I started with slackware, like a lot of people. For a while, packages seemed the wimp way of doing things - it wasn't actually accomplishing anything, for God's sake! Far too easy.
Yes, I ran slackware. I was a real hacker. I singlehandedly installed gnome from source 4 or 5 times, in its entirety, because of how often I nuked my partition and started over.
Then, I realised that I was wasting far too much time compiling things. My Pentium 166 with 32 MB of RAM, which used to be state-of-the-art, now was taking an awfully long time compiling things.
So, for a while, I ran Red Hat. I liked it, too. Its printer management meant I didn't have to futz around with magicfiltres and linuxconf meant that, generally speaking, I didn't have to worry about much.
Then, I found Debian, and my saviour - apt-get.
Simply put, apt-get makes things absurdly easy. I don't have to worry about upgrading to glibc2.1 - apt does the worrying, the downloading, and the installing for me. With Debian I've finally set up the production box I knew Linux could be.
While Slackware will always hold a special place in my heart, for "where I started," Debian is where I am, and where I'm going in the future. Even though something still tells me it's a bit too easy, I just ignore that part of me. It's just too easy to tell my box to upgrade everything on my system to ignore it.
Oh, sure, that's a stability test - but it's not a test of Linux' stability. That's a test of the stability of Realplayer and Netscape for Linux, which is a very different thing. The fact is that it takes a lot to crash Linux, at least in the stable kernel series. Even if X crashes and hoses your display, Linux is still probably working. Of course, this doesn't mean much for the end-user; if the average person's display is hosed, they'll probably just hit reset - but the fact is that Linux itself isn't crashy like Netscape and Realplayer for Linux are.
That being said, I'm not planning on installing Realplayer anyhow. The only non-free software I have on my computer is Netscape, and I've got a clear upgrade path for that, too. When something coredumps, I want to be able to fix it - and unfortunately, Realplayer and Netscape do that, sometimes more often than others - and I can't do that with proprietary software.
I didn't agree to anything. I never once said "OK" on a form or signed an agreement when I registered woot.net. Therefore, I will hereby reproduce, transfer, AND modify the data presented to me by my whois request.
You agree that you will not reproduce, sell, transfer, or modify any of the data presented in response to your search request, or use of any such data for commercial purpose, without the prior express written permission of Network Solutions.
I haven't seen GNU come out with a 64-bit compiler yet
which, of course, is because you haven't looked at all - gcc, like Linux, is 64-bit clean; otherwise, the Alpha port wouldn't work (UltraSparc too, but I think they're still using 32-bit there).
Re:The article brought up some interesting points
on
NOS Crossroads
·
· Score: 1
It pointed out that:
Linux + Apache maxes out when it gets 2000 HTTP requests a second
I would put a great deal of money that this number was, in actuality, closer to 1024, which is the total number of processes Linux 2.2 can handle by default. There are patches floating about that increase this number (significantly? I'm not sure exactly to what - though I believe it's included in the -ac series of patches).
Why did they report it as 2000? They were probably running a LOT of clients which were connecting and dying off very rapidly, and as someone mentioned above, fork() performance in Linux is stellar. It was probably a case of resolution of benchmarks not being high enough.
As for the 200 Mbps - I would guess that this has to do with the network adaptors. The capacity to have multiple adaptors for the same interface is available, and was developed as part of the Beowulf project. It's probably going to be integrated into 2.3.
In any case, these problems are probably all already fixed, but not tested enough (or wide-spread enough) to be included into a stable distribution's (or kernel's) release.
I agree, but does it make it okay to publish or otherwise make available to foreign nations the blueprints to build a laser to blind spy satellites or distrupt communications satellites? This after all is entitled to similar free speech.
In short, yes. There's no difference between someone sharing, for example, the cure for cancer and someone sharing the blueprints or other schematics for a new type of laser, which has as one of its applications cutting things in orbit. Scream "national security" all you want, but chances are, and history will back me up on this, that the other guys already know about it because their scientists just found it. (e.g., Bell and his telephone). One mustn't ignore the other fact that the cure for cancer is also going to help your enemies - their ranks will be bolstered because of fewer people dying of cancer.
