If you're using Debian or Ubuntu, you'll probably want to check out wpa_action. Just check the manpage. It allows you to control the wireless interface with the normal/etc/network/interfaces file and a wpa_supplicant config file, and in a very convenient manner allows you to set individual IP settings for different network.
I'm not really sure what you're getting at. IPv6 isn't all that complex in my view, and I'm usually the first one to complain about exaggerated complexity. At its base level, I'd even say it's less complex than IPv4, since the pure IPv6 protocol contains fewer features (no IP-level options, for instance) and has all the extra features stacked on top of it in protocols layered above.
Sure, there are lots of things that can be implemented on top of IPv6 which entail a lot of complexity, like mobility support, but in most implementations, you can just leave that out. The implementations in the larger operating systems like Linux, the BSDs and Windows are very complex indeed, because they support most of the higher-level functionality, but I see no reason why simpler implementations would have to be bothered with that. I'm very curious as to what complex features or layering violations it is that you are referring to, because I cannot think of anything in particular in the base specification.
The greatest mistake in the IPv6 standard, in my view, is the fact that it specifies IPSec as required functionality, but I don't think that will be a problem. When makers of embedded devices start implementing IPv6, there's no chance they'll be implementing IPSec, and the requirements specification will probably be split into something like "base" and "full" support, the former not requiring IPSec.
I don't know what to do if it keeps reappearing, though.
Which it does, though. RAs are usually repeated every once in a while (and Vista is no exception), and each time it is repeated, the kernel adds a new route through it again.
That, and sometimes I've encountered 5-10 Vista laptops on the subnet, and it starts becoming a pain to remove all the routes.
The real problem is that GNU/Linux doesn't detect broken IPv6 connectivity and keeps using the AAAA records it gets from DNS.
Well, I'll have to disagree with you utterly there. The real problem is that Vista announces that it will route and then drops the packets routed through it.
I, too, am using 6to4 at home in order to get rid of NAT, but lately I've been having great trouble when traveling around with my IPv6-enabled laptop (running Debian).
See, whenever I get to a public access point (which uses public IPv4 addresses, rather than a private 192.168.x.x net) it turns out that any Vista computers connected to the same link auto-configure themselves to use 6to4 and then advertise over ICMP that they are willing to route traffic through their 6to4 net. However, it turns out that they just drop the traffic! My laptop, not knowing that, though, will try to route IPv6 traffic through them nevertheless, which just makes every IPv6 site (including my own) stop working. Viva Vista!
Does anyone know why Vista does this, and whether it's possible to prevent or work around it somehow?
For purely resistive loads, yes. I can't say I know the details of how CPUs work internally with regards to power consumption, but I would not be surprised if it is also proportional to the clock speed. In that case, it would be another 1/3 to that, making 192% normal TDP, giving 240 W.
There are no intuitive interfaces.. everything it learned.
My thought exactly. Many times, I've wondered if a command-line interface isn't actually superior very much for that exact reason -- it at least makes no attempt to appear intuitive, so at least it doesn't give people the wrong idea. They realize from the start that they'll have to learn.
Most likely to make it easier for people to set up routing. If you have a/64 subnet, it's easy to take 8 or 16 bits in some steps to partition the space into the subnets you want for different department and other things, all while keeping the possibility for automatic assignment of addresses.
Many Mac users, instead of using some 'generic' WiFi access point, instead use Apple's Airport Extreme router. Per the article, Airport Extreme's have support for IPv6 built right into the router, and the router will *automatically* route IPv6 traffic using the 6to4 standard (which basically tunnels the traffic over the IPv4 connection from the ISP).
Indeed. I was quite impressed to read about that. I have been thinking for quite a while that router makes should be doing exactly that, so it's good to see that at least one of them does.
