To use them in linux, you want to look at the man pages for showkey, loadkeys (for console), and probably xev, xmodmap, and xkeycaps (for X). (I wonder if anyone has made xkeycaps know about these keyboards yet? That would be really cool. Get xkeycaps. It makes messing with your keyboard really easy.) #define X(x,y) x##y
One of the most important parts of the GNU GPL is that people (and especially big companies) can't use the fruits of your labour without making their source available. The idea is that this makes it harder for closed-source stuff to be as good as the open-source stuff, so people are more likely to use the open-source stuff. Better software for Windoze users is ok, as long as it is open source. Making a closed-source project better is very much against the rules, and very much pisses off many people. #define X(x,y) x##y
I hope we see more stories from the buy-stuff-and-give-it-away dept. in the future:) I love that move. What stuff is out there that the open-source community would like to have bought and freed? Napster? (then we could take out the part that ignores TCP congestion control, and make some other improvements, and release a new version on all platforms.) #define X(x,y) x##y
You're right that the redhat installer tends to load up your system with crap. That only happens if you just let it take the defaults for packages to install, which is what a no-clue newbie would be best doing. If you know anything at all, you'll un-select all the stuff you don't need. Basically, BSD, and other Linux distros like Slackware and Debian, use an opt-in installer, where you have to ask for anything that you want, as opposed to redhat where you have to opt-out of stuff you don't want. You are refering to the redhat way as the "linux way". Don't do that. Different distros have _very_ different flavours and philosophies. Give some others a try sometime. Stampede gives you a really raw system where you'll need to fix a lot of config files and stuff, but it's a learning experience to half-way roll your own system. AFAIK, slackware is most similar to the BSD package system (but BSD has some good stuff that slack doesn't, I think.) Debian is very concerned with having everything fit together well, and being well documented. Every package has someone specifically maintaining it, so things tend to be well set up for each other. It is also very concious of security, including out of the box security so you don't get cracked before you figure out how to pronounce "linux":) (BTW, I'm not calling you a newbie, but maybe you were when you installed redhat and had it litter your system with junk. Everyone is on their first install of any OS/distro.) #define X(x,y) x##y
Mandrake and RedHat aren't known for feeling efficient. They feel more like windoze unless you cut down on the massive GUI crap they like to set up. The side-by-side comparison won't be fair unless you're running the same window manager, with the same mouse (this makes a lot of difference to feel, IMHO), and with a comparable video card.
If you still find you like NetBSD more, then great, use it. As long as you use an OS which does what you want and you know how to use it, you've got my respect. (i.e. I'm not one of those annoying linux zealots.) #define X(x,y) x##y
RedHat is not the only linux distro out there. Unlike BSD, all linux distros use the same kernel (well, different versions of the same kernel), but they do have different package formats and apps to manage them. Give Debian a try. It has a _very_ powerful package system:) #define X(x,y) x##y
I absolutely agree with you. email and news reading also has to go. That's what mutt and slrn are for.
I've been thinking for a while about writing a web browswer that uses the gecko engine, but doesn't have all of Moz's crap plastered all over it. Now that I see some like-minded posts, I am most encouraged:) I haven't written anything big before, but I'd love to try. If anyone wants to help me write one, email me: peter@llama.nslug.ns.ca and we'll figure out what language to use, how portable to make it, and all that jazz. #define X(x,y) x##y
NS 4.72 isn't exactly quick on its feet either. Every Unix graphical browser that can handle modern HTML is huge and bloated, AFAIK.
I think the problem with Mozilla is not that Gecko is slow, but rather that the app build around it is SLOW. I'm thinking about taking the Gecko stuff and writing a small, light-weight browser (and _not_ a news+email client! damnit...) around that. I'm heartened to know that other companies are using gecko, since that means it is possible to use it without the rest of Mozilla, and that it doesn't suck. If it sucked, all those companies wouldn't be using it. Especially not the cell-phone companies. #define X(x,y) x##y
Certainly some motivation to enter into education or an -ology where all the girls are at, (and the good lookin ones too, definately not enough of those in computer science.)
