This is really funny. I made a nice HTML formatted post, but the slashdot lameness filter prevented me to include any code example, so you can enjoy the post without the code.
I beleive it make the point of the article, but in an ironic way...
Go to <http://www.catseye.mb.ca/esoteric/index.html&g t;. You'll find various extraodinary humourous and clever computer langage and progams.
Befunge, for instance. This is hello world in befunge:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.catseye.mb.ca/esoteric/befunge/src/ hello.bf>
There are hilarious INTERCAL <http://www.tuxedo.org/~esr/intercal> progams too (I spent one night programming the first INTERCAL quine, and it have been probably one of the funniest moment of my life). Every one involved in programming should at least read the manual once in its life. In particular the page related to the precedence of operators. Take a look at the niceties of the language, like the absence of GO TO (replaced by a COME FROM). Check compiler flags, in particular the NOKIDDING one..
The canonical INTERCAL program looks like:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.tuxedo.org/~esr/intercal/intercal.t xt.gz>. Go to section 5.3 (search for 'perfectly clear'). PS manual is at <http://www.tuxedo.org/~esr/intercal/intercal.p s.gz>
The IOCCC is another place to see artistic programs <http://www.ioccc.org/years.html>
The ultra-classic 1984 winner
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.ioccc.org/1984/mullender.c>
And this one. It is a C program that looks like a train. Is it funny ? No. Now look at it:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.ioccc.org/1986/marshall.c>
The ability of beeing able to see the source code made all the difference. Well, should have, but slashdot prevented me to post it. At least I can still link...
> my reluctance has been based more on device drivers, and the ability to get an X or whatever FreeBSD uses for a GUI operating with my hardware
FreeBSD uses the same XFree linux does. FreeBSD is not limited, but, of course, software is generally avalaible a few weeks after linux ones. Not a big deal, unless you want absolute cutting edge.
> I've never really had trouble playing with the source in Linux, it's all available... however I will concede that it sounds as if FreeBSD is much more organized, and that finding the source for a particular program is a much more structured process. I would assume that this is why you say it is good for understanding your system.
That's it. And when you have changed a source, 'make install' will do the right thing.
I played with linux at home a few years. But I can't do full-time unix OS hacking (I am a Objective-C hacker, I only hack OSes in my free time), and always found it painfull to do with linux. The operating system is distributed as binaries. Changing a binary means getting the source, eventually a few dependancies, and compiling it with the right flags (yes, debian is supposed to be easier, but I only played with slackware and, later redhat. I played a few days with debian recently but have not been convinced). Not a difficult process, but you can easily loose one or two hours doing it, and your hacking session is over because it is bedtime. With the kernel, it was worse. Getting the lastest version, discovering it needs updated version of a shitload of utilities, getting them and their dependancies, compiling the stuff, installing it, discovering that the system no longer work, investigating around, etc, etc. A lot of fun, but not an efficient way to hack around.
The _day_ I installed FreeBSD, I found that I did not liked the console mouse (because it lacked acceleration). Did a locate moused. Found it was in/usr/src/usr.sbin/moused. Implemented mouse acceleration (hard-coded a few values, just for the fun), make, make install. I was hooked. Then I discovered that I can rebuild the *whole* os from source within itself. I was amazed (not by the technical achivement, but by my own stupidity: how could i have spent so much time in linux ?:-) )
Last point, related to that:
This is a linux-kernel post, about the oom killer.
(http://www.uwsg.indiana.edu/hypermail/linux/ker ne l/0103.2/1138.html)
------------
>[to various people]
>
>No, ulimit does not work. (But it helps a little.)
>No,/proc/sys/vm/overcommit_memory does not work.
Entirely correct. ulimit certainly makes it much harder for a single
runaway process to take down important parts of the system - now why
doesn't $(MAJOR_DISTRO_VENDOR) set it up by default? NetBSD does. It's
not an infallible solution by any means, but it sure does help.
I just asked a friend to run my test program on his NetBSD box - it ran
into ulimit and malloc() returned 0. Setting ulimit on my RH 6.2 box -
which defaults to unlimited - also caused it to fail gracefully.
------------
It should speak for itself. The BSDs have an edge because the user-space is in sync with the kernel. When something have to be done, it is generally done the 'right' way, because the user-space can be updated at the same time. The result is (IMHO), as system that is cleaner and easier to understand.
It is not that FreeBSD is better than Linux. It is not the case. Linux have more driver support. Linux have a better scalability. But FreeBSD is much cleaner than linux, hence esier to play with. This is why I'd favor BSD for hacking.
I have a laptop with a freebsd that can rebuild itself from scratch. And I develop device drivers in my bed. Maybe not everyone definition of fun, but its mine.
> Since you are clearly using bash as your shell, wich is a GNU shell, I'd say that sometimes that the choerence you are talking about has its execptions.
Congratulations. I knew someone could pinpoint this.
