As a user, I see it as a problem if I have to modify the OS.
So do I. I'm sure that the Hurd people will do reasonable stuff by default. You should be able to run a default Debian GNU/Hurd install just as you would a default Debian GNU/Linux install (minus the packages that are Linux-only, of course).
And Linux lets you install loadable modules on the fly these days, so it's not like anyone has to recompile their kernel to use a new printer or something.
Sure, but you need to be root to install modules. Anyways, the whole point of it is in giving users choice. If some part of the OS doesn't work the way you want it to, you can replace it with something else. You don't necessarily need to program it yourself. You can use someone else's code. Some other poster in this article mentioned that the recent Linux issues about the VM wouldn't have been as big a deal under Hurd, since if you wanted, you could easily swap Rik's VM with Andrea's VM.
I would expect that most of the time, the Hurd would be used just like UNIX. But sometimes it's nice to have the extra flexibility. Of course, whether or not it's worth it is up to the admin/computer owner.
I guess I just don't see what I'm missing with MacOS X and Linux. In MacOS X, for instance, I can mount a CD or a filesystem shared by another machine, and I can do it without being root.
Well, you definitely can't have that amount of control in Linux. Unless there's already an entry in/etc/fstab, normal users cannot mount a new filesystem.
I've never used OS X, so I can't say anything about that. But how easy would it be to add support for a new filesystem (without being root)? Assuming that there is a way to add new filesystems at all. How easy is it to mount a loopback filesystem?
BTW, there are other advantages to the Hurd than just the multiple server thing. For example, the authentication mechanism allows a daemon (take telnet or ssh as an example) to start of running as no user (not to be confused with a "nobody" user) and, when it gets appropriate authentication tokens, to upgrade its permissions. This is in contrast to traditional UNIX-like OS's, where telnet and ssh need to run as root, and drop you to your own user when you log in. Hurd's method effectively prevents remote root exploits (short of guessing root's (or someone else's) password). (Actually, that may also be related to the multi-server nature of Hurd. I don't know the Hurd well enough to say.)
The last official release of the Hurd is 0.2. It has made a lot of progress since then, but no one has had the urge (yet) to declare a new version number. Hurd development is mostly done in conjunction with Debian now, so the next Hurd release will probably be named the same as the next Debian release (though I may be wrong about that).
The Hurd is aparently quite usable already. (I don't use it, so I can't give any first hand information.) XFree86 and a lot of other software already runs on it. One notable exception is GNOME, because they don't have a working POSIX threads library.
If you want to try it out, look at http://www.debian.org/ports/hurd/. Working with the Debian base, they had over half (I forget the exact number) the packages compiled and usable about two months ago. Since then, they've been working on an ABI change, so they had to purge all their old packages, but they're working on recompiling everything again.
As for working with Apple's Mach kernel, they probably could, but it would require work to port the system over. They're actually moving away from the Mach microkernel, and towards the L4 microkernel, so they probably wouldn't want to port it to Apple's Mach kernel. It may not even be good enough for the Hurd either. The Hurd takes a multi-server approach, while OS X is still a monolithic design (AFAIK). When you do multi-server, the speed of interprocess communication becomes critical, whereas it is not so important under a monolithic design.
Re HURD's many-server architechture, ok, sounds interesting, but why is it a good thing? What does it do for programmers or end-users?
Look through the "Introductory Material" at http://www.gnu.org/software/hurd/docs.html if you really want to read up on it.
The short answer is that it allows you to replace parts of the operating system easily. For programmers (OS developers in particular), you can do your own implementation of part of the OS and test it out, without disrupting any of the other users. If your bit crashes, no problem, just replace it with what was there before. For users, it allows you to do things like mount filesystems without being root or having to bug root to frob/etc/fstab, which you want to do if you want to work with an ISO image. Of course, there are other things that normal users can do with the Hurd, but this is probably the biggest.
My point was that tweakt was misunderstanding what RMS was talking about. But even so, this doesn't mean that the Hurd is "good in theory, bad in practice." It's just than nobody has bothered to code up that stuff yet. The Hurd is able to support whatever hardware you want to throw at it. You just need to code up support for it, if the Hurd doesn't already support it. Just like Linux, and any other operating system that you want to use.
