Yes but do you have proof about your discs getting stolen in the mail? Probably not. Your mail getting stolen from your building, well, that's a different story.
You are missing the point. If they were getting stolen "in the mail" (i.e., at the post office), handing off delivery to this kind of service wouldn't help. But it is precisely because mail gets stolen "from [his] building" that delivering to a different building helps.
Sure there's a possibility of it getting picked up out of the box, but we usually have someone home and as soon as it gets there, my wife gets it.
Well, how nice for you. But people who travel around the world don't have wives sitting at home snatching the mail out of the hands of thieves. That's why it helps them with identity theft: their mail isn't sitting in some unobserved mailbox for months at a time.
I don't feel comfortable letting some mailroom dude scan my mail because he has to open it first. I don't care if the POPE is running the company, I still don't trust it.
As opposed to the "mailroom dudes" that put it in the envelope? And the USPS dudes who deliver it? And the random dudes that just walk by? At least with a company, the set of people who can look at your mail is well-defined. (Oh, and I wouldn't trust the pope with my mail either, but that's an entirely different matter.)
There is actually an entire movement of people that have discovered this. Look for "simple living" on Google. "The Simple Living Guide" by Luhrs is pretty nice reading, too.
Even if you don't want to adopt frugality and simple living right now, just knowing that you could can make you worry a lot less about the future.
So, your point amounts to saying that Microsoft can buy lots of prestige and people in countries that are not economically well-off. True, but that doesn't make building a nation's computational infrastructure on top of the products of a single foreign vendor a good choice. And while Microsoft makes lots of money, it employs comparatively few people, so most IT workers still end up empty-handed.
It isn't actually that professors put their notes on-line. Many schools even put more lectures in digital video formats on-line (unless I missed the links, only a few of the OCW lectures appear to be available in video).
What seems to be new about this is that MIT has hired staff to put together a professional-looking, organized website. This means that putting a course on OCW is less work for the professors and that the end-result is more useful to students. That's in contrast to many other universities that try to "protect" their content from the outside world and give little or no assistance to their professors in putting course content on-line.
Beach parties, office parties, slumber parties, bachelor parties. Darl has to make up for a lifetime of repression from his church, and with his millions of ill-gotten gains, he finally can.
MS, according to the latest 10-Q (available at SEC), has apparently purchased those "expanded licensing options" that were mentioned in the April 30 10-Q.
Does that then mean MS has expanded its in-house use of Linux?:-)
The reason that corporations don't just go into space is because there is a mind boggling overlapse of governmental agency jurisdiction over the necessary elements of space flight (DoT, DoC, NASA, FAA, etc). No-one can get permission to -try- let alone to work on getting the price down.
The US does not have jurisdiction about what is launched into space from other nations, neither legally nor even in any practically useful sense. I'm sure the Russians, the Chinese, Central American nations, and many African nations would be happy to make it pretty easy for any corporation who wants to to launch whatever they like into space from their territory. I think it's just a lack of demand.
I can't quite tell from a quick reading of the paper, but this seems to be a user-mode file system. That is, if you call the regular POSIX "open" call, you probably can't open a file in the GoogleFS. It appears that some library code linked directly into the application handles all file system operations. A number of distributed file systems take that approach--it can be more efficient.
I wonder how it compares to PVFS. It seems like GoogleFS deals more aggressively with component failure. Any ideas?
I have three words for you... accurate colour management.
Color management is not primarily a function of the photo editing software you use, it's a separate part of the operating system. In fact, it's not clear that photo editing software should know anything about color management at all. Photoshop's integration of ICC is the usual bloated kitchen-sink approach of Windows/Mac programs.
In any case, there are several color management solutions for Linux, and at least one of them works with the Gimp.
(Personally, I think you should start working with images yourself before passing comment on my technique or workflow.)
I have worked professionally with digital images for 20 years--far longer than Photoshop has even been around. Maybe you should start learning a bit more about digital imaging beyond Photoshop yourself before accusing people of being non-professionals just because they don't use Photoshop.
This one software package is single-handedly keeping me from migrating to Linux. For those who say "But what about Gimp? It's just as good..." Those people have also never done professional graphics for print, video or even the web.