As you guessed:
bash-2.03$ which bash
/usr/local/bin/bash
So this bash is not from the base OS (nothing from the base OS will end in/usr/local). It comes from:
/usr/ports/shells/bash2
Anyway, there is no contradiction between beeing GNU software and beeing in the base freebsd:
bash-2.03$ bc --version
bc 1.05
Copyright 1991, 1992, 1993, 1994, 1997, 1998 Free Software Foundation, Inc.
bash-2.03$ which bc
/usr/bin/bc
bash-2.03$ ls -l/usr/src/usr.bin/bc
ls:/usr/src/usr.bin/bc: No such file or directory
bash-2.03$ ls -l/usr/src/gnu/usr.bin/bc
total 4
drwxrwxr-x 2 2035 207 512 Aug 31 2000 CVS
-rw-r--r-- 1 2035 207 726 Jan 10 2000 Makefile
-rw-r--r-- 1 2035 207 2009 Jan 16 2000 config.h
bash-2.03$ ls -l/usr/src/contrib/bc
total 175
-rw-r--r-- 1 root wheel 117 Apr 29 1998 AUTHORS
-rw-r--r-- 1 root wheel 17984 Apr 29 1998 COPYING
-rw-r--r-- 1 root wheel 27227 Jan 16 2000 ChangeLog
drwxr-xr-x 2 root wheel 512 Aug 31 2000 Examples
-rw-r--r-- 1 root wheel 371 Jan 16 2000 FREEBSD-upgrade
-rw-r--r-- 1 root wheel 7463 Apr 29 1998 INSTALL
-rw-r--r-- 1 root wheel 367 Apr 29 1998 Makefile.am
-rw-r--r-- 1 root wheel 9809 Jan 16 2000 Makefile.in
-rw-r--r-- 1 root wheel 1430 Jan 16 2000 NEWS
-rw-r--r-- 1 root wheel 2578 Apr 29 1998 README
drwxr-xr-x 2 root wheel 512 Aug 31 2000 Test
-rw-r--r-- 1 root wheel 308 Apr 29 1998 acconfig.h
-rw-r--r-- 1 root wheel 4789 Jan 16 2000 aclocal.m4
drwxr-xr-x 2 root wheel 512 Aug 31 2000 bc
-rw-r--r-- 1 root wheel 1765 Jan 16 2000 config.h.in
-rwxr-xr-x 1 root wheel 74880 Jan 16 2000 configure
-rw-r--r-- 1 root wheel 2088 Jan 16 2000 configure.in
drwxr-xr-x 2 root wheel 512 Aug 31 2000 dc
drwxr-xr-x 2 root wheel 512 Aug 31 2000 doc
drwxr-xr-x 2 root wheel 512 Aug 31 2000 h
-rwxr-xr-x 1 root wheel 4771 Apr 29 1998 install-sh
drwxr-xr-x 2 root wheel 512 Aug 31 2000 lib
-rwxr-xr-x 1 root wheel 4639 Apr 29 1998 missing
-rwxr-xr-x 1 root wheel 730 Apr 29 1998 mkinstalldirs
-rw-r--r-- 1 root wheel 10 Apr 29 1998 stamp-h.in
bash-2.03$
So yes, the 'coherence' I hyped in the original mail is not exactly as simple as I presented it, but is still predictable (there is contributed code in freebsd base system, so GNU bc is freebsd's bc).
> What exactly is the differences between BSD and Linux?
Let's hope that I don't reply to a troll.
I'd talk about FreeBSD vs Linux.
The most important difference between linux and FreeBSD is that linux is a kernel while FreeBSD is a server operating system.
This is a serious difference. Linus torvald, or alan cox have zero power on deciding what initialisation scripts should looks like, or what the cron program should be. This is decided by 'distributors'. In general, people use a set of GNU tools on top of the linux kernel, and the resulting is called GNU/linux. But a linux distribution contains much more than only GNU tools. No one really knows what is and what is not linux-the-operating-system.
On the opposite, FreeBSD is a kernel plus a user-land. The kernel is designed to run with this userland, and is distributing it separately would make no sense. The linux concept of updating the kernel is alien to FreeBSD, in which you would upgrade the whole system (You should take into account that upgrading the whole FreeBSD system is probably easier than updating the linux kernel).
The result is that there are many flavors of linux, while there is only one FreeBSD. It is a good, and a bad point. The good side, is that a FreeBSD system is orders of magnitude more coherent than a linux system. It is much more easy to learn and tweak, because sources to the whole system are in/usr/src, not only the sources of the kernel. The bad side is that it is probably more boring. The other bad side is that there is a distinction between say Mozilla and top under FreeBSD (Mozilla is not part of freebsd, while top is). Under linux there is no hard disctinctions.
I said that freebsd is more coherent. Let's give you a couple of random examples:
bash-2.03$ which ls
/bin/ls
This means that ls is a FreeBSD command. It is not your everyday linux ls is, it is the freebsd one. Its sources are located in/usr/src/bin/ls:
No configure/autoconf, no README, no LICENSE. Just the meat, plain and simple. The makefile is trivial. Change. make. run. make install. FreeBSD is a joy to hack.
Okay. another example.
bash-2.03$ man -k ATAPI
ata(4), acd(4), ad(4), afd(4), ast(4) - Generic ATA/ATAPI disk controller driver
burncd(8) - control the ATAPI CD-R/RW driver
wfd(4) - ATAPI floppy driver (LS-120 floppy driver)
wst(4) - ATAPI Tape drive
Yep. Every (okay, mosts) device driver have its own man page.
Yet another example:
bash-2.03$ man 9 intro | head
INTRO(9) FreeBSD Kernel Developer's Manual INTRO(9)
NAME
intro - introduction to system kernel interfaces
DESCRIPTION
This section contains information about the interfaces and subroutines in
the kernel.
[...]
Most important kernel routines have their man pages, with usage and example.
NAME
uio, uiomove - device driver IO routines
[...]
This coherence is visible in the configuration of freebsd. Configuration is made in flat-files, in a pure unix way, but those are coherent. For instance:
Last thing, to update the whole system to latest version, you use 'cvsup -L 2 stable-supfile' and your sources are up-to-date. Then you do a single 'make installworld' and your system is up to date. From source. Every single bit (that is considered in the system, of course).
OpenBSD and NetBSD are different operating systems, but share the same spirit as FreeBSD. OpenBSD is target at security, while NetBSD is targeted at portability.
At the bottom line:
* Linux is great if you want a binary cutting-edge unix-like OS.