I'll admit that I'm not an OS guru, so most of what I say is second hand, but you should look at Thomas Bushnell's paper at http://www.gnu.org/software/hurd/hurd-paper.html, which should give you and idea of how the Hurd and UNIX are different. If you're too lazy to read through the whole thing, jump down to "Why this is so different" or thereabouts. The beginning of part 1 is actually pretty interesting too, and is worth reading if you want to see how the Hurd differs from UNIX.
But why would preventing a user from mounting a loopback filesystem be a policy decision? The file is already owned by the user, so he/she can already do whatever he/she wants with it. He/she can make his/her own versions of cp/ls/etc. à la mtools to read from and write to the filesystem. Preventing the user from mounting doesn't do anything from a security standpoint, and just makes it harder for the user to do useful work.
You obviously don't know much about the architecture of the Hurd. Go read up on it (http://hurd.gnu.org/) and come back.
The architecture of the Hurd (not to be confused with the implementation) gives users a lot more freedom than any UNIX-based system. For example, UNIX will not let you mount a loopback encrypted filesystem unless you are root (or without bugging root to frob/etc/fstab, which he/she probably won't want to do), even if the encrypted file is owned by you, in your own directory and you want to mount it within your own home directory. This is something that the architecture prevents you from doing, so no UNIX implementation will ever let you do that (without a heavy dose of magic). Allowing stuff like that is one of the architectural features of the Hurd.
Porting, and supporting 56k modems are just implementation details, and have nothing to do with the architecture, which is what RMS is talking about.
Whoa. If you're going to correct someone, at least get your facts straight. Dvorak wasn't created until the 1920s/1930s. The Dvorak was created to increase typing speed, rather than to decrease it, by putting the most commonly used letters on the home row. I forget the exact number, but It's something over 80% of the letters typed in a normal English text would be on the home row.
I have no idea where the author of the article got that idea. It's the stupidest thing I've heard for a long while. If you look around the Fastap site, you'll find that you do actually hit the recessed number, and it manages to ignore hits that may be generated by accidentally hitting one of the surrounding letters.
Incidentally, it looks like they also have a credit-card sized QWERTY keyboard, with numbers and symbols in between the letters. That might be interesting to stick onto a PDA.
A Slashdot effect would not result in a missing index.* file, and if it does, even more reason to avoid IIS. A server can choke due to high traffic, but if it ever deletes your data, stay away from it. It might've been hacked, but if it was, whoever hacked it should have replaced the index.* file with something else, rather than just deleting it. Another possibility is that they hired some clueless admins, because they couldn't afford expensive experts.
I believe what you are talking about is already being done, through dual licensing. See, for example, the licensing for FFTW (www.fftw.org - an FFT library developed by some MIT guys). It's licensed under the GPL, and if you don't want the GPL (e.g. you want to include it into a closed-source program), you can sign a licensing deal with MIT.
ReiserFS also does the same thing, and Hans has mentioned before that RMS hasn't complained about it yet.
I agree. I love docbook. If you can write HTML, chances are you can do docbook. It's plain text, so you can edit it in whatever program you want to.
It's a bit of a pain to set up, but once it's going, it's great (much like most of Linux, I find;-) ). I use Jade -- openjade seems to be slower, and I couldn't see any advantage of openjade over Jade.
My only complaint about docbook is that it (currently) doesn't do math too well -- I use Walsh's stylesheets, and they don't seem to grok MathML yet -- so I have to stick with LaTeX for some things still. Oh, and its table support doesn't seem to be complete when using the TeX backend, so things may come out in unexpected ways.
Furthermore if we improve the experiment and let the array be of
random size...
Of course, you need to define a distribution for that to make any sense at all. Plus, the priors are unknown. In order to get anything meaningful out of Bayesian analysis, you need to know the priors, and we obviously don't.
Next, you assume that vast technical superiority = many more inhabitants. Simply not true.
The above argument also applies to your age.
Hmmm. Did you read the FAQ from the site? Question 10.
The output from a pseudo-random number generator is usually considered "random enough for practical purposes." So if you define "practically random data" as "data that is random enough for practical purposes," you can compress it by storing the random seed and the string length.;-)
I think I can beat their 100:1 compression ratio with this scheme.