Bullshit. Use whatever you like, but stop holding up you particular tool as a mark of professionalism. You are like those people who buy Hasselblads and think they are professional photographers because of it.
Personally, I think anybody working with images who actually needs more than what the Gimp provides either has bad photographic technique or a bad workflow. If anything, the Gimp actually still is overkill for day-to-day imaging work.
the capsule replacement was always intented purely to support a low earth orbit space station. a space station that congress didn't want to build.
Since neither Congress nor the scientific community wanted the space station to be built, why was it built in the first place?
The best solution is for space to become privatized.
There is nothing to be "privatized". If you have the money, you can start launching people into space right away. The reason the private sector isn't doing it on its own is because it's futile and pointless and a waste of money. If you are saying that we should stop that kind of waste by handing out tax payer money to Boeing and the like, I'm all for it.
Research in space, on the other hand, is a public endeavor. But research does not require manned space exploration (at least not at this point and not for the foreseeable future). Unmanned probes, satellites, etc., do require public funding and are money well-spent.
one should ask the question of "what's a good MUD"? I'm not trying to be cynical here, I just have never found a MUD that I wanted to spend any significant amount of time in (well, unless you consider/. a MUD).
What product for current versions of Windows are you referring to that offers disk at a time encryption.
Do a search on Google; there is plenty of software around.
Note that that means being able to operate from an encrypted boot drive, not just being able to take a big file, call it a volume, and have it be encrypted.
I fail to see the difference. You can move most of the OS to an encrypted drive with Windows just like you would with Linux or BSD.
Of course, there are also hardware-encrypted drives whose encryption is independent of the OS. With those, you can indeed boot the entire OS (Windows, Linux, BSD, whatever) from the encrypted drive.
That argument strikes me as silly. Even if you used no encryption at all, you couldn't prove that you haven't hidden data somewhere. In fact, you can hide data on a seemingly completely blank disk.
Who said anything about PC Magazine? Apple's own benchmarks show that their fastest G5 is only moderately faster than a 3GHz Xeon and about equal to Opteron.
One can argue which is slightly faster or slightly slower, but the G5 clearly does not give you breakthrough performance. Furthermore, in terms of bang-for-the-buck, G5 systems are worse than Pentium or Opteron.
Sometimes one has to wonder whether Apple even knows what's going on. I mean, all their bogus claims about the "first 64bit personal computer" and their "desktop supercomputer" claims. The company lives on hype.
StegFS uses encryption to support deniability; the encryption is a tool, not an end in itself, and is certainly not the option (performance wise) if all you want is data confidentiality.
Deniability itself greatly increases security, even if there is nobody with a rubber hose standing over you. And it is no accident that deniability requires strong cryptography: deniability is a stronger property, not merely a different property, than encryption.
I'm saying that the behavior is too complex and not transparent enough if provided at the application level.
And both you and Matt are right at that. But "application level" has nothing to do with "user mode". As I was saying, you are confusing "user mode" and "application level". "User mode" means that the code doesn't run in kernel mode. "Application level" means that applications need to know about it. The two are distinct concepts.
The filesystem is part of the kernel, and any encryption should be transparent to a process using the file.
Yes, but not everything that "should be transparent to a process using the file" needs to run in kernel mode. This knee-jerk reaction of putting everything that needs to be "transparent" into the kernel is responsible for kernel bloat, both on Linux and on other systems.
Plan 9, PlasticFS, and CODA show that you can put large parts of the file system code into user mode and yet retain transparency. Look them up, really!
While what you say sounds really GREAT, try implementing it at a company with over 100,000 nodes soon growing to 300000+.
Well-run corporate networks (and I we have one at work) don't have problems with this: you have internal firewalls, internal security scans, internal security audits, and decentralized management (i.e., someone feels responsible for every machine). Only network connections that are related to the purpose of a machine are actually permitted, and machines that misbehave are detected and can be isolated instantaneously.
If anything, that sort of thing is easier to implement on a large network because the initial, fixed costs of setting it up are spread over more machines. Of course, the IT staff needs to have a clue, which, sadly, IT staffs at many companies don't.