* BSD is great when you want a cleanly designed server system.
* BSD is great if you like to hack/understand your system
And, well, the obvious: linux is more popular, and BSD zealots are superior assholes.
OK. Let's say I remove this comment about the placement of your head.:-)
You know, I just commented on your first sentence, and I must admit nor having read the rest of your post. So much of intellectual honesty. You were such an easy target...
> I would go into a long rant here about my personal belief that unweildiness of Mozilla
That would be interesting. I find mozilla code awful, and beleive that the original sin was to make 'dynamic' code with C++. When I look at the code, I pity them, as they took great amount of pain to code in C++ things that would have been natural with Objective-C. Of course, I am biased on this:-)
> Does anyone really think that the internet would be growing this fast if it weren't for MS.
You mean, if Microsoft did not consider Internet un-important and tried to promote their own non-internet MSN service instead in Windows 95 A ?
You need to stop beleiving the bullshit redmond throws at you. Microsoft have been surprised by internet, under-estimated the phenomenon, and tried (and failed) to control it. Now they want to make you believe that credit is due, while they did their best to slow it down.
The only credit Microsoft have is the very impressive turnaround they did in 1996-98. I would never have beleived that such a big company could react that fast.
> first off, Creating something like BIND is infinitely more difficult than something like MSIE--
Gently put your head out of your ass. You obviously don't know what you are talking about. Bind is a two-banana hack compared to MSIE. MSIE have about the same complexity as Mozilla. Ever looked at mozilla source code ? Ever tried to build it ? Now take a look at BIND source code. Build it. Draw you conclusion in term of complexity.
A BIND bug is very serious because it can compromise huge segments of the network. But people that run BIND know what they are doing (or should know). And there are alternatives.
A MSIE bug is very serious because it can compromise a huge number of individual hosts. Furthermore, people don't choose to run MSIE, they have to, or they just don't know that they are running it. And you can't remove MSIE from a windows machine.
So, IMNSHO, a MSIE bug is more serious than a BIND bug.
RMS: There it is! The AOL Server of Death!
OSCAR: Oh, great.
AIM CLIENT: Look!
RMS: There's the server from 64.12.149.13!
ESR: What is he doing here?
RMS: He is the AOL Server of Death. He asks each client five questions -
AIM CLIENT: Three questions.
RMS: Three questions. He who answers the five questions -
AIM CLIENT: Three questions.
RMS: Three questions may chat in safety.
OSCAR: What if you get a question wrong?
RMS: Then you are cast into void.
OSCAR: Oh, I won't go.
???: Who's going to answer the questions?
RMS: Sir OSCAR!
OSCAR: Yes?
RMS: Brave Sir OSCAR, you go.
OSCAR: Hey! I've got a great idea. Why doesn't AIM CLIENT go?
AIM CLIENT: Yes, let me go, my liege. I will take him single-handed. I shall make a feint to the north-east -
RMS: No, no, hang on hang on hang on! Just answer the five questions -
AIM CLIENT: Three questions.
RMS: Three questions as best you can. And we shall watch... and pray.
AIM CLIENT: I understand, my liege.
RMS: Good luck, brave AIM CLIENT. God be with you.
AOL: Stop! Who would chat with the Server of Death must answer me these
questions three, 'ere the other side he see.
AIM CLIENT: Ask me the questions, bridge-AOL. I'm not afraid.
AOL: What is your name?
AIM CLIENT: My name is Sir AIM CLIENT of America Online.
AOL: What is your quest?
AIM CLIENT: To chat with Clueless People.
AOL: What is your favorite color?
AIM CLIENT: 42.
AOL: Right. Off you go.
AIM CLIENT: Oh, thank you. Thank you very much.
OSCAR: Oh that's easy!
AOL: Stop! Who approaches the Bridge of Death must answer me these questions three, 'ere the other side he see.
OSCAR: Ask me the questions, bridge-AOL. I'm not afraid.
AOL: What is your name?
OSCAR: Sir OSCAR of Open Source.
AOL: What is your quest?
OSCAR: To chat with Clueless People.
AOL: What is the MD5 of AIM.EXE ?
OSCAR: I don't know that! Auuuuuuuugh! (OSCAR get disconnected)
Requesting the MD5 of random parts of aim.exe is somewhat clever (but probably unnecessary complex. They could have achieved the very same effect by simply asking 16 bytes of the exe file)
What I find much more subtle, is the fact that they just started requested it. The thing was burried in the exe, and they were waiting before using it. Reminds me of DirectTV <http://slashdot.org/articles/01/01/25/1343218. shtml>
Corporations are getting smart. If AOL had always requested the MD5, then competitive clients could not have been developed, and they would have run the risk of getting an open-source competitor.
I find quite scary to see that closed-source software contains such hidden protocol tricks. Who knows what more is contained in the various exe people runs daily ? What's going to happen if IIS start requesting random MD5 of.exe files when the client is Internet Explorer ?
> TiVo has actually gone to lengths to make sure they CAN'T associate the data
First, it was supposed to be funny.
Second, they can match the data, with the correlation of the IP addresses between the two uploads.
Third, the comm is now crypted, so they could easily pass private info without anyone knowledge.
Fourth, as most of the recent privacy issues have proved us, they can easily change their privacy policy (for instance, but releasing individual but anonymous data, so the advertisers would not know who you are, but whould know what you watch)
> How about the industry becomes unprofitable because those in the hardware business (making chips...not selling a computer) are generally not making their cash in the services.
Nothing will prevent them to make chips. And, as a capitalist, you should think that it would even make a more efficient market and be utltimately good for the consumer.
> I am a capitalist. I believe that companies should be able to make money.
