High-end RAID cards can still be better than software RAID. They likely include battery backup, so that writes can still go out in the event of a power failure (assuming your drives also have battery backup.) This helps to mitigate some of the failure modes that striped RAID has.
Short of this, though, the only time you'd want hardware RAID is if you've got an underpowered system. This happens more than you'd think, though, particularly when people are running lots of other applications on their disk array.
I don't understand why you lost data. Were you not able to reinstall your OS and then re-access the drives?
With hardware RAID, you need to be able to replace the RAID card with an identical one in order to guarantee that you'll have access to your data. With software RAID, it would seem like all you need is a compatible OS (i.e. the one you were using before.)
If the OS trashes your data, you're screwed, but this can happen even with hardware RAID.
I guess it depends upon how the user comes across Linux. Did someone set them up with it? It could be anything (there are many distros made to be lightweight for older hardware.) Did they hear about it and decide to download and try it? If so, they probably just got Ubuntu. But did they turn on "Desktop Effects" (which changes the window manger behind the scenes, and as far as I can tell, there's no common interface to modify settings for both of the standard WMs.)
I guess that the really important stuff is pretty standardized, though. Firefox (or Iceweasel) should be very similar between Linux distributions (though it's actually got a fair number of differences between Windows and Linux--mostly menu changes.) So from an education standpoint, if you teach a user about safety with Firefox, you're teaching it for most distros. I don't mess around with file managers all that much, so honestly, I can't say for sure what dangers might be lurking there (I just did some simple testing earlier before I wrote my previous post.)
With the slow transition to web applications, the browser becomes an operating system in its own right. So application-based security controls aren't going to be enough--if IE gets compromised, then everything I do in IE is compromised. We actually need per-site granularity, but this has multiple issues, too (cross-site scripting and sites which actually need to share security settings with one-another.)
It's a very difficult problem for this reason. Per-application granularity is actually really easy to solve, though slightly harder to solve efficiently.
On linux a single application file isn't executable by default when it's downloaded straight from the web or a mail client(untaring is another matter), the user would have to set it executable manually which, while maybe a usability issue in some ways, makes things a fair bit safer by default. It also helps that binaries don't show off an internal icon in the file manager but a standard icon for all apps. "On Linux" is an almost useless statement when you're talking about user-interfaces. The file manager determines what icons are shown, and it's entirely feasible for a file to have a.jpg extension but be executable (unlike Windows, the file extension on the Linux binary loader[1] doesn't necessarily determine the operation of the file in any way.) The shell you're using may enforce some restrictions, like making it impossible to "execute" files with a commonly non-executable extension (like.jpg). The stock Gnome file manager does this, for example, but others may not.
This is one advantage that Windows actually has over Linux. Education is easier on Windows because of the homogeneous environment. With Linux, you have to worry about multiple desktop environments (KDE, Gnome, others?), multiple window managers (Metacity, Fluxbox, whatever the whiz-bang 3d window manager du jour is, etc.), multiple shells (though bash is most common, tcsh is another common Unix default), etc. Worse, you have different versions of the software with drastically different interfaces (ok, Windows has this between major versions, but I never seem to have a problem converting things over in a Windows environment--there's almost always a theme or setting to go back to the old way of doing things. I've upgraded software on Linux in the past to find a completely different interface than I was accustomed to.)
Anyway, this wasn't meant to be a rant for or against Linux, so I should probably stop now.
[1] As far as I know, there aren't any binary loaders written for Linux which enforce any restrictions on this. While it would be possible to write one, this is one time where "on Linux" actually makes sense, despite my initial statement to the contrary.
Are you suggesting that the vast majority of people get infected through vectors like this? Because while a few people might get infected this way (and in turn infect other people), my money would be on more people getting infected by opening e-mail attachments from people that they don't know.
Yes. Clearly the exact question posed in the article is not appropriate for the intent. That is, people who actually read the prompt and want a more secure connection will also choose "Yes."
The idea of a test to determine if the user is savvy or not should be applied by the ISP before letting them plug anything into the net. I've been saying for years that you should need a license to use the Internet. And an even more strict license to operate a server. Probably even by class of server (Windows, Unix, etc.)
Blocking port 25 is going to stop your average 14 year for about 60 seconds. I disagree with this. While there are ways around blocking port 25 (and 587, if you're being thorough), it does one of two things:
1) It forces the malware to go through your smtp server, which means that you can apply spam filtering and other heuristics to determine if it's legitimate. or 2) It forces the malware to tunnel, which eliminates it as the ISP's problem, and makes it easier to block, assuming there are only a few places where the malware could tunnel their SMTP connections.
