I really doubt any imposters could fool close scrutiny with the ability to do arbitrary tests (like DNA, or just checking various historical medical conditions, did Saddam have any operations?). The imposters were only to fool photographs.
If they wanted to prove it, they should attack one of the GNU or some other FSF servers. Make sure there is enough identification to show that the attack is the same type and by the same person.
Attacking SCO's servers so SCO can say "oh, poor us, being attacked by those mean old Linux hackers" is so absolutely 100% benificial to SCO that there is no question that this is being done by them. Any SCO-hater would have attacked long ago, and at times that would hurt their case, such as when they did a press release.
Sure sounds like you don't have graphics acceleration.
As you may know, you are pretty much stuck with having to get an Nvidia card or a recent ATI card to get actual hardware OpenGL on Linux. Kind of sucks but at least you will also improve the Windows performance as well, with these better cards. If you have such a card you may be missing the drivers, which is probably more of a pain to solve than replacing the card. You can get the Nvidia driver from their site, look for the file with the instructions "sh NVIDEA.something.run"
There has been SIX MONTHS of SCO pissing people off, and no DDOS that SCO felt worth reporting. But now, convienently just before the first weekend after a bad-for-SCO press release (their losing that Utah court thing), this happens. And SCO goes and makes THREE press releases in one day, pushing the bad news off the financial pages.
Then, after hundreds of people (whether they are right or wrong) say "it's fishy because the ftp server still works), suddenly the ftp server is attacked! Why would any SCO-hater do this, when they could use their apparently available resources to attack any other SCO machine, ie everything *but* the ftp server?
I'm sorry, I do NOT buy this. SCO is lying and has done this to themselves.
The "attack" did not come from any open-source symphasizers.
After 24 hours the main argument that SCO was faking this was that their ftp server was up. It was very common knowledge and you can be absolutlely certain the hacker was reading the news about the hack. What happened then? Suddenly the attack slowed to the main server and it started up with double intensity to the ftp server! Look at the damn graph and see what other conclusion you can think of.
Any Leet SCO-hating fanatic would have doubled the attacks on the main server, or perhaps attacked every machine *except* the ftp site. That would have been the most clear "I hate you SCO and I'm going to mess with you as much as possible" attack. If they hated SCO they would want their attack to match the insults being directed at SCO as much as possible.
Instead the attack suddenly switched to be as exactly as possible a refutation of the publicity about the attack.
There is no question what the motives of the "attacker" are. And it is absolutly disgusting that SCO can get positive publicity for this nasty little stunt.
By your logic Lindows should be able to call their produce "Windows". I don't think that is true in the least. You might be able to make a new airplane and call it "Windows", but not an operating system.
The "Lindows" name is an obvious ploy on the "Windows" name. Microsoft thinks it is too similar, and I do too. The only stupid and wrong thing Microsoft is doing is drawing attention to this competitor with this lawsuit, why they don't just ignore it is a mystery to me.
Seriously though, the "Lindows" name always sounded stupid to me. And now they are stuck with having to make something similar to that so people know it is the same company. Best I can think of is to call it "Lind-OS".
The fact that it is all being stored in a single block is bad, as reallocating and copying that block as it gets larger will take linear time, especially if you do something stupid like reallocate on every addition rather than predict the space ahead of time. It is also possible that ext is really stupid about adding to the data structure, possibly rearranging the names into the order of the tree, which would cause the slowness you are seeing.
However the fact that the data structure is in one block does not mean it is a linear list.
*Looking up files* in the directory takes time that is approximately the log of the number of files, indicating a tree.
A hash would be better as this may reduce it to close-to constant lookup time, though it may make adding files even slower than what you are seeing now due to the need to rehash when the hash table is reallocated (avoided by chaining hash tables but that may make lookup slower).
Apple's OS/X Safari has this problem too. The preview shows "www.microsoft.com" although it is scrolled vertically 1/2 line so there is some hint that something is wrong. The address bar ends up with %00 like it should.
After seing Safari screw up, I tried Konqueror, but it seems to work (maybe). It shows "www.microsoft.com@secunia.com/..." in both the preview and in the address bar after you click (ie the %00 seems to have disappeared).
I would agree that this bug is not IE-only. It sounds like Opera is the only one doing the right thing. I would recommend that the browser should popup a warning for any username without a password, or containing a dot or any punctuation mark (to get around really stupid users who may read "http://www_microsoft_com@nasty.site.com/u_r_ownz" as being microsoft. You say ok to add the url to a list that won't pop up again.