So that's why there are almost no girls in physics, but there's lots (and good looking at my university, too.:), in biology: Physics doesn't end with ology. Damn:(. OTOH, there are girls in chemistry. At the Canadian Physics Olympiad, 1997, there were 15 guys and no girls. The Chemistry Olympiad, at the same cegep (sp?) in Montreal, had 8 guys and 7 girls, IIRC. #define X(x,y) x##y
Why use Intel PIII CPUs as the reference? You'd need to figure out how to scale other archs, so you'd need a benchmark. Probably the most widely accepted benchmarks are the SPECint and SPECfp benchmarks. I think it would make the most sense to use straight SPECint numbers for computer speed.
I'm wondering how to include SPECfp numbers, though. Maybe we should choose a reference speed, as 47Ronin did, and assign it speed 1. Speed of other CPUs would be calculated by averaging the ratios SPECint(cpu)/SPECint(ref) and SPECfp(cpu)/SPECfp(ref). So the final formula would be: speed(cpu) = speed(ref) * ( Sint(cpu)/Sint(ref) + Sfp(cpu)/Sfp(ref) )/2
This way, a CPU has a SPECfp rating twice as high as the reference, but only 3/4 the SPECint, would have speed=1.375, which seems reasonable. Obviously, you will still need to look at real benchmarks, not just a speed number, to see if a given CPU will do what you want, but at least it would give _some_ meaning to the numbers shown in ads:)
I think it would be necessary to pick a reference with a good balance between integer and fp, so things wouldn't be skewed. I think UltraSPARCs would be good for this, since x86 CPUs have too crappy FP, and Alphas have too good FP (relative to integer) I think. I'm sure there is a better choice though; Anyone want to tell me what it is?
That actually raises an interesting question: Will integer performance rise faster, slower, or equally to floating point performance in future CPUs? Which is easier to accomplish from a hardware design standpoint? #define X(x,y) x##y
What's so big about having it GPLed? If they released it under a less restrictive license, it would be better for everyone except them. (The difference between e.g. the X license and the GPL is that you can do more stuff with X licensed code than with GPLed code. The purpose of the GPL is to defend the author's wish that the code not be used in closed projects.) I GPL stuff I write, (which isn't much to speak of yet:( ). It would be cool if they released it under the GPL, since that would really make other companies who still like Micros~1 think about going GPL to be able to use the code. Releasing X licensed code is just as big a plunge for Micros~1, though, it's the real deal, Open Source the way its meant to be. (among other things, both the GPL and X style licenses meet the DFSG-free (Debian Free Software Guidelines) requirements.) #define X(x,y) x##y
If they de-orbit the satellites, they'll do it over empty ocean. That is not cool from the point of view of seeing them burn up, but I guess the advantage of most likely not killing anybody kind of takes precedence here.
What should happen is that some TV network should get a plane to fly out there at high altitude where they can get some great footage of a burning-up satellite. I'd love to see that! I might even turn on my TV, which I haven't done for a while. (and not because I left it on:)
I think it would work. A plane doesn't cost that much to fly around in. Iridium could even recoup some money on it if they sell the TV rights or something. (well probably a couple million, which is chump change compared to what they're in the hole by, but it's a lot to me:) #define X(x,y) x##y
I agree with you, except that you're wrong about google's use of cookies. Observe: llama:~$ nc google.com 80 HEAD / HTTP/1.0
HTTP/1.0 200 OK Date: Fri, 17 Mar 2000 01:30:56 GMT Server: GWS/1.6 Connection: close Set-Cookie: ID=6011ba7756ca44bd; domain=.google.com; path=/; expires=Sun, 17-Jan-2038 19:14:07 GMT Content-Length: 1539 Content-Type: text/html Last-Modified: Wed, 15 Mar 2000 04:58:03 GMT
See the Set-Cookie header?