On the quite opposite hand, there's Vista. While the article pointed out that Vista sets up 6to4 automatically when it has a globally routable IPv4 address (which is a good thing, of course), there's an annoying other side to that coin. See, Vista announces that it routes through its 6to4 address, but then in actual fact doesn't (it just drops the packets silently). It has been annoying me quite some times when I've connected to a public WiFi access point at my university, only to see every IPv6-enabled site (including my own!) fail miserably since my Linux laptop will try to route through one of these Vista black holes. That's Microsoft for you...
You mention undo only as such, but I'd like to add that the undo functionality really is one of Emacs' killer functions. See, when you undo a string of commands in Emacs, it doesn't just discard that part of the undo history as most (all?) other editors do, but it folds it backwards onto the undo stack, which is unbelievably useful.
I often use it as a sort of short-term, local version control. I can type something, try it, undo it and type something else, try that, and then undo back to the first version if I weren't satisfied with the second. It also enables me to use undo as a sort of short-term memory extension in that I can hack around a bit, and then interrupt myself for a visit to the toilet or similar. Then, when I get back, I can check exactly what I was doing before I left by just undoing a bit, and then undo the undoings.
I don't think I'd ever be able to use an editor that doesn't have Emacs' killer undo.
This is only true because people write such terrible and incomplete manual pages.
The original Bell Labs man pages completely described the system from the point of view of an administrator or user. The only better documentation was the source.
I agree completely, but I'd like to add here that some Linux distros are better than others. I've found that Debian is one of the better -- especially since Debian packagers tend to write manpages for programs that lack them while packaging.
Of course, the BSDs (probably because of them being real Unix) beats every Linux distro by far. That was one of the things that hit me the hardest when I started using FreeBSD. "Wow, the manpages are usable!"
The current blight of wimpy, inaccurate and incomplete man pages seems to originate from the GNU developers who insist on using the terrible "info" crap, writing huge volumes of text with no real content, and the tradition is continued by Linux developers who generally provide little or no man page documentation
I agree with you to some part here, but not completely. I certainly agree that the info system should not be a replacement for manpages, but it has a very definite function in that in can describe larger systems of programs (and larger programs, like Emacs) more completely than can manpages.
Indeed, there are many info manuals that are extremely useful, such as the one for glibc and the GNU autotools. I use them very often. The one for emacs is also actually very good, but I don't find myself using it very often at all, since emacs is so good at documenting itself anyway. There are some that aren't all that good, too, though. The one for gcc comes to mind.
Yes, and don't forget C-x z for repeating a command such as C-x e. The convenient thing about it is that you can keep pressing just the z key afterwards to keep repeating.
I find it especially useful when re-indenting code blocks that are a bit to irregular to use C-c C-q: Just do C-x ( TAB C-n C-x ) C-x e C-x z z z z...
(defun switch-to-previous-buffer (numbufs) "Switches to the previous file-associated buffer in the buffer list." (interactive "p") (switch-to-buffer (get-previous-buffer numbufs (buffer-list))))
In short, it allows you to switch back and forth between your two most recently used buffers with M-RET, or between elder buffers with a prefix argument. It's a very convenient alternative to C-x b (though it doesn't replace it, obviously).
I agree as far as writing a new X11 server would probably be a good idea, but as far as I can tell from the project home page, this is not a new X server, but some completely custom display server.
Window management is largely pushed to the clients, they draw their own decorations and move and resize themselves, typically implemented in a toolkit library. More of the core desktop could be pushed into wayland, for example, stock desktop components such as the panel or the desktop background.
Everything is direct rendered and composited. No cliprects, no drawing api/protocl between server and client. No pixmaps/windows/drawables, only surfaces (essentially pixmaps). No gcs/fonts, no nested windows. OpenGL is already direct rendered, pixman may be direct rendered which adds the cairo API, or cairo may gain a GL backend.
Could be a "shell" for launching gdm X server, user session servers, safe mode xservers, graphics text console. From gdm, we could also launch a rdp session, solid ice sessions.
The wayland protocol is a async object oriented protocol. All requests are method invocations on some object. The request include an object id that uniquely identifies an object on the server. Each object implements an interface and the requests include an opcode that identifies which method in the interface to invoke.