Like rambus or like AMD ? By locking the IP or by making a better/cheaper/more attractive product ? Only if it is the former, then I understand why you would object to open-source hardware.
> The Open Source model works so wll for software while companies can sell services while making less money on the software.
Well, 'Free Processor' would still cost money. Even if silicon wants to be free.:-)
What the 'Open Source' model does is to provide everyone with the bare minimum of software. It turned the OS, the text editor and the compiler (for instance) into a commodity. Much more importantly, it releived us from 'vendor locks'.
What an 'Open Source' model on hardware would do is to provide everyone with the bare minimum of hardware compatibility, and save people from vendor locks. I expect that you understand that having standard hardware (from isa-boards to usb devices) have been a good thing for the computer industry. Escaping the IBM vendor lock on PC have been one of the best things that occured to computer users.
[crunch]
> For example, Company X, Y and Z make these cpus. Company X employs 4 people to work on the spec of this chip so they have their inhouse employees. Company Y employs 2 people. And Company Z employs none. Company Z has less overhead of employees to work on the product and can sell the same thing for less. That puts Company X out of business. In order to compete, Company Y has to lay of its employees that work on the product.
Making millions of highly clocked processor will still require a lot of skill, so I would not expect intel to be put out of business any day. Furthermore, open hardware can peacefully(?) coexist with closed hardware. The existence of GNU/linux did not prevent thousands of people waiting in line at each release of Windows (or Mac OS X).
Open hardware may even help close closed companies. For instance FreeBSD enables Apple to cget back in the competition with Microsoft.
> Where do we make our money?
Your example may as well be: Company X employs 4 people to work on the spec of this chip so they can be the first to implement next revisions with higher-clock. Company Y employs 2 people, but they work on making a different form-factor for the processor, so it can be embedded in wrist-watches. Company Z employs none and goes out of business because it don't produce anything relevant.
I don't know, it just make sense to me. Maybe I am not a capitalist...:-)
> I guess my feelings are known because I believe that hardware's design should not be completely open
Would you care to explain us why ?
I see advantages in the fact that:
1/ Everyone (with the knowledge to understand it) can see how the processor is done
2/ Everyone (with the motivation) can learn from the processor core
3/ Everyone (with the foundry) can manufacture a totally compatible processor
4/ Everyone (with the technical skils) can improve the procesor design
5/ (A few other ones)
Of course you are not going to run a plant in your backyard, but I fail to see real downsides to the idea.
Last time I checked, it was "pax". And don't get me started on this, please, as pax have, well, some rough corners... (like replacing carefully symbolic links by real hierarchies, or changing permission on directories unrelated to the install at hand)
1/ OpenSTEP capitalisation was OpenStep (the spec) or OPENSTEP (the OS)
2/ OPENSTEP relied heavily on DPS. DPS was licensed from Adobe. The story says that Adobe refused to extend license.
3/ The silver bullet was not OPENSTEP, but the YellowBox (aka OPENSTEP/Enterprise, aka OpenStep for Windows NT, a way to run OpenStep applications under NT)
Yes, CmdrTaco, this is unfair. It is 19h50 in europe. Just when european have to leave work, you post such a trollish article ? This is unfair for us, europeean trolls. (Btw, Cmdr, is it because you learn that Mac
OS X supports 2 buttons that you are okay to try it ?)
Lastly, Mac OS X is not sweet because it brings unix to the masses, but because it bring a real object oriented system to the desktop.
NeXT promised, Apple delivers.
I started to develop for NeXTstep in 1991. The moto was "NeXT have ten years of advance". I didn't expected it to be that true.
> Unfortunately even to this day I find loads of sys admins crawling all over their DMZ-level servers with Telnet
Would you be kind enough to give us more information about those individuals, please ? I'd love to have them as friends, we could do great things together...
> The only thing more frightening than your statement was the speed in which your statement was voted up.
Yeah. At first read I thought he was trying to make a point, like, they'll see that this stupidy cost $50 to $100M, but it looks like he actually beleive what he wrote./bots are probably moderating him just because of its low UID.
...but while many sites want to go pay-for-no-ads (salon, penisarcade, k5, for instance), one of the rare site that provided content I could pay for is moving into the opposite (ie: free without ad, while there was a membership fee before). (They plan innovative ways to finance the site).
Slashdot Mirror
Posting here to remove my mod and getting back the +5:Troll this deserves...
Cheers,
--fred
This is really funny. I made a nice HTML formatted post, but the slashdot lameness filter prevented me to include any code example, so you can enjoy the post without the code.
I beleive it make the point of the article, but in an ironic way...
Go to <http://www.catseye.mb.ca/esoteric/index.html&g t;. You'll find various extraodinary humourous and clever computer langage and progams.
Befunge, for instance. This is hello world in befunge:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.catseye.mb.ca/esoteric/befunge/src/ hello.bf>
There are hilarious INTERCAL <http://www.tuxedo.org/~esr/intercal> progams too (I spent one night programming the first INTERCAL quine, and it have been probably one of the funniest moment of my life). Every one involved in programming should at least read the manual once in its life. In particular the page related to the precedence of operators. Take a look at the niceties of the language, like the absence of GO TO (replaced by a COME FROM). Check compiler flags, in particular the NOKIDDING one..
The canonical INTERCAL program looks like:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.tuxedo.org/~esr/intercal/intercal.t xt.gz>. Go to section 5.3 (search for 'perfectly clear'). PS manual is at <http://www.tuxedo.org/~esr/intercal/intercal.p s.gz>
The IOCCC is another place to see artistic programs <http://www.ioccc.org/years.html>
The ultra-classic 1984 winner
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.ioccc.org/1984/mullender.c>
And this one. It is a C program that looks like a train. Is it funny ? No. Now look at it:
CODE REMOVED DUE TO DMCA^W LAMENESS FILTER.