Even so, I mostly agree with your point that there doesn't seem to be any advantage to topclick over google, since neither one uses banner adds to help some nasty company track you. OTOH, topclick is making a point about their privacy intentions, and I like that. I'll support them for that any day, as long as they're in the same league. (and they are, since they use google:). #define X(x,y) x##y
... only OS I know of that you can download a driver, download an 1 file, put it in some directory, add an interface w/ that driver, configure TCP/IP for the device, and restart the networking subsystem-- all in under 2 minutes, without restarting the computer once
Well then, meet GNU/Linux, another OS which can do what you describe for ethernet cards. You can reconfigure the kernel, make modules;make modules_install, then load the module into running kernel, and use ifconfig or dhcpcd to set it up. More than 2 minutes, but only if you compile it from source. If you had the driver module (1 file) already, then you could just load it. I have no idea what other systems can do stuff like this, but I'm sure some can. #define X(x,y) x##y
Consider this: anything that a human wrote is likely to be at least somewhat comprehensible. Reading human-generated LaTeX would be far preferable to MS Word in a text editor. Since it was written by a human, they would have made at least some effort to keep things legible for their own benefit. LaTeX can be pretty hairy sometimes, but the information is usually easy enough to find/modify, even if the formatting isn't. In any case, you can look at the electonically rendered (by xdvi for LaTeX, and by a browser for HTML), and then find the words that interest you in the source with a text editor.
(my bias: I use LaTeX for all my stuff that I write to be printed out.) #define X(x,y) x##y
You'll never know when Debian's going to decide to nuke your master boot record and keep you from booting.
I thought I knew the answer to that question, and I thought it was never. Does this happen to you or something? (I don't use any non-standard kernel that would prevent recovery by booting of a random linux CD, but that would be a serious problem (Critical level bug at least (think about a server in a building far away...)) if Debian screwed up a system like that. I've never had problems when I haven't run lilo manually.) #define X(x,y) x##y
Another reason it was not real: There were no typo's or spelling errors, yet it was signed with the mark of the 'Taco. Hmmmm, something strange about that, I must say.:) #define X(x,y) x##y
The Linux kernel uses a lot of complicated data structures. (maybe not complicated DS, but used in complicated ways, with many functions which access them). This makes the kernel not very good for teaching about e.g. virtual memory, because the VM code probably does some filesystem and/or scheduling stuff too. To completely understand (like that ever happens:) a piece of code in the kernel, you have to know about a lot of extraneous stuff. (That's been my experience looking at kernel code. I haven't looked at enough to be confident hacking it, though.)
OTOH, I found it extremely helpful for my third year operating systems course to have read lots of docs about things in the kernel work, and to have written programs which make system calls like open(2), read(2), and mmap(2). (especially mmap. It's cool:). Linux is great as background knowledge when thinking about operating system concepts, but it is too complicated to make case studies from, or try to actually teach directly. Our prof tried to give some examples of real code near the end of the term, but he had to stop and explain a bunch of things, so we didn't get into much code. He did leave top(1) running in the ssh window on the PC connected to the overhead projector, so the class got to watch the system being seriously overloaded (load av. 40--70 on a 4way SMP UltraSPARC, 1GB RAM:). Some guy with a laptop ran a process which showed up as "hey_cs3120" or something like that for a few seconds:) #define X(x,y) x##y
That was obviously some other limp-brained/fool, since the first guy couldn't moderate after posting unless he posted really AC, not just _clicking_ the AC button. (i.e. from a browser without his cookie.) He could have been bullshitting about forgetting that posting reversed his moderating actions, but if so, there's no reasoning with him. (I'm assuming male here, (AC, not doomy). If you're insulted, then good..) BTW, I though that your first post was a good call, doomy. #define X(x,y) x##y
Slashdot Mirror
To use them in linux, you want to look at the man pages for showkey, loadkeys (for console), and probably xev, xmodmap, and xkeycaps (for X). (I wonder if anyone has made xkeycaps know about these keyboards yet? That would be really cool. Get xkeycaps. It makes messing with your keyboard really easy.)