Why would you want to eliminate X from computer games? Wouldn't you normally want them to be managed by your normal window manager, so that you can switch to or from them or play them in windowed mode? I know I do.
What does it even mean that it "scales to 256 cores", though? I'd interpret it to say that it has support for 256-processor SMP scheduling, but in that case, my reaction would have to be "not until now?". Seriously, does Windows Datacenter Edition (oh, the oxymoron that it is) not support that many CPUs already? Or does it mean that the standard edition of Windows 7 will just be licensed to be able to use 256 CPUs?
Also, how many CPUs does Linux support these days? Last I read about it a couple of years ago, it supported 1024 processors.
And if/when some sinister DNS poisoning shows up to the party, it's gonna be big trouble.
Not with SSL client certificate authentication, it won't.
Or heck, someone simply gets hold of your master password. Oops someone just got full access to ALL your favorite websites...
Seeing how most people use the same stock password for most sites anyway, it seems the same problem exists anyway. If you're actually one of the few (like me, though;) who actually use different passwords, then the smart thing to do would obviously be to have different OpenID accounts corresponding to those different passwords and trust levels. Or, again, just use SSL authentication.
isn't this a major single point of failure issue or have I missed something?
Well, to begin with, you'll have to realize that OpenID is mostly intended to facilitate such simple things as posting comments to on-line forums like Slashdot, various blog engines or phpBB-like bulletin boards. If you just find a blog and wish to comment, it's very tedious to have to register an account just for that. OpenID fixes that problem. You are not meant to be logging into your bank account with OpenID.
Unless I've missed the point somehow and there's some way to know if the site you're on is accredited.
You have indeed missed the point, and even more than you think. You don't enter your OpenID password on the site you're authenticating to, at all. Ever. You just enter your OpenID username, and it redirects you to your actual OpenID provider, and there you enter your password (or, even better, use the SSL certificate installed in your browser, or your Kerberos credentials, or similar) to authenticate to it. It then redirects you back to the actual site with a cryptographic cookie that verifies your identity.
If you're worried about phishing, that's a very different issue. Certainly a real one, though, but not anything you wouldn't be subjected to anyway. And, if you authenticate with something like an SSL certificate, it won't be a problem anyway.
Well, I'm sorry for misconstruing your words, then. I do think that you could have been a bit clearer, though---after all, it's not like there isn't any source; you're just doubting its validity.
What a load of crap. The ONLY thing a kill switch is good for is to take control of choice of the apps away from the customer, and to set up and maintain a monopoly.
It's really this simple: You don't have to get your programs through Google. Hence, there's no monopoly. Hence, there's no danger.
By your logic, all ISP's and Microsoft should install a kill-switch on your computer to protect the Internet. And they should be responsible for what's on your computer, and not you.
Now I'm curious. Just how do would you use my logic to get to that conclusion? Please reply.
Is there a source for this statement? People in the comment threads have said this a dozen times, but nobody's mentioned why they believe this is true.
Strange as this may sound, if you look hard enough in the summary, you'll find that some words are underlined. The fact is, that if you click on these words, your web browser will take you elsewhere, and even stranger is that one of these "links" (as we call a consequent group of such words leading to the same destination) will lead you to a site other than Slashdot. We call that place the "article" in layman's terms ("TFA" in common Slashdot parlance).
Now, of course, I wouldn't expect you, or many others, to actually know these secrets, but some would consider them a source for points in the discussion of, well, an article.
Slashdot Mirror
If you're using Debian or Ubuntu, you'll probably want to check out wpa_action. Just check the manpage. It allows you to control the wireless interface with the normal /etc/network/interfaces file and a wpa_supplicant config file, and in a very convenient manner allows you to set individual IP settings for different network.
I'm not really sure what you're getting at. IPv6 isn't all that complex in my view, and I'm usually the first one to complain about exaggerated complexity. At its base level, I'd even say it's less complex than IPv4, since the pure IPv6 protocol contains fewer features (no IP-level options, for instance) and has all the extra features stacked on top of it in protocols layered above.