And the larger the network gets, the more important it is. With 300000+ machines, it's more likely that the tooth fairy exists than that all of them will remain virus-free.
We all know that Intel is going to come out with something faster and better than the current P4. But what matters is what is shipping, not rumors companies place with tech column writers in order to scare people away from their competitors. Apple made a lot of noise about the G5, which ended up basically just keeping up with the x86 world in terms of speed, and Intel made a lot of noise about the P4 and that was a disappointment, too. When the P5 ships and how well it will perform remains to be seen.
Let it go. Linux is big enough. The primary problem with Windows is, in fact, that it tries to be everything to everybody. Let's not have Linux fall into that same trap.
If you don't see the benefits of Linux and want to pay Bill Gates for Windows 2003, go right ahead--I really don't care. But I certainly don't want Linux to become any more like Windows 2003 in order to grab a little more marketshare.
If you are irrationally obsessed with driving Microsoft out of business through open source, I think a much better competitor for Windows 2003 than Linux would be ReactOS running Mono/.NET. The Windows nerds can wallow in all that GDI+, MFC,.NET goodness, they can click on dialog boxes all day long, and us UNIX/Linux users don't have to live through that.
There really isn't a single right answer for everybody. Let's not make the same mistake Gates is making. We need a lot more variety in operating systems. Linux taking over everything would be just as bad as Windows taking over everything.
To be fair, Carbon and Cocoa applications use the same standard widgets,
I don't think so; maybe they share the same graphics code but they certainly don't behave the same.
In fact, I've found that Mac OS X is the most consistent desktop environment I've used.
I think you made my point for me: OS X has a completely disparate set of underlying toolkits and technologies. Yet, people like you believe the UI to be highly consistent. It doesn't even sway your opinion that some toolkits put menus at the top of the screen while others put them inside windows. Obviously, enforcing the use of a single toolkit is not necessary to achieve consistency (and as Apple's non-standard shiny metal applications show, it isn't sufficient either).
Unlike Macintosh, however, if you run something like KDE on X11, you really do get a single toolkit and a completely consistent user interface. And unlike Macintosh, where you really are forced to run a mix of toolkits and GUI styles in order to get all the functionality you need, an environment like KDE really does implement everything you need using a single toolkit. Frankly, I slightly prefer the Macintosh UI to KDE, but that is in spite of its inconsistency and use of many toolkits.
While it is certainly possible to easily implement file encryption at the user/application layer, I disagree that it should be. Matt Blaze pointed out a number of reasons why in his CFS paper back in 1993.
You are confusing implementation at the application layer with implementation in user mode. The crypto-code need not run in kernel mode in order to give users the same behavior and security as a kernel-based implementation; the kernel just needs to provide the right hooks.
besides, the goal of stegFS isn't necessarily to support encryption; it is meant to support plausible deniability of file ownership, and those two goals are very different.
StegFS supports encryption and deniability. Its approach to deniability also happens to greatly increase security.
After the returned to work it spread across the unpatched systems and caused so much network traffic that everything was down for days (some areas didn't have IT on sight to clean up the problems). Really makes you think just how vunerable you are to these.
Yes, and if you ponder that a little longer, you'll see that the way to deal with that is to make your systems robust against virus attacks from the inside, no matter whether the machines are up-to-date and no matter who plugs in where. With halfway modern networking hardware, that is not difficult to do.
I actually cut a CEO's network cable in half (in front of him and his just-about-to-faint secretary) for doing something quite similar.
If you can't deal with a CEO plugging his virus-infected laptop into your network, that only goes to show that your internal security and antivirus measures suck. Your network won't be secure and reliable unless you can prevent virus infections from spreading internally.
X11 toolkits are pluggable--you plug whatever toolkit you want into the client.
If you put anything "pluggable" into the server, you need an architecture-independent runtime for it and you end up with DisplayPostscript, NeWS, or Java. Is that what you want? Then go right ahead and use them: that's what they are there for.
Slashdot Mirror
Given how much bad stuff there is in paper and ink, that is cruel. Don't give your hamster printed materials to chew on.