There is, after all, a difference between sharing ideas about things which are potentially damaging to you with your enemies, and outfitting your enemies with guns. The thought that really underlies proprietary software and crypto regulations in the USA is "we're the only ones who can do this," which is blatantly false, and nieve as well. Sharing thoughts, ideas, and sometimes even implementations is what has changed us from a society of relatively weak technology to a society of high technology in little over 100 years.
I suppose, of course, that you've tried to manage it and failed, and this is why you're saying that?
What you have to realise, of course, is that moving to microkernel would not eliminate code, simply modularise it - and Linux is already mostly modularised. Also, Linus and Alan Cox have a very good idea of what's going on in all areas of the kernel - ie, it hasn't gotten to large for them to manage it.
The fact is that moving Linux to a microkernel is a very non-trivial job, and wouldn't really pay off: it's quite portable and works just fine as it is.
Something just occurred to me: it has been ported to a microkernel - thus producing mklinux.
I'm not sure to what you're referring. First, new bugs are always introduced when you add new code - there is no such thing as a bug-free piece of software. Maybe even your bugfixes have bugs, but less severe - it's a trade-off. I, however, haven't heard of many major bugs standing for more than a week - without an explanation, at least, of why it can't be fixed [in this timeframe, without breaking other stuff]. So-far you're batting.000.
Distributions focussing on ease-of-use? Isn't that kind of their point? Of course you can go out, hex-edit your kernel and boot directly off it, or bootstrap using loadlin or whatever. But the fact is that's non-trivial and for most people very hard. Making an easy installation, or at least one that allows you to not have to download everything, is what a distribution is all about.
As for bloatware, I haven't seen it. Unless you consider things like X, KDE and GNOME bloatware?
When it comes down to setting up a server you use what's right for the job. If that job is serving up.asp pages, you'll use Windows NT. If that job is doing Oracle, you'll probably choose Linux or Solaris. Saying that Linux is completely inappropriate for any server job is kind of like saying that BSD or Solaris is completely appropriate for every server job - ie, it's not true. I don't know about BSD or Solaris, but with most Unices on stable kernels, stability is more a function of hardware reliability and electricity than it is of software. I've heard all kinds of stuff about "[BSD,Solaris] is more stable and faster" but I've never, ever seen anything to back it up. Maybe this was true at one point, but I very much doubt it now.
I really, really want to go to this conference, but two things hold me back: 1, it's in Ottawa (meaning I'd have to take time off work) and 2, it costs $225 - $275, and I probably will have to take the latter seeing as I don't start work till July. Grr, but Alan Cox will be there... why can't it be Toronto!
I've said it once and I'll say it again: Binary-only is not good enough.
I've got a plan for the next computer I build. It will have a Matrox G200, because a) it's probably the best 2D card out there, and b) Matrox released all the specs for the G200. I will be sending a snail mail letter to Matrox & ATI, and possibly Nvidia (depending on their TNT 3D path - source or no source) telling them exactly why I bought what I did. The same thing goes for the sound card. I'll be buying one based on the Trident 4D wave chip - because Trident themselves developed an ALSA driver for it. Creative and Trident wil be getting letters on that.
Why am I planning to go to such lengths? Because I want people to know I support companies who support us. I might never modify my GLX module (or whatever) for the Matrox G200. But I want the ability to do that. I want to know that if I buy myself an Alpha, that it will probably work on that too - and if it doesn't, I can probably make it work there.
Creative: If you want Linux people to buy your cards, give the source of this to Alan Cox for inclusion in his -ac series. It will rapidly get tested, bugfixed, and then I will consider taking you off my 'blacklist' - where you currently enjoy the company of corporations such as ATI. I influence the purchasing habits of a large group of people, and I consistently tell them not to buy ATI cards - because better cards abound. I reccomend against Creative, too.
You can have your intellectual property - but you can't have my money.
Actually, Linux doesn't support over 900-some-odd MB of usable memory without patches, and no more than 2 GB with patches, at least yet. There is talk of allowing the 36-bit memory map of recent Pentium processors to be used, but it hasn't been implemented in a released kernel afaik.
What I'm trying to say is, testing has nothing to do with it. It's well-known that Linux can't address that much memory - on i386 at least.