I thought that the GPLv3 pushing masses away was implied by the vague circumstances you listed in your post.
Just look at some of their actions recently, they attacked novell for something that only existed in their mind in order to push a license not many people were interested in and they done this at a time when nobody liked Vista and the switch to vista could have been just as dramatic as the switch to linux. So instead of embracing Vista's rejection, they created an unneeded controversy to keep the masses at bay. I figured that the center of the controversy was GPLv3 and its changes to how patents are addressed, but even if it wasn't, I highly doubt that the masses are paying one bit of attention to the FSF or their controversies. People stay away from Linux out of ignorance and fear, and very occasionally the fears are justified (the fear that their software won't run, or that there won't be comparable software.)
Mine was a Palm IIIxe. I loved it! The best thing (compared to PDAs of today) is that it ran on standard batteries. Starting to go dead? Pick up more. No need to let it sit charging before you can continue to use it.
Sure. The point was that making blanket statments about a medium as varied as Anime just doesn't work. Nor does saying, "I'd like to get into anime, so where do I start?" have an answer applicable to everyone.
Anime (or more broadly, animation) may have elements that make certain things more feasible (such as the freedom to play around with casting, etc.) but that doesn't mean that all of them do it. I pointed out Sailor Moon and DBZ precisely because they're commonly known outside of fandom, and they're perfect counterexamples to some of the common claims made about anime's superiority.
I'm not interested in a fight or anything. I tried Cowboy Beebop (sounded perfect for my tastes) and it just didn't grab me. I think that I have subconscious expectations (from my culture) when it comes to animation that are just too hard to get over. Objectively, I can look at an anime or manga and say that it has certain qualities[1], and sometimes qualities that I'll look for in other mediums, but when I sit down to watch/read, I just tend to get put off.
[1] And more importantly, I can say that just because a story is presented in a certain medium doesn't mean that it's marketed to a specific age group. Contrast this to people who think that all comics are for kids, and thus comics with excessive sex and violence are clearly trying to corrupt our youth. I can objectively realize this, but that doesn't mean that subconsciously, when I sit down to watch it, I'm not put off by the perceived conflation of adult themes and "cartoony" rendering. It's a personal failing, but at least I don't try to impose my preconceptions upon society.
Who needs a paper backup when you can always re-download if something happens? (Isn't the absence of paper rather the whole *point*?) Yeah, that worked really well in the past. I don't really expect that Kindle is going to work out, so I have to wonder if the Kindle store will continue. At least with Kindle, you can store your purchases externally, and there's no indication that the device checks in with Amazon before it lets you read, but the "delete any time, since you can just re-download" line is asking for trouble if they ever stop supporting the service.
Orson Scott Card (which is my current reading material). Man, if it was a bit cheaper, and the Ender saga was on there, I might buy one just for the experience of reading Ender's Game on one of these:)
I believe that the Kindle allows for annotation by the reader. I don't know if it's written (with a stylus) or typed, but it's definitely one of the features that I think they got right.
I'm not inclined to give much credence to the complaints of folks who haven't actually held one of these in their hands. A reasonable position, but it outlines another flaw with the Kindle. Amazon isn't putting them into stores, so we can't test them out. We have to spend $400 basically blind in order to see whether we like the feel of the device. Not going to happen.
when you wish you had something to read with you I've pretty much always got a book that I'm currently reading. Whenever I wish that I had "something to read", it's always, "I wish I'd brought my book with me." Kindle doesn't really solve this problem--it just changes it to, "I wish I'd brought my Kindle with me." The only way to address this is to make the Kindle small enough to fit in my pocket, and that utterly destroys its value to most people (who want something with a larger screen.) Foldable e-paper would do the trick, or holographic projection, but then we're entering the realm of sci-fi.
or you didn't know in advance that you wanted to read just THAT book This is even rarer for me. It's really rare that I want THIS BOOK RIGHT NOW for older books, particularly ones that I've already read. Compare to music (because Kindle+Amazon is similar to iPod+iTMS) where your arguments apply to me much more.
That said, it may just be that the Kindle isn't for me, but other people will flock to it. We'll see, I guess.
Before your comment on O'Reilly books, I was really giving the Kindle a very "meh" response. It looks like Amazon is trying to follow the Apple iTMS+iPod model, but they've overlooked a few things.