Sure, there are lots of things that can be implemented on top of IPv6 which entail a lot of complexity, like mobility support, but in most implementations, you can just leave that out. The implementations in the larger operating systems like Linux, the BSDs and Windows are very complex indeed, because they support most of the higher-level functionality, but I see no reason why simpler implementations would have to be bothered with that. I'm very curious as to what complex features or layering violations it is that you are referring to, because I cannot think of anything in particular in the base specification.
The greatest mistake in the IPv6 standard, in my view, is the fact that it specifies IPSec as required functionality, but I don't think that will be a problem. When makers of embedded devices start implementing IPv6, there's no chance they'll be implementing IPSec, and the requirements specification will probably be split into something like "base" and "full" support, the former not requiring IPSec.
I don't know what to do if it keeps reappearing, though.
Which it does, though. RAs are usually repeated every once in a while (and Vista is no exception), and each time it is repeated, the kernel adds a new route through it again.
That, and sometimes I've encountered 5-10 Vista laptops on the subnet, and it starts becoming a pain to remove all the routes.
The real problem is that GNU/Linux doesn't detect broken IPv6 connectivity and keeps using the AAAA records it gets from DNS.
Well, I'll have to disagree with you utterly there. The real problem is that Vista announces that it will route and then drops the packets routed through it.
What crazy configuration are you referring to? I'm just using a plain Debian installation with regards to network settings.
I, too, am using 6to4 at home in order to get rid of NAT, but lately I've been having great trouble when traveling around with my IPv6-enabled laptop (running Debian).
See, whenever I get to a public access point (which uses public IPv4 addresses, rather than a private 192.168.x.x net) it turns out that any Vista computers connected to the same link auto-configure themselves to use 6to4 and then advertise over ICMP that they are willing to route traffic through their 6to4 net. However, it turns out that they just drop the traffic! My laptop, not knowing that, though, will try to route IPv6 traffic through them nevertheless, which just makes every IPv6 site (including my own) stop working. Viva Vista!
Does anyone know why Vista does this, and whether it's possible to prevent or work around it somehow?
Power usage grows as the square of the voltage.
For purely resistive loads, yes. I can't say I know the details of how CPUs work internally with regards to power consumption, but I would not be surprised if it is also proportional to the clock speed. In that case, it would be another 1/3 to that, making 192% normal TDP, giving 240 W.
There are no intuitive interfaces.. everything it learned.
My thought exactly. Many times, I've wondered if a command-line interface isn't actually superior very much for that exact reason -- it at least makes no attempt to appear intuitive, so at least it doesn't give people the wrong idea. They realize from the start that they'll have to learn.
[...]MS's anticompetitive^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H incentive[...]
Are you aware that you can erase an entire word with ^W instead?
Most likely to make it easier for people to set up routing. If you have a /64 subnet, it's easy to take 8 or 16 bits in some steps to partition the space into the subnets you want for different department and other things, all while keeping the possibility for automatic assignment of addresses.
Many Mac users, instead of using some 'generic' WiFi access point, instead use Apple's Airport Extreme router. Per the article, Airport Extreme's have support for IPv6 built right into the router, and the router will *automatically* route IPv6 traffic using the 6to4 standard (which basically tunnels the traffic over the IPv4 connection from the ISP).
Indeed. I was quite impressed to read about that. I have been thinking for quite a while that router makes should be doing exactly that, so it's good to see that at least one of them does.
On the quite opposite hand, there's Vista. While the article pointed out that Vista sets up 6to4 automatically when it has a globally routable IPv4 address (which is a good thing, of course), there's an annoying other side to that coin. See, Vista announces that it routes through its 6to4 address, but then in actual fact doesn't (it just drops the packets silently). It has been annoying me quite some times when I've connected to a public WiFi access point at my university, only to see every IPv6-enabled site (including my own!) fail miserably since my Linux laptop will try to route through one of these Vista black holes. That's Microsoft for you...