Is that the evolution of rendering algorithms, of computer systems, or the fact that the Matrix required less computational power than Titanic? After all, weren't 100+ Alphas used for Titanic?
While this is true, Mozilla will fully support the alpha channel in PNG, meaning that in addition to CSS1, XML, etc, we can look forward to using another great image format - very cool.
...that just because mozilla.org pulled it as a request of Netscape, that doesn't mean that mozilla.org has anything to do with Netscape. Netscape developed and (?incorporated?) the pluggable API, and then asked mozilla.org to remove it, for whatever reason. The same thing would happen if Adam Lock were to ask for the ActiveX control to be removed.
However, this is a neat feature that, when Netscape brings it back in, will make Mozilla that much better. A pluggable chat API, so you can write.so:s for AIM, ICQ, or even IRC - and optional, to boot. Yes, Mozilla will kick some serious ass.
I don't care how long it takes. I want the best and only the best for gecko, and if that includes necko then that's fine for me. Navigator is working damned well thus far and I don't anticipate it breaking suddenly. Far more gains will be made if Gecko is nearly perfect, and thus easy to extend and understand, than if it's rushed just to get a "usable product."
You can have old libc5 run-time libraries as well as glibc2.1 installed on the same system; in fact it's pretty much suicide not to. The only thing that saying a distribution is "glibc2.1" does is tell you that all of the programs are, by default, linked with glibc2.1. The old libraries are still there, though, just not used by the majority of programs.
better package manager:apt is pretty damned amazing. apt-get install <package> and you're set. While ports might be as good as that, apt is not inferior in any way.
KDE: currently depends on a non-free Qt. Once Qt 2.0 goes gold, it can be included in main, because the QPL meets the requirements of the DFSG, but until then it stays in contrib and non-free.
more stable: I've never seen anyone show me any sort of evidence either way. When you're dealing with Linux and FreeBSD on PC-style hardware, it is sort of a function of the crap hardware (or not), now isn't it?
FreeBSD may be good, but my pick is Debian. Try both.
In order to run SMP, you have to recompile your kernel, but it's not an ordeal by any means. In the new 2.2 kernels, it's actually an option in make {X,menu}config. While generally speaking, SMP kernels will work on UP boxen, you're going to have some amount of overhead (probably no more than 30%), and so it's good to only compile SMP kernels on SMP boxen.
...assuming, of course, that they keep their mirror up-to-date. ftp.gnome.org is always swamped, and the newest packages that I'm trying to get are not on the mirrors.
No. I'm saying that generally, code generation bugs - which is what gcc 2.7.2.3 has - can be programmed around, but unsupported features of a language - such as namespaces, which I believe is where one of the the real problems lie - are not a simple thing around which to work.
Egcs supports the standard of C++, and gcc doesn't. This is a no-brainer, really.
Check freshmeat for proof. Most of the software announced there is licensed under the GPL; indeed, stuff not under the GPL is the exception.
"Headlining" software is simply software which big companies are releasing as Open Source, and isn't necessarily Free Software - an example is the Apple source license, which allows Apple to terminate your rights to the code - a decidedly non-free aspect of the license.
In any case, most [new] free software projects aren't announced on Slashdot, but most of them do use the GPL.
Harmony is (not was, it's resurrected!) the Free (LGPL) implementation of the library Qt, originally by Troll Tech. It was created because of the license problems (which still exist) with Qt - specifically, distributing binaries of programs linked to Qt, which is a no-no unless Qt is part of your base operating system. In any case, Harmony originally 'died' because of the QPL (Troll Tech's new Open Source (but not Free) license) and the fact that Troll Tech never guaranteed that they wouldn't sue over Harmony.
ROTT was the only game I ever lost a _lot_ of sleep over - because I was up, playing it over the modem with a friend of mine. Truly, that was (and is!) one of the best multiplayer games ever.
Slashdot Mirror
Yes, I ran slackware. I was a real hacker. I singlehandedly installed gnome from source 4 or 5 times, in its entirety, because of how often I nuked my partition and started over.
Then, I realised that I was wasting far too much time compiling things. My Pentium 166 with 32 MB of RAM, which used to be state-of-the-art, now was taking an awfully long time compiling things.
