be doing at work anyway.
It would certainly eliminate repeat offenders; nobody would be convicted twice.
However, this would have the greatest effect when the police corps is well funded and pervasive enough to locate all criminals. If crime is a certain method of suicide, only suicidal people will be criminals.
"Maksim Orlovich discovered an incorrect bounds check in kjs,
the JavaScript interpreter engine used by Konqueror and other
parts of KDE, that allows a heap based buffer overflow
when decoding specially crafted UTF-8 encoded URI sequences."
Modifying a bounds check is a small change, though important; it doesn't break anyone's JavaScript (except the malicious stuff). If it were a problem in a specific function and the fix could modify legitimate return values or place additional constraints on the arguments, then there would be trouble. In that case, the function in question should add an additional error check, and the full fix would be rolled out in the next release.
As for WMF stuff, that's just a failed graphics format with no reason to be included in Windows at all.
I'm using Firebird 0.7, you insensitive clod!
(Well, actually, I am. Had this computer at uni in 2003, used by parents on dialup since.)
Why didn't she try the back of her finger? Or did that also have changing markings?
GNU Communis Publicas Potestas
Omnes exempla ad verbum transcribire distribuereque posunt, sed mutare non sinerint.
Exordium.
Permissiones plurimi
And that's where it goes to pieces. I had a link to a technical Latin dictionary, concentrating on computing terms; I seem to have lost it, though.
No, if the package has dependencies that *are* GPL, you still don't have to release your package under the GPL.
Take Crossover Office, for example. It depends on the Linux kernel, X, libstdc/libstdc++, and almost certainly glibc. Each of those are GPL. But Crossover Office is released under a commercial license.
Dependencies are irrelevant to the GPL; it's a matter of being a derivative work. If a GPL library offers certain functions, my code doesn't incorporate the body of those functions if I use that library; so I can distribute that library under the GPL and distribute my code under whatever license I choose. However, if I copy a portion of that library into my code, I have to GPL my code, provided I distribute it.
Now, if I have a non-GPL dependency for my code, I can still GPL my code. For instance, there are GPL applications for Windows, which is of course closed-source and proprietary.
The logic is this:
My code doesn't include any part of the library, only an interface. If the library's released under the GPL, it doesn't matter unless I distribute it, in which case I need to distribute it under the GPL. If I have an application that uses it, it doesn't include it; the OS does the linking, but the user has to get it under the GPL's terms. All I have is an interface to that library.
On the other hand, if it's a commercial library or dependency, then it's still the user's responsibility to get it and follow the license. If the license forbids using it as a dependency, then it will violate the license to install and use my program. But that's not my responsibility.
It's much easier to segregate particular packages to use the GPL than most people think. And most software corporations try to propagate that myth. This isn't the place to do that.
Still, people would complain about the dependencies not being open sourced if the complete version history was released. I say keep it to yourself.
Hm. Strike that second clause.
I really should preview.
Whoa, now. !(Netbeans == sucks). Rather, Netbeans \subset sucks. Likewise, Netbeans \subset sucks.
Let's see. Most desktop Linux distributions take up between 1 and 3GB, I believe; Ubuntu takes about 1.8GB but requires a root partition of 4GB. RHEL takes 1.8GB as well. (Neither of these feature development tools, though--even getting a sane build environment with Ubuntu takes half an hour or so.) WinXP is slightly heavier and gives you less; it's perhaps 2.5GB, but doesn't have such arcana as a Lisp interpreter built into Notepad. Instead, it has a number of poorly understood services which, disabled, may crash your computer or make it run slightly faster.
However, I've seen custom Windows installations that take up less than 500MB. It's a matter of ease--with Linux, someone else has done this and released their distribution, even has a website for it; with Windows, someone else has done this and released the results on BitTorrent without even a forum for support.
The article cited drivers as their main complaint, but also claim that they fix these errors. I hope that they send the relevant patches or snippets to the maintainers, and at least bug reports about everything else.
If Slashdot has at least 100,000 members (well, it's closer to ten times that, but....) and half of them owe $100 due to this patent at $0.25/pound, that's 50,000 people sharing 400 pounds of fat between them. That's 0.008 pounds of fat each.
Given an average body weight of, say, 130 pounds (averaged between male and female), that's less than 0.01% fat each. How many Slashdotters are starving to death as we speak?
If you used only long filenames, would the Windows drivers be able to deal with it? Probably. But you do need a space for the driver to write the short filename if necessary; just bzero it in the factory.
Really, the only issue is with organizations wanting to implement FAT--embedded systems with user-writeable media. If I want to sell flash drives pre-formatted with FAT32, I'm not writing drivers, so I'm not violating patents. And if someone else decides that they want to put files with long filenames on the drive I sold them, then that's their affair--as long as their driver supports it, Bob's their uncle.
Who suffers? Manufacturers of MP3 players, mainly, who either need to license FAT, add a Windows driver, or use a form of FAT that implements *exactly one* filename system (long or short, doesn't matter) but still allows both to be written (for Windows interoperability).
And the Linux kernel, of course.
You could also implement only the long filename support and not fall under this patent. Who needs short filename support these days? People using Win95 probably aren't buying MP3 players en masse.