The most notable thing they've overlooked is that books are not like CDs. iPods work well because they don't deviate all that much from the previous portable music models, and where they do deviate, they make noticeable improvements. With the iPod, for example, you still listen to the music through headphones. The primary listening interface has not changed at all. However with the Kindle, they've changed the primary reading interface to a screen. Now that might not seem too important, but it's going to turn a lot of people off right away. Then you've got things like flipping through a book--it's really hard to recreate this feeling with a device like this, and though I haven't played with a Kindle, I find it hard to imagine that they could have succeeded. Then, of course, there's portability. While the Kindle would be useful on my desk or coffee-table, I'm not going to carry one around with me. For reading novels, then, I'm basically no better off than if I had just bought the novel, since I'm probably never going to want to carry a dozen novels around with me at all times (unlike music.) And with the novel, I can choose to loan it to a friend. I can't do that with the Kindle without loaning them all of my novels, and depriving myself of them (though this is an issue with any DRM, including iTunes, it's especially applicable to my circle of friends, who loan each other novels all of the time, but rarely loan music.)
I look at this device and all I can see is the number of novels I could buy for $400 instead, have a better reading experience, and have them separate from each other. A specialized solution for O'Reilly books (like the Safari "bookshelf" in portable form) would be very nice to be sure, but I don't know that I could justify $400+book costs for it.
You even quoted all of the relevant parts. The entire portion of text that you qutoed:
I didn't believe it until I listened to some jazz Vorbis files over an M-Audio 5.1 Revolution card with Sennheiser HD 540 Open-Aire headphones (neither of which are expensive enough to be truly "audiophile"), and I noticed some distortion in the high frequency sounds. Playing the same song from the original CD was significantly better, and I assume that playing a FLAC (or Apple or MS equivalent) would sound near identical. When you're called out for not reading the post before posting a huge rant, you might actually go back and re-read the post just to make sure that you were wrong, rather than assuming that you were right. That's politician-like behavior.
Not at all. The malicious FLAC starts up a daemon that sits and waits for you to use su/sudo, then it performs its ptrace injection. Maybe it modifies your bash_login file so that it runs every time you log in. It could even probably spoof an update-manager window, making you think that there are updates to install (which will then prompt you to enter your password.)
That's the thing about security. If the OS allows the user full control, then your security implementation is only as good as your user is knowledgeable. If you moved most Windows users over to Linux, it wouldn't take a week for malware authors to start spoofing gksudo windows.
Actually, the admin account on Vista doesn't have to enter a password--they just have to click through the UAC control. If you try to perform an administrative action as a regular user in Vista, you'll have to enter a password as part of the UAC control.
I guess that it depends upon the definition of "end users." Am I an "end user" of gcc, even though I've never compiled anything in my life? After all, I'm using software that was compiled with gcc.
It's the same thing with any web service software and the AGPL. If I download web service software and install it to my server, I feel as though I am the user of the software, and my web visitors are users of my site (which is the output of the software--specifically excluded from being a derivative work under the GPL, but not the AGPL in the specific case of web services.)
Put more succinctly, the GPL never inhibits use of software (that is, running software.) The AGPL does. That's why the AGPL applies to the users of software (the people running the actual software), not just distributors.
The theory is the same as with any EULA (the AGPL is truly a EULA, unlike the GPL.) The AGPL allows you to use the code if you give the code back to any of your users. If you refuse to do that, you are violating the license, and thus you have no right to use the software.
It's the same thing as Microsoft saying that by using their software, you absolve them of any wrongdoing. It's part of the license. If you try to sue them, you're in violation, and thus never had a right to use the software in the first place.
This is why EULAs are bad. I'm very sad to see the FSF going to one, though maybe it will drive people to more open licenses (like BSD.)
So that file doesn't count as a larger work that then must be distributed under the AGPL. If so, then the AGPL is worthless. I can just include my changes to the AGPL project and distribute a clean version of the source, but with the "include \"mystuff.php\"" and a few new function calls. My changes don't go out (they were just included!)
The whole thing is sticky and not nearly as well thought out as the GPL.
It is a EULA, which is a great departure from the GPL of old.
Of course, web applications break the model. There are two users--the host is using the application to interact with the user, and the user is using the web application to interact with the host. It's a really weird situation. That said, I kinda like the separation of the old days, because frankly, where does this end? Should the user get a copy of Apache just for visiting a website? If the website interacts with a database, should they get a copy of the database? What if the database is proprietary, but the front-end (web server/php server) is AGPL?
This really muddies the waters. Services traditionally haven't been touched by copyleft, and if this thing catches on, it will have some very interesting (and probably unfortunate) ramifications.
Slashdot Mirror
High-end RAID cards can still be better than software RAID. They likely include battery backup, so that writes can still go out in the event of a power failure (assuming your drives also have battery backup.) This helps to mitigate some of the failure modes that striped RAID has.
Short of this, though, the only time you'd want hardware RAID is if you've got an underpowered system. This happens more than you'd think, though, particularly when people are running lots of other applications on their disk array.