So, for a while, I ran Red Hat. I liked it, too. Its printer management meant I didn't have to futz around with magicfiltres and linuxconf meant that, generally speaking, I didn't have to worry about much.
Then, I found Debian, and my saviour - apt-get.
Simply put, apt-get makes things absurdly easy. I don't have to worry about upgrading to glibc2.1 - apt does the worrying, the downloading, and the installing for me. With Debian I've finally set up the production box I knew Linux could be.
While Slackware will always hold a special place in my heart, for "where I started," Debian is where I am, and where I'm going in the future. Even though something still tells me it's a bit too easy, I just ignore that part of me. It's just too easy to tell my box to upgrade everything on my system to ignore it.
That being said, I'm not planning on installing Realplayer anyhow. The only non-free software I have on my computer is Netscape, and I've got a clear upgrade path for that, too. When something coredumps, I want to be able to fix it - and unfortunately, Realplayer and Netscape do that, sometimes more often than others - and I can't do that with proprietary software.
Why did they report it as 2000? They were probably running a LOT of clients which were connecting and dying off very rapidly, and as someone mentioned above, fork() performance in Linux is stellar. It was probably a case of resolution of benchmarks not being high enough.
As for the 200 Mbps - I would guess that this has to do with the network adaptors. The capacity to have multiple adaptors for the same interface is available, and was developed as part of the Beowulf project. It's probably going to be integrated into 2.3.
In any case, these problems are probably all already fixed, but not tested enough (or wide-spread enough) to be included into a stable distribution's (or kernel's) release.
There is, after all, a difference between sharing ideas about things which are potentially damaging to you with your enemies, and outfitting your enemies with guns. The thought that really underlies proprietary software and crypto regulations in the USA is "we're the only ones who can do this," which is blatantly false, and nieve as well. Sharing thoughts, ideas, and sometimes even implementations is what has changed us from a society of relatively weak technology to a society of high technology in little over 100 years.
What you have to realise, of course, is that moving to microkernel would not eliminate code, simply modularise it - and Linux is already mostly modularised. Also, Linus and Alan Cox have a very good idea of what's going on in all areas of the kernel - ie, it hasn't gotten to large for them to manage it.
The fact is that moving Linux to a microkernel is a very non-trivial job, and wouldn't really pay off: it's quite portable and works just fine as it is.
Something just occurred to me: it has been ported to a microkernel - thus producing mklinux.
Distributions focussing on ease-of-use? Isn't that kind of their point? Of course you can go out, hex-edit your kernel and boot directly off it, or bootstrap using loadlin or whatever. But the fact is that's non-trivial and for most people very hard. Making an easy installation, or at least one that allows you to not have to download everything, is what a distribution is all about.
As for bloatware, I haven't seen it. Unless you consider things like X, KDE and GNOME bloatware?
When it comes down to setting up a server you use what's right for the job. If that job is serving up .asp pages, you'll use Windows NT. If that job is doing Oracle, you'll probably choose Linux or Solaris. Saying that Linux is completely inappropriate for any server job is kind of like saying that BSD or Solaris is completely appropriate for every server job - ie, it's not true. I don't know about BSD or Solaris, but with most Unices on stable kernels, stability is more a function of hardware reliability and electricity than it is of software. I've heard all kinds of stuff about "[BSD,Solaris] is more stable and faster" but I've never, ever seen anything to back it up. Maybe this was true at one point, but I very much doubt it now.
I really, really want to go to this conference, but two things hold me back: 1, it's in Ottawa (meaning I'd have to take time off work) and 2, it costs $225 - $275, and I probably will have to take the latter seeing as I don't start work till July. Grr, but Alan Cox will be there ... why can't it be Toronto!
I've got a plan for the next computer I build. It will have a Matrox G200, because a) it's probably the best 2D card out there, and b) Matrox released all the specs for the G200. I will be sending a snail mail letter to Matrox & ATI, and possibly Nvidia (depending on their TNT 3D path - source or no source) telling them exactly why I bought what I did. The same thing goes for the sound card. I'll be buying one based on the Trident 4D wave chip - because Trident themselves developed an ALSA driver for it. Creative and Trident wil be getting letters on that.