See <http://www.ioccc.org/1986/marshall.c>
The ability of beeing able to see the source code made all the difference. Well, should have, but slashdot prevented me to post it. At least I can still link...
Cheers,
--fred
is a state of mind
-- #1000 --
(Posting logged)
> my reluctance has been based more on device drivers, and the ability to get an X or whatever FreeBSD uses for a GUI operating with my hardware
/usr/src/usr.sbin/moused. Implemented mouse acceleration (hard-coded a few values, just for the fun), make, make install. I was hooked. Then I discovered that I can rebuild the *whole* os from source within itself. I was amazed (not by the technical achivement, but by my own stupidity: how could i have spent so much time in linux ? :-) )
r ne l/0103.2/1138.html)
/proc/sys/vm/overcommit_memory does not work.
FreeBSD uses the same XFree linux does. FreeBSD is not limited, but, of course, software is generally avalaible a few weeks after linux ones. Not a big deal, unless you want absolute cutting edge.
> I've never really had trouble playing with the source in Linux, it's all available... however I will concede that it sounds as if FreeBSD is much more organized, and that finding the source for a particular program is a much more structured process. I would assume that this is why you say it is good for understanding your system.
That's it. And when you have changed a source, 'make install' will do the right thing.
I played with linux at home a few years. But I can't do full-time unix OS hacking (I am a Objective-C hacker, I only hack OSes in my free time), and always found it painfull to do with linux. The operating system is distributed as binaries. Changing a binary means getting the source, eventually a few dependancies, and compiling it with the right flags (yes, debian is supposed to be easier, but I only played with slackware and, later redhat. I played a few days with debian recently but have not been convinced). Not a difficult process, but you can easily loose one or two hours doing it, and your hacking session is over because it is bedtime. With the kernel, it was worse. Getting the lastest version, discovering it needs updated version of a shitload of utilities, getting them and their dependancies, compiling the stuff, installing it, discovering that the system no longer work, investigating around, etc, etc. A lot of fun, but not an efficient way to hack around.
The _day_ I installed FreeBSD, I found that I did not liked the console mouse (because it lacked acceleration). Did a locate moused. Found it was in
Last point, related to that:
This is a linux-kernel post, about the oom killer.
(http://www.uwsg.indiana.edu/hypermail/linux/ke
------------
>[to various people]
>
>No, ulimit does not work. (But it helps a little.)
>No,
Entirely correct. ulimit certainly makes it much harder for a single
runaway process to take down important parts of the system - now why
doesn't $(MAJOR_DISTRO_VENDOR) set it up by default? NetBSD does. It's
not an infallible solution by any means, but it sure does help.
I just asked a friend to run my test program on his NetBSD box - it ran
into ulimit and malloc() returned 0. Setting ulimit on my RH 6.2 box -
which defaults to unlimited - also caused it to fail gracefully.
------------
It should speak for itself. The BSDs have an edge because the user-space is in sync with the kernel. When something have to be done, it is generally done the 'right' way, because the user-space can be updated at the same time. The result is (IMHO), as system that is cleaner and easier to understand.
It is not that FreeBSD is better than Linux. It is not the case. Linux have more driver support. Linux have a better scalability. But FreeBSD is much cleaner than linux, hence esier to play with. This is why I'd favor BSD for hacking.
I have a laptop with a freebsd that can rebuild itself from scratch. And I develop device drivers in my bed. Maybe not everyone definition of fun, but its mine.
Cheers,
--fred
> Since you are clearly using bash as your shell, wich is a GNU shell, I'd say that sometimes that the choerence you are talking about has its execptions.
/usr/local). It comes from:
/usr/src/usr.bin/bc
/usr/src/usr.bin/bc: No such file or directory
/usr/src/gnu/usr.bin/bc
/usr/src/contrib/bc
Congratulations. I knew someone could pinpoint this.
As you guessed:
bash-2.03$ which bash
/usr/local/bin/bash
So this bash is not from the base OS (nothing from the base OS will end in
/usr/ports/shells/bash2
Anyway, there is no contradiction between beeing GNU software and beeing in the base freebsd:
bash-2.03$ bc --version
bc 1.05
Copyright 1991, 1992, 1993, 1994, 1997, 1998 Free Software Foundation, Inc.