Also web sites should reject url's with usernames, rather than accept them by default. This will get rid of:"joke" redirections, though it won't help if the redirector also controls their web site.
It is possible that when dithering image data it should use it's predefined colors rather than the pixel values in the image. My only examples do this, they only do the exact-color solution when a solid fill color is chosen by the application.
However I'm unsure if the bias is any problem. Once the first color is chosen it will be used for all colors that land in that "cell". You won't end up with most of the cells used for slightly different colors, which is a problem with most schemes.
I'd like to know why ALL of their DDOS attacks happen at times like 4:20 in the morning when they won't hurt their business? Can anybody come up with an explanation other than "they are doing it to themselves"?
If I understand this right, there was a bug. Maybe this bug was introduced by the previous patch, or maybe the previous patch did not work as expected, or whatever, but no matter what the reason, there was a bug, they could fix it, and they sent out a patch. That is the correct behavior.
They were probably being pretty stupid to say "no new patches". Due to Murphy's law, that guarantees that a problem will come up within days. Probably if they said "we are going to issue more patches than ever" then suddenly all their programmers would start have trouble finding bugs or figuring out how to fix them...
Anyway we can laugh at marketing for the "no new patches" but technically they did the right thing.
I did not know the 16-bit devices allowed you to change the colormap. Was the set of colors much larger than 16, however? I suspect they allowed changes to support blinking, not to make more colors. Only hardware I am familiar with like this is 1-bit displays (where they were black & white), some 2-bit displays where there was a "colormap" but only black and white so the colormap is only useful for blinking, and EGA which I thought had a fixed colormap.
If the colormap will be fixed, why manufacture devices that allow it to be changed at all?
The X server can change the colormap. However it is now it's job to manage the colors being requested by the programs and display them as best as possible. A simple algorithm that works pretty good is to design a fixed colormap, but when a color is requested, round it to the nearest colormap entry, and if that colormap entry has not been used, set it to *exactly* the requested color, rather than the color you intended for the colormap. Then fix your dithering algorithims to work with the colors as assigned in the colormap. This tends to reduce dithering a great deal for most programs. This requires a writable colormap without having to present it in the API.
Readline has to be *linked* with the program to be useful. This is totally different from just having seen or worked with Readline before. Plenty of closed source is written by people "tainted" by Readline and FSS/RMS cannot and will not do a thing about it. The GPL is not viral no matter how much you want to pretend it is.
I agree RMS is being an ass in putting useful shareable functionality under the GPL. The end result is that all Linux command-line programs do not use readline and thus lack editing capabilities. It probably has forced *zero* software to be GPL.
It appears that Slashdot just removes posts that are below your threshold and "reparents" responses to them so they look like they are responses to the parent.
I would prefer that Slashdot either remove the post and *all* responses, or have it act like the score is the maximum of itself and all responses to it. Ie if somebody posts a down-rated comment, but somebody else posts and high-scoring response, that probably means the down-rated comment is interesting.
Systems with 4 bits usually do not have a colormap. They have a fixed set of 16 colors.
It is pretty obvious that with any kind of intelligent design, an 8-bit display can present a nice full-color display without exposing the internal colormap.
I agree with the original poster that colormaps should have been scrapped long ago. This can be done by making the X server always present a single TrueColor (or maybe DirectColor) visual, and emulating it using a fixed colormap on displays with less than 24 bits per pixel.
One good thing about XFree86 is that it only allowed one visual, so that managed to get rid of all the programs that *required* a colormap. We should not lose the one good thing they did!
The ctrl+/- changes the video resolution (the monitor gets a different number of pixels). It does not change the virtual screen (the area that programs think is visible).
Making a way to change the virtual screen, not the video resolution, is what is wanted, and is currently missing.
I am rather annoyed that RandR is so complex. Why couldn't they just send a ConfigureNotify event to the root window? I would think most window managers could be easily rewritten to use that, instead of inventing a whole new protocol. In addition it would be nice if attempts to resize the root window caused the X server to pick the nearest resolution and switch to that. See, it can all be done without adding to the interface!.
Anybody who does not believe that forking happens in the internal development of closed-source is totally ignorant of the corporate software development process. I have worked at several large compainies and in no cases was the code I worked on not a fork. I would estimate there were 3 or 4 simultaneous forks at all times. In the end one would probably be selected and 3/4 of the work people did was lost. Sometimes it would fork and diverge into totally incompatable products for different platforms, or for "professional" versus "personal" use.
I agree with the "removing the catalytic converter is the same as modifying the source code" person.