I don't understand why you lost data. Were you not able to reinstall your OS and then re-access the drives?
With hardware RAID, you need to be able to replace the RAID card with an identical one in order to guarantee that you'll have access to your data. With software RAID, it would seem like all you need is a compatible OS (i.e. the one you were using before.)
If the OS trashes your data, you're screwed, but this can happen even with hardware RAID.
I guess it depends upon how the user comes across Linux. Did someone set them up with it? It could be anything (there are many distros made to be lightweight for older hardware.) Did they hear about it and decide to download and try it? If so, they probably just got Ubuntu. But did they turn on "Desktop Effects" (which changes the window manger behind the scenes, and as far as I can tell, there's no common interface to modify settings for both of the standard WMs.)
I guess that the really important stuff is pretty standardized, though. Firefox (or Iceweasel) should be very similar between Linux distributions (though it's actually got a fair number of differences between Windows and Linux--mostly menu changes.) So from an education standpoint, if you teach a user about safety with Firefox, you're teaching it for most distros. I don't mess around with file managers all that much, so honestly, I can't say for sure what dangers might be lurking there (I just did some simple testing earlier before I wrote my previous post.)
Hard to say.
With the slow transition to web applications, the browser becomes an operating system in its own right. So application-based security controls aren't going to be enough--if IE gets compromised, then everything I do in IE is compromised. We actually need per-site granularity, but this has multiple issues, too (cross-site scripting and sites which actually need to share security settings with one-another.)
It's a very difficult problem for this reason. Per-application granularity is actually really easy to solve, though slightly harder to solve efficiently.
This is one advantage that Windows actually has over Linux. Education is easier on Windows because of the homogeneous environment. With Linux, you have to worry about multiple desktop environments (KDE, Gnome, others?), multiple window managers (Metacity, Fluxbox, whatever the whiz-bang 3d window manager du jour is, etc.), multiple shells (though bash is most common, tcsh is another common Unix default), etc. Worse, you have different versions of the software with drastically different interfaces (ok, Windows has this between major versions, but I never seem to have a problem converting things over in a Windows environment--there's almost always a theme or setting to go back to the old way of doing things. I've upgraded software on Linux in the past to find a completely different interface than I was accustomed to.)
Anyway, this wasn't meant to be a rant for or against Linux, so I should probably stop now.
[1] As far as I know, there aren't any binary loaders written for Linux which enforce any restrictions on this. While it would be possible to write one, this is one time where "on Linux" actually makes sense, despite my initial statement to the contrary.
Are you suggesting that the vast majority of people get infected through vectors like this? Because while a few people might get infected this way (and in turn infect other people), my money would be on more people getting infected by opening e-mail attachments from people that they don't know.
1) It forces the malware to go through your smtp server, which means that you can apply spam filtering and other heuristics to determine if it's legitimate.
or 2) It forces the malware to tunnel, which eliminates it as the ISP's problem, and makes it easier to block, assuming there are only a few places where the malware could tunnel their SMTP connections.
Both cases reduce the spam problem.
That's what you think.
Sorry if I assumed too much.
Mine was a Palm IIIxe. I loved it! The best thing (compared to PDAs of today) is that it ran on standard batteries. Starting to go dead? Pick up more. No need to let it sit charging before you can continue to use it.
Fair points. I don't usually think of returning items just because I didn't like them.
Sure. The point was that making blanket statments about a medium as varied as Anime just doesn't work. Nor does saying, "I'd like to get into anime, so where do I start?" have an answer applicable to everyone.
Anime (or more broadly, animation) may have elements that make certain things more feasible (such as the freedom to play around with casting, etc.) but that doesn't mean that all of them do it. I pointed out Sailor Moon and DBZ precisely because they're commonly known outside of fandom, and they're perfect counterexamples to some of the common claims made about anime's superiority.
I'm not interested in a fight or anything. I tried Cowboy Beebop (sounded perfect for my tastes) and it just didn't grab me. I think that I have subconscious expectations (from my culture) when it comes to animation that are just too hard to get over. Objectively, I can look at an anime or manga and say that it has certain qualities[1], and sometimes qualities that I'll look for in other mediums, but when I sit down to watch/read, I just tend to get put off.
[1] And more importantly, I can say that just because a story is presented in a certain medium doesn't mean that it's marketed to a specific age group. Contrast this to people who think that all comics are for kids, and thus comics with excessive sex and violence are clearly trying to corrupt our youth. I can objectively realize this, but that doesn't mean that subconsciously, when I sit down to watch it, I'm not put off by the perceived conflation of adult themes and "cartoony" rendering. It's a personal failing, but at least I don't try to impose my preconceptions upon society.