Yes but do you have proof about your discs getting stolen in the mail? Probably not. Your mail getting stolen from your building, well, that's a different story.
You are missing the point. If they were getting stolen "in the mail" (i.e., at the post office), handing off delivery to this kind of service wouldn't help. But it is precisely because mail gets stolen "from [his] building" that delivering to a different building helps.
Sure there's a possibility of it getting picked up out of the box, but we usually have someone home and as soon as it gets there, my wife gets it.
Well, how nice for you. But people who travel around the world don't have wives sitting at home snatching the mail out of the hands of thieves. That's why it helps them with identity theft: their mail isn't sitting in some unobserved mailbox for months at a time.
I don't feel comfortable letting some mailroom dude scan my mail because he has to open it first. I don't care if the POPE is running the company, I still don't trust it.
As opposed to the "mailroom dudes" that put it in the envelope? And the USPS dudes who deliver it? And the random dudes that just walk by? At least with a company, the set of people who can look at your mail is well-defined. (Oh, and I wouldn't trust the pope with my mail either, but that's an entirely different matter.)
There is actually an entire movement of people that have discovered this. Look for "simple living" on Google. "The Simple Living Guide" by Luhrs is pretty nice reading, too.
Even if you don't want to adopt frugality and simple living right now, just knowing that you could can make you worry a lot less about the future.
So, your point amounts to saying that Microsoft can buy lots of prestige and people in countries that are not economically well-off. True, but that doesn't make building a nation's computational infrastructure on top of the products of a single foreign vendor a good choice. And while Microsoft makes lots of money, it employs comparatively few people, so most IT workers still end up empty-handed.
It isn't actually that professors put their notes on-line. Many schools even put more lectures in digital video formats on-line (unless I missed the links, only a few of the OCW lectures appear to be available in video).
What seems to be new about this is that MIT has hired staff to put together a professional-looking, organized website. This means that putting a course on OCW is less work for the professors and that the end-result is more useful to students. That's in contrast to many other universities that try to "protect" their content from the outside world and give little or no assistance to their professors in putting course content on-line.
Beach parties, office parties, slumber parties, bachelor parties. Darl has to make up for a lifetime of repression from his church, and with his millions of ill-gotten gains, he finally can.
MS, according to the latest 10-Q (available at SEC), has apparently purchased those "expanded licensing options" that were mentioned in the April 30 10-Q.
:-)
Does that then mean MS has expanded its in-house use of Linux?
The reason that corporations don't just go into space is because there is a mind boggling overlapse of governmental agency jurisdiction over the necessary elements of space flight (DoT, DoC, NASA, FAA, etc). No-one can get permission to -try- let alone to work on getting the price down.
The US does not have jurisdiction about what is launched into space from other nations, neither legally nor even in any practically useful sense. I'm sure the Russians, the Chinese, Central American nations, and many African nations would be happy to make it pretty easy for any corporation who wants to to launch whatever they like into space from their territory. I think it's just a lack of demand.
I can't quite tell from a quick reading of the paper, but this seems to be a user-mode file system. That is, if you call the regular POSIX "open" call, you probably can't open a file in the GoogleFS. It appears that some library code linked directly into the application handles all file system operations. A number of distributed file systems take that approach--it can be more efficient.
I wonder how it compares to PVFS. It seems like GoogleFS deals more aggressively with component failure. Any ideas?
I have three words for you ... accurate colour management.
Color management is not primarily a function of the photo editing software you use, it's a separate part of the operating system. In fact, it's not clear that photo editing software should know anything about color management at all. Photoshop's integration of ICC is the usual bloated kitchen-sink approach of Windows/Mac programs.
In any case, there are several color management solutions for Linux, and at least one of them works with the Gimp.
(Personally, I think you should start working with images yourself before passing comment on my technique or workflow.)
I have worked professionally with digital images for 20 years--far longer than Photoshop has even been around. Maybe you should start learning a bit more about digital imaging beyond Photoshop yourself before accusing people of being non-professionals just because they don't use Photoshop.