The argument that people can violate FCC regulations because they have the source code and this would make the provider of the source code liable is stupid. You can violate FCC regulations easily and much worse with a bunch of parts you can buy from Radio Shack, and the store and manufacturers of those parts are not worried about liability.
I think Linus meant copying parts of that header file (ie the implementation of an inline function).
Not absolutely certain, now, however. But it would otherwise seem that he is saying that non-GPL modules are impossible (how can you possilby make a useful module without #include of at least one kernel header?) All his other comments seem to indicate that he is ok with binary modules, so I think this sentence is being mis-read.
1. Linus's wording is extremely bad, and I'm not sure why people don't seem to notice it. They either read it the way he intended and realize that binary modules are fine by him, or they read it wrongly (the way I did) and think he is saying that binary modules are illegal. He said something along the lines of "code the uses any part of the kernel source, including the headers, must be GPL". By "use" he meant that a big chunk was copied and pasted or otherwise edited into the module. Unfortunatly a lot of people (like me) think that a piece of code that #includes a header file "uses" that header file, leading to the shocking conclusing that Linus was saying binary modules are not allowed at all (completely contradictory to everything else he said).
2. The first poster wrote the "PWC" code. Even though he GPL'd it, it is still *his* code. So copying some of it for the non-GPL "PWCX" is perfectly fine. Sending a copy to Phillips and asking them to write a non-GPL thing from it is also fine.
a company won't put trade secrets into a binary driver to run the risk of losing these trade secrets because of a lawsuit based on the GPL which states that the binary only driver has to be opened up
This is NOT going to happen. You are talking about a copyright violation. The worst that will happen is the company will be told to stop distributing their driver.
Okay, possibly the manufacturer may be liable to damages from some copyright holder of a part of Linux. That is going to end up being money, it is not going to be a requirement that the company lose their copyright. If you can find a single case anywhere where a copyright violation has caused a company to lose it's other copyrights I would be interested in hearing about it.
The "you may lose your copyright" boogyman argument is FALSE and is FUD being perpetuations by those who will gain if GPL code is discouraged.
Slashdot Mirror
I really doubt any imposters could fool close scrutiny with the ability to do arbitrary tests (like DNA, or just checking various historical medical conditions, did Saddam have any operations?). The imposters were only to fool photographs.
If they wanted to prove it, they should attack one of the GNU or some other FSF servers. Make sure there is enough identification to show that the attack is the same type and by the same person.
Attacking SCO's servers so SCO can say "oh, poor us, being attacked by those mean old Linux hackers" is so absolutely 100% benificial to SCO that there is no question that this is being done by them. Any SCO-hater would have attacked long ago, and at times that would hurt their case, such as when they did a press release.
According to the page source it is PostNuke.
Sure sounds like you don't have graphics acceleration.
As you may know, you are pretty much stuck with having to get an Nvidia card or a recent ATI card to get actual hardware OpenGL on Linux. Kind of sucks but at least you will also improve the Windows performance as well, with these better cards. If you have such a card you may be missing the drivers, which is probably more of a pain to solve than replacing the card. You can get the Nvidia driver from their site, look for the file with the instructions "sh NVIDEA.something.run"
There has been SIX MONTHS of SCO pissing people off, and no DDOS that SCO felt worth reporting. But now, convienently just before the first weekend after a bad-for-SCO press release (their losing that Utah court thing), this happens. And SCO goes and makes THREE press releases in one day, pushing the bad news off the financial pages.
Then, after hundreds of people (whether they are right or wrong) say "it's fishy because the ftp server still works), suddenly the ftp server is attacked! Why would any SCO-hater do this, when they could use their apparently available resources to attack any other SCO machine, ie everything *but* the ftp server?
I'm sorry, I do NOT buy this. SCO is lying and has done this to themselves.
The "attack" did not come from any open-source symphasizers.
After 24 hours the main argument that SCO was faking this was that their ftp server was up. It was very common knowledge and you can be absolutlely certain the hacker was reading the news about the hack. What happened then? Suddenly the attack slowed to the main server and it started up with double intensity to the ftp server! Look at the damn graph and see what other conclusion you can think of.
Any Leet SCO-hating fanatic would have doubled the attacks on the main server, or perhaps attacked every machine *except* the ftp site. That would have been the most clear "I hate you SCO and I'm going to mess with you as much as possible" attack. If they hated SCO they would want their attack to match the insults being directed at SCO as much as possible.
Instead the attack suddenly switched to be as exactly as possible a refutation of the publicity about the attack.