bash-2.03$ which bc
/usr/bin/bc
bash-2.03$ ls -l
ls:
bash-2.03$ ls -l
total 4
drwxrwxr-x 2 2035 207 512 Aug 31 2000 CVS
-rw-r--r-- 1 2035 207 726 Jan 10 2000 Makefile
-rw-r--r-- 1 2035 207 2009 Jan 16 2000 config.h
bash-2.03$ ls -l
total 175
-rw-r--r-- 1 root wheel 117 Apr 29 1998 AUTHORS
-rw-r--r-- 1 root wheel 17984 Apr 29 1998 COPYING
-rw-r--r-- 1 root wheel 27227 Jan 16 2000 ChangeLog
drwxr-xr-x 2 root wheel 512 Aug 31 2000 Examples
-rw-r--r-- 1 root wheel 371 Jan 16 2000 FREEBSD-upgrade
-rw-r--r-- 1 root wheel 7463 Apr 29 1998 INSTALL
-rw-r--r-- 1 root wheel 367 Apr 29 1998 Makefile.am
-rw-r--r-- 1 root wheel 9809 Jan 16 2000 Makefile.in
-rw-r--r-- 1 root wheel 1430 Jan 16 2000 NEWS
-rw-r--r-- 1 root wheel 2578 Apr 29 1998 README
drwxr-xr-x 2 root wheel 512 Aug 31 2000 Test
-rw-r--r-- 1 root wheel 308 Apr 29 1998 acconfig.h
-rw-r--r-- 1 root wheel 4789 Jan 16 2000 aclocal.m4
drwxr-xr-x 2 root wheel 512 Aug 31 2000 bc
-rw-r--r-- 1 root wheel 1765 Jan 16 2000 config.h.in
-rwxr-xr-x 1 root wheel 74880 Jan 16 2000 configure
-rw-r--r-- 1 root wheel 2088 Jan 16 2000 configure.in
drwxr-xr-x 2 root wheel 512 Aug 31 2000 dc
drwxr-xr-x 2 root wheel 512 Aug 31 2000 doc
drwxr-xr-x 2 root wheel 512 Aug 31 2000 h
-rwxr-xr-x 1 root wheel 4771 Apr 29 1998 install-sh
drwxr-xr-x 2 root wheel 512 Aug 31 2000 lib
-rwxr-xr-x 1 root wheel 4639 Apr 29 1998 missing
-rwxr-xr-x 1 root wheel 730 Apr 29 1998 mkinstalldirs
-rw-r--r-- 1 root wheel 10 Apr 29 1998 stamp-h.in
bash-2.03$
So yes, the 'coherence' I hyped in the original mail is not exactly as simple as I presented it, but is still predictable (there is contributed code in freebsd base system, so GNU bc is freebsd's bc).
Cheers,
--fred
> What exactly is the differences between BSD and Linux?
/usr/src, not only the sources of the kernel. The bad side is that it is probably more boring. The other bad side is that there is a distinction between say Mozilla and top under FreeBSD (Mozilla is not part of freebsd, while top is). Under linux there is no hard disctinctions.
/usr/src/bin/ls:
/usr/src/bin/ls
/etc/rc.conf
:-)
Let's hope that I don't reply to a troll.
I'd talk about FreeBSD vs Linux.
The most important difference between linux and FreeBSD is that linux is a kernel while FreeBSD is a server operating system.
This is a serious difference. Linus torvald, or alan cox have zero power on deciding what initialisation scripts should looks like, or what the cron program should be. This is decided by 'distributors'. In general, people use a set of GNU tools on top of the linux kernel, and the resulting is called GNU/linux. But a linux distribution contains much more than only GNU tools. No one really knows what is and what is not linux-the-operating-system.
On the opposite, FreeBSD is a kernel plus a user-land. The kernel is designed to run with this userland, and is distributing it separately would make no sense. The linux concept of updating the kernel is alien to FreeBSD, in which you would upgrade the whole system (You should take into account that upgrading the whole FreeBSD system is probably easier than updating the linux kernel).
The result is that there are many flavors of linux, while there is only one FreeBSD. It is a good, and a bad point. The good side, is that a FreeBSD system is orders of magnitude more coherent than a linux system. It is much more easy to learn and tweak, because sources to the whole system are in
I said that freebsd is more coherent. Let's give you a couple of random examples:
bash-2.03$ which ls
/bin/ls
This means that ls is a FreeBSD command. It is not your everyday linux ls is, it is the freebsd one. Its sources are located in
bash-2.03$ cd
bash-2.03$ ls -l
total 61
-rw-r--r-- 1 root wheel 274 Jun 17 2000 Makefile
-rw-r--r-- 1 root wheel 3066 Aug 28 1999 cmp.c
-rw-r--r-- 1 root wheel 2798 Jul 5 2000 extern.h
-rw-r--r-- 1 root wheel 13984 Feb 13 10:50 ls.1
-rw-r--r-- 1 root wheel 18557 Aug 13 2000 ls.c
-rw-r--r-- 1 root wheel 3172 Jul 5 2000 ls.h
-rw-r--r-- 1 root wheel 11618 Jul 5 2000 print.c
-rw-r--r-- 1 root wheel 4579 Jul 22 2000 util.c
bash-2.03$
No configure/autoconf, no README, no LICENSE. Just the meat, plain and simple. The makefile is trivial. Change. make. run. make install. FreeBSD is a joy to hack.
Okay. another example.
bash-2.03$ man -k ATAPI
ata(4), acd(4), ad(4), afd(4), ast(4) - Generic ATA/ATAPI disk controller driver
burncd(8) - control the ATAPI CD-R/RW driver
wfd(4) - ATAPI floppy driver (LS-120 floppy driver)
wst(4) - ATAPI Tape drive
Yep. Every (okay, mosts) device driver have its own man page.
Yet another example:
bash-2.03$ man 9 intro | head
INTRO(9) FreeBSD Kernel Developer's Manual INTRO(9)
NAME
intro - introduction to system kernel interfaces
DESCRIPTION
This section contains information about the interfaces and subroutines in
the kernel.
[...]
Most important kernel routines have their man pages, with usage and example.
bash-2.03$ man 9 uio
Formatting page, please wait...Done.
UIO(9) FreeBSD Kernel Developer's Manual UIO(9)
NAME
uio, uiomove - device driver IO routines
[...]
This coherence is visible in the configuration of freebsd. Configuration is made in flat-files, in a pure unix way, but those are coherent. For instance:
cat
[...]
nfs_server_enable="YES"
sendmail_enable="NO"
check_quotas="NO"
portmap_enable="YES"
inetd_enable="NO"
allscreens_flags="-m on -g 100x37 VESA_800x600"
moused_port="/dev/cuaa0"
moused_flags=""
moused_type="mouseman"
moused_enable="YES"
ntpdate_enable="YES"
ntpdate_flags="ntp.apple.com"
sshd_enable="YES"
[...]
I think you get the idea.