#define X(x,y) x##y
Damn, I should have stopped after the first sentence. A DM7400 is a quad NAND TTL chip. A G4 is a Motorola 7400.
#define X(x,y) x##y
pgcc can generate MMX code now. Check out The PGCC FAQ
#define X(x,y) x##y
7400? Isn't that a quad TTL NAND gate? :) Is a G4 actually a PPC 740? (G3 = PPC 750, right?)
#define X(x,y) x##y
One of the most important parts of the GNU GPL is that people (and especially big companies) can't use the fruits of your labour without making their source available. The idea is that this makes it harder for closed-source stuff to be as good as the open-source stuff, so people are more likely to use the open-source stuff. Better software for Windoze users is ok, as long as it is open source. Making a closed-source project better is very much against the rules, and very much pisses off many people.
#define X(x,y) x##y
I hope we see more stories from the buy-stuff-and-give-it-away dept. in the future :) I love that move. What stuff is out there that the open-source community would like to have bought and freed? Napster? (then we could take out the part that ignores TCP congestion control, and make some other improvements, and release a new version on all platforms.)
#define X(x,y) x##y
s/biff/mesg/
:)
Other than that, thumbs up. Never used Multics, so I can't say whether
that part's right, but it sounds plausible
#define X(x,y) x##y
You're right that the redhat installer tends to load up your system with crap. That only happens if you just let it take the defaults for packages to install, which is what a no-clue newbie would be best doing. If you know anything at all, you'll un-select all the stuff you don't need. Basically, BSD, and other Linux distros like Slackware and Debian, use an opt-in installer, where you have to ask for anything that you want, as opposed to redhat where you have to opt-out of stuff you don't want. :)
You are refering to the redhat way as the "linux way". Don't do that. Different distros have _very_ different flavours and philosophies. Give some others a try sometime. Stampede gives you a really raw system where you'll need to fix a lot of config files and stuff, but it's a learning experience to half-way roll your own system. AFAIK, slackware is most similar to the BSD package system (but BSD has some good stuff that slack doesn't, I think.) Debian is very concerned with having everything fit together well, and being well documented. Every package has someone specifically maintaining it, so things tend to be well set up for each other. It is also very concious of security, including out of the box security so you don't get cracked before you figure out how to pronounce "linux"
(BTW, I'm not calling you a newbie, but maybe you were when you installed redhat and had it litter your system with junk. Everyone is on their first install of any OS/distro.)
#define X(x,y) x##y
> It just felt more efficient...
Mandrake and RedHat aren't known for feeling efficient. They feel more like windoze unless you cut down on the massive GUI crap they like to set up. The side-by-side comparison won't be fair unless you're running the same window manager, with the same mouse (this makes a lot of difference to feel, IMHO), and with a comparable video card.
If you still find you like NetBSD more, then great, use it. As long as you use an OS which does what you want and you know how to use it, you've got my respect. (i.e. I'm not one of those annoying linux zealots.)
#define X(x,y) x##y
RedHat is not the only linux distro out there. Unlike BSD, all linux distros use the same kernel (well, different versions of the same kernel), but they do have different package formats and apps to manage them. Give Debian a try. It has a _very_ powerful package system :)
#define X(x,y) x##y
I absolutely agree with you. email and news reading also has to go. That's what mutt and slrn are for.
:) I haven't written anything big before, but I'd love to try. If anyone wants to help me write one, email me: peter@llama.nslug.ns.ca and we'll figure out what language to use, how portable to make it, and all that jazz.