There is no question what the motives of the "attacker" are. And it is absolutly disgusting that SCO can get positive publicity for this nasty little stunt.
By your logic Lindows should be able to call their produce "Windows". I don't think that is true in the least. You might be able to make a new airplane and call it "Windows", but not an operating system.
The "Lindows" name is an obvious ploy on the "Windows" name. Microsoft thinks it is too similar, and I do too. The only stupid and wrong thing Microsoft is doing is drawing attention to this competitor with this lawsuit, why they don't just ignore it is a mystery to me.
"Butt-head monopolist"
Seriously though, the "Lindows" name always sounded stupid to me. And now they are stuck with having to make something similar to that so people know it is the same company. Best I can think of is to call it "Lind-OS".
The fact that it is all being stored in a single block is bad, as reallocating and copying that block as it gets larger will take linear time, especially if you do something stupid like reallocate on every addition rather than predict the space ahead of time. It is also possible that ext is really stupid about adding to the data structure, possibly rearranging the names into the order of the tree, which would cause the slowness you are seeing.
However the fact that the data structure is in one block does not mean it is a linear list.
*Looking up files* in the directory takes time that is approximately the log of the number of files, indicating a tree.
A hash would be better as this may reduce it to close-to constant lookup time, though it may make adding files even slower than what you are seeing now due to the need to rehash when the hash table is reallocated (avoided by chaining hash tables but that may make lookup slower).
Apple's OS/X Safari has this problem too. The preview shows "www.microsoft.com" although it is scrolled vertically 1/2 line so there is some hint that something is wrong. The address bar ends up with %00 like it should.
" as being microsoft. You say ok to add the url to a list that won't pop up again.
:"joke" redirections, though it won't help if the redirector also controls their web site.
After seing Safari screw up, I tried Konqueror, but it seems to work (maybe). It shows "www.microsoft.com@secunia.com/..." in both the preview and in the address bar after you click (ie the %00 seems to have disappeared).
I would agree that this bug is not IE-only. It sounds like Opera is the only one doing the right thing. I would recommend that the browser should popup a warning for any username without a password, or containing a dot or any punctuation mark (to get around really stupid users who may read "http://www_microsoft_com@nasty.site.com/u_r_ownz
Also web sites should reject url's with usernames, rather than accept them by default. This will get rid of
It is possible that when dithering image data it should use it's predefined colors rather than the pixel values in the image. My only examples do this, they only do the exact-color solution when a solid fill color is chosen by the application.
However I'm unsure if the bias is any problem. Once the first color is chosen it will be used for all colors that land in that "cell". You won't end up with most of the cells used for slightly different colors, which is a problem with most schemes.
I'd like to know why ALL of their DDOS attacks happen at times like 4:20 in the morning when they won't hurt their business? Can anybody come up with an explanation other than "they are doing it to themselves"?
If I understand this right, there was a bug. Maybe this bug was introduced by the previous patch, or maybe the previous patch did not work as expected, or whatever, but no matter what the reason, there was a bug, they could fix it, and they sent out a patch. That is the correct behavior.
They were probably being pretty stupid to say "no new patches". Due to Murphy's law, that guarantees that a problem will come up within days. Probably if they said "we are going to issue more patches than ever" then suddenly all their programmers would start have trouble finding bugs or figuring out how to fix them...
Anyway we can laugh at marketing for the "no new patches" but technically they did the right thing.
I did not know the 16-bit devices allowed you to change the colormap. Was the set of colors much larger than 16, however? I suspect they allowed changes to support blinking, not to make more colors. Only hardware I am familiar with like this is 1-bit displays (where they were black & white), some 2-bit displays where there was a "colormap" but only black and white so the colormap is only useful for blinking, and EGA which I thought had a fixed colormap.
If the colormap will be fixed, why manufacture devices that allow it to be changed at all?
The X server can change the colormap. However it is now it's job to manage the colors being requested by the programs and display them as best as possible. A simple algorithm that works pretty good is to design a fixed colormap, but when a color is requested, round it to the nearest colormap entry, and if that colormap entry has not been used, set it to *exactly* the requested color, rather than the color you intended for the colormap. Then fix your dithering algorithims to work with the colors as assigned in the colormap. This tends to reduce dithering a great deal for most programs. This requires a writable colormap without having to present it in the API.
Attach the moon leash to the North Pole! Easy!
You need a swivel connector, though.