Last thing, to update the whole system to latest version, you use 'cvsup -L 2 stable-supfile' and your sources are up-to-date. Then you do a single 'make installworld' and your system is up to date. From source. Every single bit (that is considered in the system, of course).
OpenBSD and NetBSD are different operating systems, but share the same spirit as FreeBSD. OpenBSD is target at security, while NetBSD is targeted at portability.
At the bottom line:
* Linux is great if you want a binary cutting-edge unix-like OS.
* BSD is great when you want a cleanly designed server system.
* BSD is great if you like to hack/understand your system
And, well, the obvious: linux is more popular, and BSD zealots are superior assholes.
Ooops, forgot to talk about the port tree...
Cheers,
--fred
> OK. My apologies. :)
:-)
:-)
OK. Let's say I remove this comment about the placement of your head.
You know, I just commented on your first sentence, and I must admit nor having read the rest of your post. So much of intellectual honesty. You were such an easy target...
> I would go into a long rant here about my personal belief that unweildiness of Mozilla
That would be interesting. I find mozilla code awful, and beleive that the original sin was to make 'dynamic' code with C++. When I look at the code, I pity them, as they took great amount of pain to code in C++ things that would have been natural with Objective-C. Of course, I am biased on this
Cheers,
--fred
* This user has donated to the EFF and written his congressman.
You would have had much bigger influence if you only wrote to the EFF and donated to your congressman.
> Does anyone really think that the internet would be growing this fast if it weren't for MS.
You mean, if Microsoft did not consider Internet un-important and tried to promote their own non-internet MSN service instead in Windows 95 A ?
You need to stop beleiving the bullshit redmond throws at you. Microsoft have been surprised by internet, under-estimated the phenomenon, and tried (and failed) to control it. Now they want to make you believe that credit is due, while they did their best to slow it down.
The only credit Microsoft have is the very impressive turnaround they did in 1996-98. I would never have beleived that such a big company could react that fast.
Cheers,
--fred
> first off, Creating something like BIND is infinitely more difficult than something like MSIE--
Gently put your head out of your ass. You obviously don't know what you are talking about. Bind is a two-banana hack compared to MSIE. MSIE have about the same complexity as Mozilla. Ever looked at mozilla source code ? Ever tried to build it ? Now take a look at BIND source code. Build it. Draw you conclusion in term of complexity.
A BIND bug is very serious because it can compromise huge segments of the network. But people that run BIND know what they are doing (or should know). And there are alternatives.
A MSIE bug is very serious because it can compromise a huge number of individual hosts. Furthermore, people don't choose to run MSIE, they have to, or they just don't know that they are running it. And you can't remove MSIE from a windows machine.
So, IMNSHO, a MSIE bug is more serious than a BIND bug.
Cheers,
--fred
RMS: There it is! The AOL Server of Death!
OSCAR: Oh, great.
AIM CLIENT: Look!
RMS: There's the server from 64.12.149.13!
ESR: What is he doing here?
RMS: He is the AOL Server of Death. He asks each client five questions -
AIM CLIENT: Three questions.
RMS: Three questions. He who answers the five questions -
AIM CLIENT: Three questions.
RMS: Three questions may chat in safety.
OSCAR: What if you get a question wrong?
RMS: Then you are cast into void.
OSCAR: Oh, I won't go.
???: Who's going to answer the questions?
RMS: Sir OSCAR!
OSCAR: Yes?
RMS: Brave Sir OSCAR, you go.
OSCAR: Hey! I've got a great idea. Why doesn't AIM CLIENT go?
AIM CLIENT: Yes, let me go, my liege. I will take him single-handed. I shall make a feint to the north-east -
RMS: No, no, hang on hang on hang on! Just answer the five questions -
AIM CLIENT: Three questions.
RMS: Three questions as best you can. And we shall watch... and pray.
AIM CLIENT: I understand, my liege.
RMS: Good luck, brave AIM CLIENT. God be with you.
AOL: Stop! Who would chat with the Server of Death must answer me these
questions three, 'ere the other side he see.
AIM CLIENT: Ask me the questions, bridge-AOL. I'm not afraid.
AOL: What is your name?
AIM CLIENT: My name is Sir AIM CLIENT of America Online.
AOL: What is your quest?
AIM CLIENT: To chat with Clueless People.
AOL: What is your favorite color?
AIM CLIENT: 42.
AOL: Right. Off you go.
AIM CLIENT: Oh, thank you. Thank you very much.
OSCAR: Oh that's easy!
AOL: Stop! Who approaches the Bridge of Death must answer me these questions three, 'ere the other side he see.
OSCAR: Ask me the questions, bridge-AOL. I'm not afraid.
AOL: What is your name?
OSCAR: Sir OSCAR of Open Source.
AOL: What is your quest?
OSCAR: To chat with Clueless People.
AOL: What is the MD5 of AIM.EXE ?
OSCAR: I don't know that! Auuuuuuuugh! (OSCAR get disconnected)
Requesting the MD5 of random parts of aim.exe is somewhat clever (but probably unnecessary complex. They could have achieved the very same effect by simply asking 16 bytes of the exe file)
.exe files when the client is Internet Explorer ?
What I find much more subtle, is the fact that they just started requested it. The thing was burried in the exe, and they were waiting before using it. Reminds me of DirectTV <http://slashdot.org/articles/01/01/25/1343218. shtml>
Corporations are getting smart. If AOL had always requested the MD5, then competitive clients could not have been developed, and they would have run the risk of getting an open-source competitor.
I find quite scary to see that closed-source software contains such hidden protocol tricks. Who knows what more is contained in the various exe people runs daily ? What's going to happen if IIS start requesting random MD5 of
Cheers,
--fred
> TiVo has actually gone to lengths to make sure they CAN'T associate the data
First, it was supposed to be funny.