Sure. Windows NT 5.0 will never be released.;-) (Unless they decide to confuse everyone and change the numbering scheme back, which we all know will never happen...)
As their name implies, Foreverbright lights are rated for up to 200,000 hours of use - or almost 23 years.
Sounds like they changed the definition of "forever" while I wasn't paying attention.
The principle behind LEDs, while complicated, can be made easy to understand. An LED is a special type of diode and is similar to a transistor.
And then it goes on to talk about N and P silicon, and so forth. I guess they changed the definition of "easy to understand" too.
Re:I need this like I need colonic irrigation
on
Galeon 1.0 Released
·
· Score: 4, Informative
Galeon's home page uses DIV and SPAN tags everywhere. I get one column about an inch wide on the left with all text and images. My CPU sounds like it's about to puke. Turn JS off, and I get a 1994-style page with gray background. None of the web pages I create have this problem, but I'm not trying to awe people with my mastery of Dreamweaver. If it can't be done in Vi or Emacs, it's not worth doing.
It's because of NS4's buggy CSS rendering. If you turn JS off in NS4, it also turns off CSS handling (silly Netscape).
The Galeon page is nothing fancy. It's just using normal HTML -- no JavaScript required. If they did their HTML properly, the page was meant to degrade gracefully. That is, if you view it in a browser that correctly adheres to the standards that it claims to, it should be usable. Their HTML looks pretty simple, and I wouldn't be surprised if they did do it in vi or emacs.
The problem that you have is that NS4 doesn't do CSS correctly, but it likes to pretend that it does. It's a problem that many web designers face, and often the decision is: "Screw Netscape" because NS4's CSS handling can be quite unpredictable.
Why would people use CSS? Ironically, because it's supposed to allow them to create a nice-looking page that will be usable in older browsers. It's unfortunate that Netscape had to screw it up.
Re:I need this like I need colonic irrigation
on
Galeon 1.0 Released
·
· Score: 1
If you want something that's fast, use Skipstone. It has less features than Galeon, but it's much faster. Particularly on startup.
But you probably want to give Galeon a try anyway.
Two problems (that haven't been mentioned yet): firstly, you now have the problem that the program needs to know what the piece of clay looked like initially. Unless you can always start with a perfectly spherical ball of clay... Secondly, you would need to model the physics of the clay, such as sagging. For all that trouble, you might as well just go with force feedback.
Slashdot Mirror
So do I. I'm sure that the Hurd people will do reasonable stuff by default. You should be able to run a default Debian GNU/Hurd install just as you would a default Debian GNU/Linux install (minus the packages that are Linux-only, of course).
Sure, but you need to be root to install modules. Anyways, the whole point of it is in giving users choice. If some part of the OS doesn't work the way you want it to, you can replace it with something else. You don't necessarily need to program it yourself. You can use someone else's code. Some other poster in this article mentioned that the recent Linux issues about the VM wouldn't have been as big a deal under Hurd, since if you wanted, you could easily swap Rik's VM with Andrea's VM.
I would expect that most of the time, the Hurd would be used just like UNIX. But sometimes it's nice to have the extra flexibility. Of course, whether or not it's worth it is up to the admin/computer owner.
Well, you definitely can't have that amount of control in Linux. Unless there's already an entry in /etc/fstab, normal users cannot mount a new filesystem.
I've never used OS X, so I can't say anything about that. But how easy would it be to add support for a new filesystem (without being root)? Assuming that there is a way to add new filesystems at all. How easy is it to mount a loopback filesystem?
BTW, there are other advantages to the Hurd than just the multiple server thing. For example, the authentication mechanism allows a daemon (take telnet or ssh as an example) to start of running as no user (not to be confused with a "nobody" user) and, when it gets appropriate authentication tokens, to upgrade its permissions. This is in contrast to traditional UNIX-like OS's, where telnet and ssh need to run as root, and drop you to your own user when you log in. Hurd's method effectively prevents remote root exploits (short of guessing root's (or someone else's) password). (Actually, that may also be related to the multi-server nature of Hurd. I don't know the Hurd well enough to say.)