Why am I planning to go to such lengths? Because I want people to know I support companies who support us . I might never modify my GLX module (or whatever) for the Matrox G200. But I want the ability to do that. I want to know that if I buy myself an Alpha, that it will probably work on that too - and if it doesn't, I can probably make it work there.
Creative: If you want Linux people to buy your cards, give the source of this to Alan Cox for inclusion in his -ac series. It will rapidly get tested, bugfixed, and then I will consider taking you off my 'blacklist' - where you currently enjoy the company of corporations such as ATI. I influence the purchasing habits of a large group of people, and I consistently tell them not to buy ATI cards - because better cards abound. I reccomend against Creative, too.
You can have your intellectual property - but you can't have my money.
What I'm trying to say is, testing has nothing to do with it. It's well-known that Linux can't address that much memory - on i386 at least.
way to plug yourself, Mr. Connor ;)
Is that the evolution of rendering algorithms, of computer systems, or the fact that the Matrix required less computational power than Titanic? After all, weren't 100+ Alphas used for Titanic?
While this is true, Mozilla will fully support the alpha channel in PNG, meaning that in addition to CSS1, XML, etc, we can look forward to using another great image format - very cool.
However, this is a neat feature that, when Netscape brings it back in, will make Mozilla that much better. A pluggable chat API, so you can write .so:s for AIM, ICQ, or even IRC - and optional, to boot. Yes, Mozilla will kick some serious ass.
I don't care how long it takes. I want the best and only the best for gecko, and if that includes necko then that's fine for me. Navigator is working damned well thus far and I don't anticipate it breaking suddenly. Far more gains will be made if Gecko is nearly perfect, and thus easy to extend and understand, than if it's rushed just to get a "usable product."
You can have old libc5 run-time libraries as well as glibc2.1 installed on the same system; in fact it's pretty much suicide not to. The only thing that saying a distribution is "glibc2.1" does is tell you that all of the programs are, by default, linked with glibc2.1. The old libraries are still there, though, just not used by the majority of programs.
- better package manager: apt is pretty damned amazing. apt-get install <package> and you're set. While ports might be as good as that, apt is not inferior in any way.
- KDE: currently depends on a non-free Qt. Once Qt 2.0 goes gold, it can be included in main, because the QPL meets the requirements of the DFSG, but until then it stays in contrib and non-free.
- more stable: I've never seen anyone show me any sort of evidence either way. When you're dealing with Linux and FreeBSD on PC-style hardware, it is sort of a function of the crap hardware (or not), now isn't it?
FreeBSD may be good, but my pick is Debian. Try both.Well, after all, GNOME is a GNU project; you'd kind of expect them to call it by that full name. :)
In order to run SMP, you have to recompile your kernel, but it's not an ordeal by any means. In the new 2.2 kernels, it's actually an option in make {X,menu}config. While generally speaking, SMP kernels will work on UP boxen, you're going to have some amount of overhead (probably no more than 30%), and so it's good to only compile SMP kernels on SMP boxen.
...assuming, of course, that they keep their mirror up-to-date. ftp.gnome.org is always swamped, and the newest packages that I'm trying to get are not on the mirrors.
Egcs supports the standard of C++, and gcc doesn't. This is a no-brainer, really.
"Headlining" software is simply software which big companies are releasing as Open Source, and isn't necessarily Free Software - an example is the Apple source license, which allows Apple to terminate your rights to the code - a decidedly non-free aspect of the license.
In any case, most [new] free software projects aren't announced on Slashdot, but most of them do use the GPL.
Harmony is (not was, it's resurrected!) the Free (LGPL) implementation of the library Qt, originally by Troll Tech. It was created because of the license problems (which still exist) with Qt - specifically, distributing binaries of programs linked to Qt, which is a no-no unless Qt is part of your base operating system. In any case, Harmony originally 'died' because of the QPL (Troll Tech's new Open Source (but not Free) license) and the fact that Troll Tech never guaranteed that they wouldn't sue over Harmony.
ROTT was the only game I ever lost a _lot_ of sleep over - because I was up, playing it over the modem with a friend of mine. Truly, that was (and is!) one of the best multiplayer games ever.