Second, they can match the data, with the correlation of the IP addresses between the two uploads.
Third, the comm is now crypted, so they could easily pass private info without anyone knowledge.
Fourth, as most of the recent privacy issues have proved us, they can easily change their privacy policy (for instance, but releasing individual but anonymous data, so the advertisers would not know who you are, but whould know what you watch)
Cheers,
--fred
It probably means that they need to do a SQL join to associate viewing habits to a particular customer.
Cheers,
--fred
> How about the industry becomes unprofitable because those in the hardware business (making chips...not selling a computer) are generally not making their cash in the services.
:-)
:-)
Nothing will prevent them to make chips. And, as a capitalist, you should think that it would even make a more efficient market and be utltimately good for the consumer.
> I am a capitalist. I believe that companies should be able to make money.
Like rambus or like AMD ? By locking the IP or by making a better/cheaper/more attractive product ? Only if it is the former, then I understand why you would object to open-source hardware.
> The Open Source model works so wll for software while companies can sell services while making less money on the software.
Well, 'Free Processor' would still cost money. Even if silicon wants to be free.
What the 'Open Source' model does is to provide everyone with the bare minimum of software. It turned the OS, the text editor and the compiler (for instance) into a commodity. Much more importantly, it releived us from 'vendor locks'.
What an 'Open Source' model on hardware would do is to provide everyone with the bare minimum of hardware compatibility, and save people from vendor locks. I expect that you understand that having standard hardware (from isa-boards to usb devices) have been a good thing for the computer industry. Escaping the IBM vendor lock on PC have been one of the best things that occured to computer users.
[crunch]
> For example, Company X, Y and Z make these cpus. Company X employs 4 people to work on the spec of this chip so they have their inhouse employees. Company Y employs 2 people. And Company Z employs none. Company Z has less overhead of employees to work on the product and can sell the same thing for less. That puts Company X out of business. In order to compete, Company Y has to lay of its employees that work on the product.
Making millions of highly clocked processor will still require a lot of skill, so I would not expect intel to be put out of business any day. Furthermore, open hardware can peacefully(?) coexist with closed hardware. The existence of GNU/linux did not prevent thousands of people waiting in line at each release of Windows (or Mac OS X).
Open hardware may even help close closed companies. For instance FreeBSD enables Apple to cget back in the competition with Microsoft.
> Where do we make our money?
Your example may as well be: Company X employs 4 people to work on the spec of this chip so they can be the first to implement next revisions with higher-clock. Company Y employs 2 people, but they work on making a different form-factor for the processor, so it can be embedded in wrist-watches. Company Z employs none and goes out of business because it don't produce anything relevant.
I don't know, it just make sense to me. Maybe I am not a capitalist...
Cheers,
--fred
> I guess my feelings are known because I believe that hardware's design should not be completely open
Would you care to explain us why ?
I see advantages in the fact that:
1/ Everyone (with the knowledge to understand it) can see how the processor is done
2/ Everyone (with the motivation) can learn from the processor core
3/ Everyone (with the foundry) can manufacture a totally compatible processor
4/ Everyone (with the technical skils) can improve the procesor design
5/ (A few other ones)
Of course you are not going to run a plant in your backyard, but I fail to see real downsides to the idea.
Cheers,
--fred
Last time I checked, it was "pax". And don't get me started on this, please, as pax have, well, some rough corners... (like replacing carefully symbolic links by real hierarchies, or changing permission on directories unrelated to the install at hand)
Cheers,
--fred
1/ OpenSTEP capitalisation was OpenStep (the spec) or OPENSTEP (the OS)
2/ OPENSTEP relied heavily on DPS. DPS was licensed from Adobe. The story says that Adobe refused to extend license.
3/ The silver bullet was not OPENSTEP, but the YellowBox (aka OPENSTEP/Enterprise, aka OpenStep for Windows NT, a way to run OpenStep applications under NT)
Cheers,
--fred
Yes, CmdrTaco, this is unfair. It is 19h50 in europe. Just when european have to leave work, you post such a trollish article ? This is unfair for us, europeean trolls. (Btw, Cmdr, is it because you learn that Mac
OS X supports 2 buttons that you are okay to try it ?)
Lastly, Mac OS X is not sweet because it brings unix to the masses, but because it bring a real object oriented system to the desktop.
NeXT promised, Apple delivers.
I started to develop for NeXTstep in 1991. The moto was "NeXT have ten years of advance". I didn't expected it to be that true.
Cheers,
--fred
> Unfortunately even to this day I find loads of sys admins crawling all over their DMZ-level servers with Telnet
Would you be kind enough to give us more information about those individuals, please ? I'd love to have them as friends, we could do great things together...
Cheers,
--fred
>> This will save a lot of people from typing 'ln -s /usr/bin/secsh /usr/bin/ssh'.
/usr/bin/ssh' is enough. Now, the IETF should probably reconsider its decision, as the trouble have been vastly overestimated...
> Whew! I can't tell you how much that relieves me.
Well, 'ln -s secsh
Cheers,
--fred
> The only thing more frightening than your statement was the speed in which your statement was voted up.
/bots are probably moderating him just because of its low UID.
Yeah. At first read I thought he was trying to make a point, like, they'll see that this stupidy cost $50 to $100M, but it looks like he actually beleive what he wrote.
Fun to see.
Cheers,
--fred
Note that they even understand something about intellectual property and had great internal fights about it...
Cheers,
--fred
Bla bla, hate filter. content. content. bla-bla
Thanks for the link. I don't know why this is modded Offtopic, as it could also help for Earthlink extra HTTP header.
Cheers,
--fred