The last official release of the Hurd is 0.2. It has made a lot of progress since then, but no one has had the urge (yet) to declare a new version number. Hurd development is mostly done in conjunction with Debian now, so the next Hurd release will probably be named the same as the next Debian release (though I may be wrong about that).
The Hurd is aparently quite usable already. (I don't use it, so I can't give any first hand information.) XFree86 and a lot of other software already runs on it. One notable exception is GNOME, because they don't have a working POSIX threads library.
If you want to try it out, look at http://www.debian.org/ports/hurd/. Working with the Debian base, they had over half (I forget the exact number) the packages compiled and usable about two months ago. Since then, they've been working on an ABI change, so they had to purge all their old packages, but they're working on recompiling everything again.
As for working with Apple's Mach kernel, they probably could, but it would require work to port the system over. They're actually moving away from the Mach microkernel, and towards the L4 microkernel, so they probably wouldn't want to port it to Apple's Mach kernel. It may not even be good enough for the Hurd either. The Hurd takes a multi-server approach, while OS X is still a monolithic design (AFAIK). When you do multi-server, the speed of interprocess communication becomes critical, whereas it is not so important under a monolithic design.
Look through the "Introductory Material" at http://www.gnu.org/software/hurd/docs.html if you really want to read up on it.
The short answer is that it allows you to replace parts of the operating system easily. For programmers (OS developers in particular), you can do your own implementation of part of the OS and test it out, without disrupting any of the other users. If your bit crashes, no problem, just replace it with what was there before. For users, it allows you to do things like mount filesystems without being root or having to bug root to frob /etc/fstab, which you want to do if you want to work with an ISO image. Of course, there are other things that normal users can do with the Hurd, but this is probably the biggest.
My point was that tweakt was misunderstanding what RMS was talking about. But even so, this doesn't mean that the Hurd is "good in theory, bad in practice." It's just than nobody has bothered to code up that stuff yet. The Hurd is able to support whatever hardware you want to throw at it. You just need to code up support for it, if the Hurd doesn't already support it. Just like Linux, and any other operating system that you want to use.
I'll admit that I'm not an OS guru, so most of what I say is second hand, but you should look at Thomas Bushnell's paper at http://www.gnu.org/software/hurd/hurd-paper.html, which should give you and idea of how the Hurd and UNIX are different. If you're too lazy to read through the whole thing, jump down to "Why this is so different" or thereabouts. The beginning of part 1 is actually pretty interesting too, and is worth reading if you want to see how the Hurd differs from UNIX.
But why would preventing a user from mounting a loopback filesystem be a policy decision? The file is already owned by the user, so he/she can already do whatever he/she wants with it. He/she can make his/her own versions of cp/ls/etc. à la mtools to read from and write to the filesystem. Preventing the user from mounting doesn't do anything from a security standpoint, and just makes it harder for the user to do useful work.
You obviously don't know much about the architecture of the Hurd. Go read up on it (http://hurd.gnu.org/) and come back.
/etc/fstab, which he/she probably won't want to do), even if the encrypted file is owned by you, in your own directory and you want to mount it within your own home directory. This is something that the architecture prevents you from doing, so no UNIX implementation will ever let you do that (without a heavy dose of magic). Allowing stuff like that is one of the architectural features of the Hurd.
The architecture of the Hurd (not to be confused with the implementation) gives users a lot more freedom than any UNIX-based system. For example, UNIX will not let you mount a loopback encrypted filesystem unless you are root (or without bugging root to frob
Porting, and supporting 56k modems are just implementation details, and have nothing to do with the architecture, which is what RMS is talking about.
Whoa. If you're going to correct someone, at least get your facts straight. Dvorak wasn't created until the 1920s/1930s. The Dvorak was created to increase typing speed, rather than to decrease it, by putting the most commonly used letters on the home row. I forget the exact number, but It's something over 80% of the letters typed in a normal English text would be on the home row.
I have no idea where the author of the article got that idea. It's the stupidest thing I've heard for a long while. If you look around the Fastap site, you'll find that you do actually hit the recessed number, and it manages to ignore hits that may be generated by accidentally hitting one of the surrounding letters.