I've been thinking for a while about writing a web browswer that uses the gecko engine, but doesn't have all of Moz's crap plastered all over it. Now that I see some like-minded posts, I am most encouraged
#define X(x,y) x##y
NS 4.72 isn't exactly quick on its feet either. Every Unix graphical browser that can handle modern HTML is huge and bloated, AFAIK.
I think the problem with Mozilla is not that Gecko is slow, but rather that the app build around it is SLOW. I'm thinking about taking the Gecko stuff and writing a small, light-weight browser (and _not_ a news+email client! damnit...) around that. I'm heartened to know that other companies are using gecko, since that means it is possible to use it without the rest of Mozilla, and that it doesn't suck. If it sucked, all those companies wouldn't be using it. Especially not the cell-phone companies.
#define X(x,y) x##y
#define X(x,y) x##y
Why use Intel PIII CPUs as the reference? You'd need to figure out how to scale other archs, so you'd need a benchmark. Probably the most widely accepted benchmarks are the SPECint and SPECfp benchmarks. I think it would make the most sense to use straight SPECint numbers for computer speed.
:)
I'm wondering how to include SPECfp numbers, though. Maybe we should choose a reference speed, as 47Ronin did, and assign it speed 1. Speed of other CPUs would be calculated by averaging the ratios SPECint(cpu)/SPECint(ref) and SPECfp(cpu)/SPECfp(ref). So the final formula would be:
speed(cpu) = speed(ref) * ( Sint(cpu)/Sint(ref) + Sfp(cpu)/Sfp(ref) )/2
This way, a CPU has a SPECfp rating twice as high as the reference, but only 3/4 the SPECint, would have speed=1.375, which seems reasonable. Obviously, you will still need to look at real benchmarks, not just a speed number, to see if a given CPU will do what you want, but at least it would give _some_ meaning to the numbers shown in ads
I think it would be necessary to pick a reference with a good balance between integer and fp, so things wouldn't be skewed. I think UltraSPARCs would be good for this, since x86 CPUs have too crappy FP, and Alphas have too good FP (relative to integer) I think. I'm sure there is a better choice though; Anyone want to tell me what it is?
That actually raises an interesting question: Will integer performance rise faster, slower, or equally to floating point performance in future CPUs? Which is easier to accomplish from a hardware design standpoint?
#define X(x,y) x##y
#define X(x,y) x##y
What's so big about having it GPLed? If they released it under a less restrictive license, it would be better for everyone except them. (The difference between e.g. the X license and the GPL is that you can do more stuff with X licensed code than with GPLed code. The purpose of the GPL is to defend the author's wish that the code not be used in closed projects.) I GPL stuff I write, (which isn't much to speak of yet :( ). It would be cool if they released it under the GPL, since that would really make other companies who still like Micros~1 think about going GPL to be able to use the code. Releasing X licensed code is just as big a plunge for Micros~1, though, it's the real deal, Open Source the way its meant to be. (among other things, both the GPL and X style licenses meet the DFSG-free (Debian Free Software Guidelines) requirements.)
#define X(x,y) x##y
What should happen is that some TV network should get a plane to fly out there at high altitude where they can get some great footage of a burning-up satellite. I'd love to see that! I might even turn on my TV, which I haven't done for a while. (and not because I left it on :)
I think it would work. A plane doesn't cost that much to fly around in. Iridium could even recoup some money on it if they sell the TV rights or something. (well probably a couple million, which is chump change compared to what they're in the hole by, but it's a lot to me :)
#define X(x,y) x##y
I agree with you, except that you're wrong about google's use of cookies. Observe:
llama:~$ nc google.com 80
HEAD / HTTP/1.0
HTTP/1.0 200 OK
Date: Fri, 17 Mar 2000 01:30:56 GMT
Server: GWS/1.6
Connection: close
Set-Cookie: ID=6011ba7756ca44bd; domain=.google.com; path=/; expires=Sun, 17-Jan-2038 19:14:07 GMT
Content-Length: 1539
Content-Type: text/html
Last-Modified: Wed, 15 Mar 2000 04:58:03 GMT
See the Set-Cookie header?