I think so, but maybe you know more about it than I do. I don't see how it helps me switch buffers in a single keypress (or two, for older buffers).
(global-set-key "\C-z" 'undo)
You mention undo only as such, but I'd like to add that the undo functionality really is one of Emacs' killer functions. See, when you undo a string of commands in Emacs, it doesn't just discard that part of the undo history as most (all?) other editors do, but it folds it backwards onto the undo stack, which is unbelievably useful.
I often use it as a sort of short-term, local version control. I can type something, try it, undo it and type something else, try that, and then undo back to the first version if I weren't satisfied with the second. It also enables me to use undo as a sort of short-term memory extension in that I can hack around a bit, and then interrupt myself for a visit to the toilet or similar. Then, when I get back, I can check exactly what I was doing before I left by just undoing a bit, and then undo the undoings.
I don't think I'd ever be able to use an editor that doesn't have Emacs' killer undo.
This is only true because people write such terrible and incomplete manual pages.
The original Bell Labs man pages completely described the system from the point of view of an administrator or user. The only better documentation was the source.
I agree completely, but I'd like to add here that some Linux distros are better than others. I've found that Debian is one of the better -- especially since Debian packagers tend to write manpages for programs that lack them while packaging.
Of course, the BSDs (probably because of them being real Unix) beats every Linux distro by far. That was one of the things that hit me the hardest when I started using FreeBSD. "Wow, the manpages are usable!"
The current blight of wimpy, inaccurate and incomplete man pages seems to originate from the GNU developers who insist on using the terrible "info" crap, writing huge volumes of text with no real content, and the tradition is continued by Linux developers who generally provide little or no man page documentation
I agree with you to some part here, but not completely. I certainly agree that the info system should not be a replacement for manpages, but it has a very definite function in that in can describe larger systems of programs (and larger programs, like Emacs) more completely than can manpages.
Indeed, there are many info manuals that are extremely useful, such as the one for glibc and the GNU autotools. I use them very often. The one for emacs is also actually very good, but I don't find myself using it very often at all, since emacs is so good at documenting itself anyway. There are some that aren't all that good, too, though. The one for gcc comes to mind.
Yes, and don't forget C-x z for repeating a command such as C-x e. The convenient thing about it is that you can keep pressing just the z key afterwards to keep repeating.
I find it especially useful when re-indenting code blocks that are a bit to irregular to use C-c C-q: Just do C-x ( TAB C-n C-x ) C-x e C-x z z z z...
Of all the elisp I've written, this is the one I miss the most when I don't use my own EMACS:
(defun get-previous-buffer (numbufs blist)
(if (not blist) (signal 'no-such-buffer ()))
(if (not (buffer-file-name (car blist)))
(get-previous-buffer numbufs (cdr blist))
(if (> numbufs 0)
(get-previous-buffer (1- numbufs) (cdr blist))
(car blist))))
(defun switch-to-previous-buffer (numbufs)
"Switches to the previous file-associated buffer in the buffer
list."
(interactive "p")
(switch-to-buffer (get-previous-buffer numbufs (buffer-list))))
(global-set-key "\M-\C-m" 'switch-to-previous-buffer)
In short, it allows you to switch back and forth between your two most recently used buffers with M-RET, or between elder buffers with a prefix argument. It's a very convenient alternative to C-x b (though it doesn't replace it, obviously).
I agree as far as writing a new X11 server would probably be a good idea, but as far as I can tell from the project home page, this is not a new X server, but some completely custom display server.
Some excerpts from the documentation:
Window management is largely pushed to the clients, they draw their own decorations and move and resize themselves, typically implemented in a toolkit library. More of the core desktop could be pushed into wayland, for example, stock desktop components such as the panel or the desktop background.
Everything is direct rendered and composited. No cliprects, no drawing api/protocl between server and client. No pixmaps/windows/drawables, only surfaces (essentially pixmaps). No gcs/fonts, no nested windows. OpenGL is already direct rendered, pixman may be direct rendered which adds the cairo API, or cairo may gain a GL backend.