I believe that the Kindle allows for annotation by the reader. I don't know if it's written (with a stylus) or typed, but it's definitely one of the features that I think they got right.
That said, it may just be that the Kindle isn't for me, but other people will flock to it. We'll see, I guess.
Before your comment on O'Reilly books, I was really giving the Kindle a very "meh" response. It looks like Amazon is trying to follow the Apple iTMS+iPod model, but they've overlooked a few things.
The most notable thing they've overlooked is that books are not like CDs. iPods work well because they don't deviate all that much from the previous portable music models, and where they do deviate, they make noticeable improvements. With the iPod, for example, you still listen to the music through headphones. The primary listening interface has not changed at all. However with the Kindle, they've changed the primary reading interface to a screen. Now that might not seem too important, but it's going to turn a lot of people off right away. Then you've got things like flipping through a book--it's really hard to recreate this feeling with a device like this, and though I haven't played with a Kindle, I find it hard to imagine that they could have succeeded. Then, of course, there's portability. While the Kindle would be useful on my desk or coffee-table, I'm not going to carry one around with me. For reading novels, then, I'm basically no better off than if I had just bought the novel, since I'm probably never going to want to carry a dozen novels around with me at all times (unlike music.) And with the novel, I can choose to loan it to a friend. I can't do that with the Kindle without loaning them all of my novels, and depriving myself of them (though this is an issue with any DRM, including iTunes, it's especially applicable to my circle of friends, who loan each other novels all of the time, but rarely loan music.)
I look at this device and all I can see is the number of novels I could buy for $400 instead, have a better reading experience, and have them separate from each other. A specialized solution for O'Reilly books (like the Safari "bookshelf" in portable form) would be very nice to be sure, but I don't know that I could justify $400+book costs for it.
Not at all. The malicious FLAC starts up a daemon that sits and waits for you to use su/sudo, then it performs its ptrace injection. Maybe it modifies your bash_login file so that it runs every time you log in. It could even probably spoof an update-manager window, making you think that there are updates to install (which will then prompt you to enter your password.)
That's the thing about security. If the OS allows the user full control, then your security implementation is only as good as your user is knowledgeable. If you moved most Windows users over to Linux, it wouldn't take a week for malware authors to start spoofing gksudo windows.
Actually, the admin account on Vista doesn't have to enter a password--they just have to click through the UAC control. If you try to perform an administrative action as a regular user in Vista, you'll have to enter a password as part of the UAC control.
I guess that it depends upon the definition of "end users." Am I an "end user" of gcc, even though I've never compiled anything in my life? After all, I'm using software that was compiled with gcc.
It's the same thing with any web service software and the AGPL. If I download web service software and install it to my server, I feel as though I am the user of the software, and my web visitors are users of my site (which is the output of the software--specifically excluded from being a derivative work under the GPL, but not the AGPL in the specific case of web services.)
Put more succinctly, the GPL never inhibits use of software (that is, running software.) The AGPL does. That's why the AGPL applies to the users of software (the people running the actual software), not just distributors.
The theory is the same as with any EULA (the AGPL is truly a EULA, unlike the GPL.) The AGPL allows you to use the code if you give the code back to any of your users. If you refuse to do that, you are violating the license, and thus you have no right to use the software.
It's the same thing as Microsoft saying that by using their software, you absolve them of any wrongdoing. It's part of the license. If you try to sue them, you're in violation, and thus never had a right to use the software in the first place.
This is why EULAs are bad. I'm very sad to see the FSF going to one, though maybe it will drive people to more open licenses (like BSD.)
So that file doesn't count as a larger work that then must be distributed under the AGPL. If so, then the AGPL is worthless. I can just include my changes to the AGPL project and distribute a clean version of the source, but with the "include \"mystuff.php\"" and a few new function calls. My changes don't go out (they were just included!)
The whole thing is sticky and not nearly as well thought out as the GPL.
If you think that GPLv3 pushed the masses away from using Linux, you may want to seek psychiatric help.
It is a EULA, which is a great departure from the GPL of old.
Of course, web applications break the model. There are two users--the host is using the application to interact with the user, and the user is using the web application to interact with the host. It's a really weird situation. That said, I kinda like the separation of the old days, because frankly, where does this end? Should the user get a copy of Apache just for visiting a website? If the website interacts with a database, should they get a copy of the database? What if the database is proprietary, but the front-end (web server/php server) is AGPL?
This really muddies the waters. Services traditionally haven't been touched by copyleft, and if this thing catches on, it will have some very interesting (and probably unfortunate) ramifications.