This one software package is single-handedly keeping me from migrating to Linux. For those who say "But what about Gimp? It's just as good..." Those people have also never done professional graphics for print, video or even the web.
Bullshit. Use whatever you like, but stop holding up you particular tool as a mark of professionalism. You are like those people who buy Hasselblads and think they are professional photographers because of it.
Personally, I think anybody working with images who actually needs more than what the Gimp provides either has bad photographic technique or a bad workflow. If anything, the Gimp actually still is overkill for day-to-day imaging work.
the capsule replacement was always intented purely to support a low earth orbit space station. a space station that congress didn't want to build.
Since neither Congress nor the scientific community wanted the space station to be built, why was it built in the first place?
The best solution is for space to become privatized.
There is nothing to be "privatized". If you have the money, you can start launching people into space right away. The reason the private sector isn't doing it on its own is because it's futile and pointless and a waste of money. If you are saying that we should stop that kind of waste by handing out tax payer money to Boeing and the like, I'm all for it.
Research in space, on the other hand, is a public endeavor. But research does not require manned space exploration (at least not at this point and not for the foreseeable future). Unmanned probes, satellites, etc., do require public funding and are money well-spent.
one should ask the question of "what's a good MUD"? I'm not trying to be cynical here, I just have never found a MUD that I wanted to spend any significant amount of time in (well, unless you consider /. a MUD).
What product for current versions of Windows are you referring to that offers disk at a time encryption.
Do a search on Google; there is plenty of software around.
Note that that means being able to operate from an encrypted boot drive, not just being able to take a big file, call it a volume, and have it be encrypted.
I fail to see the difference. You can move most of the OS to an encrypted drive with Windows just like you would with Linux or BSD.
Of course, there are also hardware-encrypted drives whose encryption is independent of the OS. With those, you can indeed boot the entire OS (Windows, Linux, BSD, whatever) from the encrypted drive.
That argument strikes me as silly. Even if you used no encryption at all, you couldn't prove that you haven't hidden data somewhere. In fact, you can hide data on a seemingly completely blank disk.
Who said anything about PC Magazine? Apple's own benchmarks show that their fastest G5 is only moderately faster than a 3GHz Xeon and about equal to Opteron.
One can argue which is slightly faster or slightly slower, but the G5 clearly does not give you breakthrough performance. Furthermore, in terms of bang-for-the-buck, G5 systems are worse than Pentium or Opteron.
Sometimes one has to wonder whether Apple even knows what's going on. I mean, all their bogus claims about the "first 64bit personal computer" and their "desktop supercomputer" claims. The company lives on hype.
StegFS uses encryption to support deniability; the encryption is a tool, not an end in itself, and is certainly not the option (performance wise) if all you want is data confidentiality.
Deniability itself greatly increases security, even if there is nobody with a rubber hose standing over you. And it is no accident that deniability requires strong cryptography: deniability is a stronger property, not merely a different property, than encryption.
I'm saying that the behavior is too complex and not transparent enough if provided at the application level.
And both you and Matt are right at that. But "application level" has nothing to do with "user mode". As I was saying, you are confusing "user mode" and "application level". "User mode" means that the code doesn't run in kernel mode. "Application level" means that applications need to know about it. The two are distinct concepts.
The filesystem is part of the kernel, and any encryption should be transparent to a process using the file.
Yes, but not everything that "should be transparent to a process using the file" needs to run in kernel mode. This knee-jerk reaction of putting everything that needs to be "transparent" into the kernel is responsible for kernel bloat, both on Linux and on other systems.
Plan 9, PlasticFS, and CODA show that you can put large parts of the file system code into user mode and yet retain transparency. Look them up, really!
While what you say sounds really GREAT, try implementing it at a company with over 100,000 nodes soon growing to 300000+.
Well-run corporate networks (and I we have one at work) don't have problems with this: you have internal firewalls, internal security scans, internal security audits, and decentralized management (i.e., someone feels responsible for every machine). Only network connections that are related to the purpose of a machine are actually permitted, and machines that misbehave are detected and can be isolated instantaneously.