Incidentally, it looks like they also have a credit-card sized QWERTY keyboard, with numbers and symbols in between the letters. That might be interesting to stick onto a PDA.
If they port it to the Cocoa layer, it may be compatible with GNUStep.
A Slashdot effect would not result in a missing index.* file, and if it does, even more reason to avoid IIS. A server can choke due to high traffic, but if it ever deletes your data, stay away from it. It might've been hacked, but if it was, whoever hacked it should have replaced the index.* file with something else, rather than just deleting it. Another possibility is that they hired some clueless admins, because they couldn't afford expensive experts.
I wonder if they have all their patches applied. It would be even funnier if they got hit by Nimda/Code Red.
Bah! Anyone who uses Lynx is a weenie. Real geeks use netcat.
Hold on. Let me get this straight. First you say that Linux is about choice, and then you bash Sharp for using Opera instead of Mozilla?
And what is wrong with selling GPL'ed software for money? Even the FSF does that.
I believe what you are talking about is already being done, through dual licensing. See, for example, the licensing for FFTW (www.fftw.org - an FFT library developed by some MIT guys). It's licensed under the GPL, and if you don't want the GPL (e.g. you want to include it into a closed-source program), you can sign a licensing deal with MIT.
ReiserFS also does the same thing, and Hans has mentioned before that RMS hasn't complained about it yet.
Perhaps figuring out the navigation to their site is another part of the challenge. ;-)
I agree. I love docbook. If you can write HTML, chances are you can do docbook. It's plain text, so you can edit it in whatever program you want to.
;-) ). I use Jade -- openjade seems to be slower, and I couldn't see any advantage of openjade over Jade.
It's a bit of a pain to set up, but once it's going, it's great (much like most of Linux, I find
My only complaint about docbook is that it (currently) doesn't do math too well -- I use Walsh's stylesheets, and they don't seem to grok MathML yet -- so I have to stick with LaTeX for some things still. Oh, and its table support doesn't seem to be complete when using the TeX backend, so things may come out in unexpected ways.
Of course, you need to define a distribution for that to make any sense at all. Plus, the priors are unknown. In order to get anything meaningful out of Bayesian analysis, you need to know the priors, and we obviously don't.
Next, you assume that vast technical superiority = many more inhabitants. Simply not true.
Hmmm. Did you read the FAQ from the site? Question 10.
The output from a pseudo-random number generator is usually considered "random enough for practical purposes." So if you define "practically random data" as "data that is random enough for practical purposes," you can compress it by storing the random seed and the string length. ;-)
I think I can beat their 100:1 compression ratio with this scheme.
Not only would you need to save it, you'd also have to chmod it and make it executable before you explicitly execute it.
Sure. Windows NT 5.0 will never be released.
Sounds like they changed the definition of "forever" while I wasn't paying attention.
And then it goes on to talk about N and P silicon, and so forth. I guess they changed the definition of "easy to understand" too.
It's because of NS4's buggy CSS rendering. If you turn JS off in NS4, it also turns off CSS handling (silly Netscape).
The Galeon page is nothing fancy. It's just using normal HTML -- no JavaScript required. If they did their HTML properly, the page was meant to degrade gracefully. That is, if you view it in a browser that correctly adheres to the standards that it claims to, it should be usable. Their HTML looks pretty simple, and I wouldn't be surprised if they did do it in vi or emacs.
The problem that you have is that NS4 doesn't do CSS correctly, but it likes to pretend that it does. It's a problem that many web designers face, and often the decision is: "Screw Netscape" because NS4's CSS handling can be quite unpredictable.
Why would people use CSS? Ironically, because it's supposed to allow them to create a nice-looking page that will be usable in older browsers. It's unfortunate that Netscape had to screw it up.
If you want something that's fast, use Skipstone. It has less features than Galeon, but it's much faster. Particularly on startup.
But you probably want to give Galeon a try anyway.
Two problems (that haven't been mentioned yet): firstly, you now have the problem that the program needs to know what the piece of clay looked like initially. Unless you can always start with a perfectly spherical ball of clay... Secondly, you would need to model the physics of the clay, such as sagging. For all that trouble, you might as well just go with force feedback.