Even so, I mostly agree with your point that there doesn't seem to be any advantage to topclick over google, since neither one uses banner adds to help some nasty company track you. OTOH, topclick is making a point about their privacy intentions, and I like that. I'll support them for that any day, as long as they're in the same league. (and they are, since they use google:).
#define X(x,y) x##y
Well then, meet GNU/Linux, another OS which can do what you describe for ethernet cards. You can reconfigure the kernel, make modules;make modules_install, then load the module into running kernel, and use ifconfig or dhcpcd to set it up. More than 2 minutes, but only if you compile it from source. If you had the driver module (1 file) already, then you could just load it. I have no idea what other systems can do stuff like this, but I'm sure some can.
#define X(x,y) x##y
Consider this: anything that a human wrote is likely to be at least somewhat comprehensible. Reading human-generated LaTeX would be far preferable to MS Word in a text editor. Since it was written by a human, they would have made at least some effort to keep things legible for their own benefit. LaTeX can be pretty hairy sometimes, but the information is usually easy enough to find/modify, even if the formatting isn't. In any case, you can look at the electonically rendered (by xdvi for LaTeX, and by a browser for HTML), and then find the words that interest you in the source with a text editor.
(my bias: I use LaTeX for all my stuff that I write to be printed out.)
#define X(x,y) x##y
I thought I knew the answer to that question, and I thought it was never. Does this happen to you or something? (I don't use any non-standard kernel that would prevent recovery by booting of a random linux CD, but that would be a serious problem (Critical level bug at least (think about a server in a building far away...)) if Debian screwed up a system like that. I've never had problems when I haven't run lilo manually.)
#define X(x,y) x##y
Another reason it was not real: There were no typo's or spelling errors, yet it was signed with the mark of the 'Taco. Hmmmm, something strange about that, I must say. :)
#define X(x,y) x##y
Of course they didn't find the tanks. I got 'em first, and sold them to Fidel Castro. Suck on that, Bill Clinton. hehehehe :)
#define X(x,y) x##y
The Linux kernel uses a lot of complicated data structures. (maybe not complicated DS, but used in complicated ways, with many functions which access them). This makes the kernel not very good for teaching about e.g. virtual memory, because the VM code probably does some filesystem and/or scheduling stuff too. To completely understand (like that ever happens :) a piece of code in the kernel, you have to know about a lot of extraneous stuff. (That's been my experience looking at kernel code. I haven't looked at enough to be confident hacking it, though.)
:). Linux is great as background knowledge when thinking about operating system concepts, but it is too complicated to make case studies from, or try to actually teach directly. Our prof tried to give some examples of real code near the end of the term, but he had to stop and explain a bunch of things, so we didn't get into much code. He did leave top(1) running in the ssh window on the PC connected to the overhead projector, so the class got to watch the system being seriously overloaded (load av. 40--70 on a 4way SMP UltraSPARC, 1GB RAM :). Some guy with a laptop ran a process which showed up as "hey_cs3120" or something like that for a few seconds :)
OTOH, I found it extremely helpful for my third year operating systems course to have read lots of docs about things in the kernel work, and to have written programs which make system calls like open(2), read(2), and mmap(2). (especially mmap. It's cool
#define X(x,y) x##y
That was obviously some other limp-brained /fool, since the first guy couldn't moderate after posting unless he posted really AC, not just _clicking_ the AC button. (i.e. from a browser without his cookie.) He could have been bullshitting about forgetting that posting reversed his moderating actions, but if so, there's no reasoning with him. (I'm assuming male here, (AC, not doomy). If you're insulted, then good..)
BTW, I though that your first post was a good call, doomy.
#define X(x,y) x##y