Readline has to be *linked* with the program to be useful. This is totally different from just having seen or worked with Readline before. Plenty of closed source is written by people "tainted" by Readline and FSS/RMS cannot and will not do a thing about it. The GPL is not viral no matter how much you want to pretend it is.
I agree RMS is being an ass in putting useful shareable functionality under the GPL. The end result is that all Linux command-line programs do not use readline and thus lack editing capabilities. It probably has forced *zero* software to be GPL.
It appears that Slashdot just removes posts that are below your threshold and "reparents" responses to them so they look like they are responses to the parent.
I would prefer that Slashdot either remove the post and *all* responses, or have it act like the score is the maximum of itself and all responses to it. Ie if somebody posts a down-rated comment, but somebody else posts and high-scoring response, that probably means the down-rated comment is interesting.
Systems with 4 bits usually do not have a colormap. They have a fixed set of 16 colors.
It is pretty obvious that with any kind of intelligent design, an 8-bit display can present a nice full-color display without exposing the internal colormap.
I agree with the original poster that colormaps should have been scrapped long ago. This can be done by making the X server always present a single TrueColor (or maybe DirectColor) visual, and emulating it using a fixed colormap on displays with less than 24 bits per pixel.
One good thing about XFree86 is that it only allowed one visual, so that managed to get rid of all the programs that *required* a colormap. We should not lose the one good thing they did!
Um, network transparency is the entire reason they want to add the sound. So yes it will be network transparent.
Actually you have the terminology backwards.
The ctrl+/- changes the video resolution (the monitor gets a different number of pixels). It does not change the virtual screen (the area that programs think is visible).
Making a way to change the virtual screen, not the video resolution, is what is wanted, and is currently missing.
I am rather annoyed that RandR is so complex. Why couldn't they just send a ConfigureNotify event to the root window? I would think most window managers could be easily rewritten to use that, instead of inventing a whole new protocol. In addition it would be nice if attempts to resize the root window caused the X server to pick the nearest resolution and switch to that. See, it can all be done without adding to the interface!.
Anybody who does not believe that forking happens in the internal development of closed-source is totally ignorant of the corporate software development process. I have worked at several large compainies and in no cases was the code I worked on not a fork. I would estimate there were 3 or 4 simultaneous forks at all times. In the end one would probably be selected and 3/4 of the work people did was lost. Sometimes it would fork and diverge into totally incompatable products for different platforms, or for "professional" versus "personal" use.
I agree with the "removing the catalytic converter is the same as modifying the source code" person.
The argument that people can violate FCC regulations because they have the source code and this would make the provider of the source code liable is stupid. You can violate FCC regulations easily and much worse with a bunch of parts you can buy from Radio Shack, and the store and manufacturers of those parts are not worried about liability.
I think Linus meant copying parts of that header file (ie the implementation of an inline function).
Not absolutely certain, now, however. But it would otherwise seem that he is saying that non-GPL modules are impossible (how can you possilby make a useful module without #include of at least one kernel header?) All his other comments seem to indicate that he is ok with binary modules, so I think this sentence is being mis-read.
I think he is perfectly fine.
1. Linus's wording is extremely bad, and I'm not sure why people don't seem to notice it. They either read it the way he intended and realize that binary modules are fine by him, or they read it wrongly (the way I did) and think he is saying that binary modules are illegal. He said something along the lines of "code the uses any part of the kernel source, including the headers, must be GPL". By "use" he meant that a big chunk was copied and pasted or otherwise edited into the module. Unfortunatly a lot of people (like me) think that a piece of code that #includes a header file "uses" that header file, leading to the shocking conclusing that Linus was saying binary modules are not allowed at all (completely contradictory to everything else he said).
2. The first poster wrote the "PWC" code. Even though he GPL'd it, it is still *his* code. So copying some of it for the non-GPL "PWCX" is perfectly fine. Sending a copy to Phillips and asking them to write a non-GPL thing from it is also fine.
a company won't put trade secrets into a binary driver to run the risk of losing these trade secrets because of a lawsuit based on the GPL which states that the binary only driver has to be opened up
This is NOT going to happen. You are talking about a copyright violation. The worst that will happen is the company will be told to stop distributing their driver.
Okay, possibly the manufacturer may be liable to damages from some copyright holder of a part of Linux. That is going to end up being money, it is not going to be a requirement that the company lose their copyright. If you can find a single case anywhere where a copyright violation has caused a company to lose it's other copyrights I would be interested in hearing about it.
The "you may lose your copyright" boogyman argument is FALSE and is FUD being perpetuations by those who will gain if GPL code is discouraged.