If anything, that sort of thing is easier to implement on a large network because the initial, fixed costs of setting it up are spread over more machines. Of course, the IT staff needs to have a clue, which, sadly, IT staffs at many companies don't.
And the larger the network gets, the more important it is. With 300000+ machines, it's more likely that the tooth fairy exists than that all of them will remain virus-free.
We all know that Intel is going to come out with something faster and better than the current P4. But what matters is what is shipping, not rumors companies place with tech column writers in order to scare people away from their competitors. Apple made a lot of noise about the G5, which ended up basically just keeping up with the x86 world in terms of speed, and Intel made a lot of noise about the P4 and that was a disappointment, too. When the P5 ships and how well it will perform remains to be seen.
Let it go. Linux is big enough. The primary problem with Windows is, in fact, that it tries to be everything to everybody. Let's not have Linux fall into that same trap.
.NET goodness, they can click on dialog boxes all day long, and us UNIX/Linux users don't have to live through that.
If you don't see the benefits of Linux and want to pay Bill Gates for Windows 2003, go right ahead--I really don't care. But I certainly don't want Linux to become any more like Windows 2003 in order to grab a little more marketshare.
If you are irrationally obsessed with driving Microsoft out of business through open source, I think a much better competitor for Windows 2003 than Linux would be ReactOS running Mono/.NET. The Windows nerds can wallow in all that GDI+, MFC,
There really isn't a single right answer for everybody. Let's not make the same mistake Gates is making. We need a lot more variety in operating systems. Linux taking over everything would be just as bad as Windows taking over everything.
To be fair, Carbon and Cocoa applications use the same standard widgets,
I don't think so; maybe they share the same graphics code but they certainly don't behave the same.
In fact, I've found that Mac OS X is the most consistent desktop environment I've used.
I think you made my point for me: OS X has a completely disparate set of underlying toolkits and technologies. Yet, people like you believe the UI to be highly consistent. It doesn't even sway your opinion that some toolkits put menus at the top of the screen while others put them inside windows. Obviously, enforcing the use of a single toolkit is not necessary to achieve consistency (and as Apple's non-standard shiny metal applications show, it isn't sufficient either).
Unlike Macintosh, however, if you run something like KDE on X11, you really do get a single toolkit and a completely consistent user interface. And unlike Macintosh, where you really are forced to run a mix of toolkits and GUI styles in order to get all the functionality you need, an environment like KDE really does implement everything you need using a single toolkit. Frankly, I slightly prefer the Macintosh UI to KDE, but that is in spite of its inconsistency and use of many toolkits.
While it is certainly possible to easily implement file encryption at the user/application layer, I disagree that it should be. Matt Blaze pointed out a number of reasons why in his CFS paper back in 1993.
You are confusing implementation at the application layer with implementation in user mode. The crypto-code need not run in kernel mode in order to give users the same behavior and security as a kernel-based implementation; the kernel just needs to provide the right hooks.
besides, the goal of stegFS isn't necessarily to support encryption; it is meant to support plausible deniability of file ownership, and those two goals are very different.
StegFS supports encryption and deniability. Its approach to deniability also happens to greatly increase security.
After the returned to work it spread across the unpatched systems and caused so much network traffic that everything was down for days (some areas didn't have IT on sight to clean up the problems). Really makes you think just how vunerable you are to these.
Yes, and if you ponder that a little longer, you'll see that the way to deal with that is to make your systems robust against virus attacks from the inside, no matter whether the machines are up-to-date and no matter who plugs in where. With halfway modern networking hardware, that is not difficult to do.
I actually cut a CEO's network cable in half (in front of him and his just-about-to-faint secretary) for doing something quite similar.
If you can't deal with a CEO plugging his virus-infected laptop into your network, that only goes to show that your internal security and antivirus measures suck. Your network won't be secure and reliable unless you can prevent virus infections from spreading internally.
X11 toolkits are pluggable--you plug whatever toolkit you want into the client.
If you put anything "pluggable" into the server, you need an architecture-independent runtime for it and you end up with DisplayPostscript, NeWS, or Java. Is that what you want? Then go right ahead and use them: that's what they are there for.