Could be a "shell" for launching gdm X server, user session servers, safe mode xservers, graphics text console. From gdm, we could also launch a rdp session, solid ice sessions.
The wayland protocol is a async object oriented protocol. All requests are method invocations on some object. The request include an object id that uniquely identifies an object on the server. Each object implements an interface and the requests include an opcode that identifies which method in the interface to invoke.
Why would you want to eliminate X from computer games? Wouldn't you normally want them to be managed by your normal window manager, so that you can switch to or from them or play them in windowed mode? I know I do.
What does it even mean that it "scales to 256 cores", though? I'd interpret it to say that it has support for 256-processor SMP scheduling, but in that case, my reaction would have to be "not until now?". Seriously, does Windows Datacenter Edition (oh, the oxymoron that it is) not support that many CPUs already? Or does it mean that the standard edition of Windows 7 will just be licensed to be able to use 256 CPUs?
Also, how many CPUs does Linux support these days? Last I read about it a couple of years ago, it supported 1024 processors.
And if/when some sinister DNS poisoning shows up to the party, it's gonna be big trouble.
Not with SSL client certificate authentication, it won't.
Or heck, someone simply gets hold of your master password. Oops someone just got full access to ALL your favorite websites...
Seeing how most people use the same stock password for most sites anyway, it seems the same problem exists anyway. If you're actually one of the few (like me, though ;) who actually use different passwords, then the smart thing to do would obviously be to have different OpenID accounts corresponding to those different passwords and trust levels. Or, again, just use SSL authentication.
isn't this a major single point of failure issue or have I missed something?
Well, to begin with, you'll have to realize that OpenID is mostly intended to facilitate such simple things as posting comments to on-line forums like Slashdot, various blog engines or phpBB-like bulletin boards. If you just find a blog and wish to comment, it's very tedious to have to register an account just for that. OpenID fixes that problem. You are not meant to be logging into your bank account with OpenID.
Unless I've missed the point somehow and there's some way to know if the site you're on is accredited.
You have indeed missed the point, and even more than you think. You don't enter your OpenID password on the site you're authenticating to, at all. Ever. You just enter your OpenID username, and it redirects you to your actual OpenID provider, and there you enter your password (or, even better, use the SSL certificate installed in your browser, or your Kerberos credentials, or similar) to authenticate to it. It then redirects you back to the actual site with a cryptographic cookie that verifies your identity.
If you're worried about phishing, that's a very different issue. Certainly a real one, though, but not anything you wouldn't be subjected to anyway. And, if you authenticate with something like an SSL certificate, it won't be a problem anyway.
Well, I'm sorry for misconstruing your words, then. I do think that you could have been a bit clearer, though---after all, it's not like there isn't any source; you're just doubting its validity.
What a load of crap. The ONLY thing a kill switch is good for is to take control of choice of the apps away from the customer, and to set up and maintain a monopoly.
It's really this simple: You don't have to get your programs through Google. Hence, there's no monopoly. Hence, there's no danger.
By your logic, all ISP's and Microsoft should install a kill-switch on your computer to protect the Internet. And they should be responsible for what's on your computer, and not you.
Now I'm curious. Just how do would you use my logic to get to that conclusion? Please reply.
Is there a source for this statement? People in the comment threads have said this a dozen times, but nobody's mentioned why they believe this is true.
Strange as this may sound, if you look hard enough in the summary, you'll find that some words are underlined. The fact is, that if you click on these words, your web browser will take you elsewhere, and even stranger is that one of these "links" (as we call a consequent group of such words leading to the same destination) will lead you to a site other than Slashdot. We call that place the "article" in layman's terms ("TFA" in common Slashdot parlance).
Now, of course, I wouldn't expect you, or many others, to actually know these secrets, but some would consider them a source for points in the discussion of, well, an article.
Well, maybe, maybe not. In the end, it doesn't really matter. What matters is that it's possible to get the program you want, and that you can.