Or rather, why did rms et al dislike UNIX? It seems to come up a lot that they basically settled for a UNIX clone since it was widespread. They certainly didn't care for the philosophy.
Name:
Anonymous2013-02-13 11:39
Well... It WAS widespread... what were the alternatives? Make something brand new so people would be even extra cautious to try it out? And I don't think they were consciously against the philosophy, probably just thought it sounded good and later on swayed from the path in accordance to popular demand.
Name:
Anonymous2013-02-13 12:12
>>2 I don't think they were consciously against the philosophy
Richard Stallman told me he never cared for it in an email. He mentioned that the Emacs logo used be an overflowing sink because he basically cares more about having lots of features than sticking to the UNIX way. Which is fair enough.
Name:
Anonymous2013-02-13 12:14
The source-sharing tradition of the Unix world has always been friendly to code reuse (this is why the GNU project chose Unix as a base OS, in spite of serious reservations about the OS itself).
esr pls explain
Name:
Anonymous2013-02-13 12:34
That's why GNU is shite.
Name:
Anonymous2013-02-13 13:28
This actually explains a lot.
Name:
Anonymous2013-02-13 14:06
>>1
He played with the idea of a free OS based on the lisp-machines but deprecated the idea.
In order to run safer LISP code on them you had to had a microcode for the machine or the type-inference wouldn't work. Also the performance of IBM machines on LISP was getting better than in lisp-machines.
He admits that never played with unix before the GNU project, but thinks that overall the solution was correct
weird how old rms was more pragmatic than now
Name:
Anonymous2013-02-13 14:27
>>7
LLLLEEEEEEEEEEELLLLLLLLLLLLLLLLLLLLLLL
>30 YEARS AGO RMS WAS OLDER THAN NOW!!
>EGIN
I never used Unix (not even for a minute) until after I decided to develop a free replacement for it (the GNU system). I chose that design to follow because it was portable and seemed fairly clean. I was never a fan of Unix; I had some criticisms of it too. But it was ok overall as a model.
- http://stallman.org/stallman-computing.html
Prior to GNU, Stallman was working on a proprietary OS being developed by MIT, which I can imagine was probably superior to Unix in some ways.
I personally use Unix-likes because they're better than Windows, not because they're perfect.
Name:
Anonymous2013-02-13 15:15
>>9
Obviously nothing is perfect, but describe something better as a model.
Name:
Anonymous2013-02-13 17:00
>stallman hates UNIX
dear god, it all makes sense now
all the gnu bloat
all the emacs jokes
he just fucking hates the UNIX philosophy
Woah Woah Woah, rewind a little. Who built the OS? Linus Torvalds and friends. Who built shitty things into the OS? RMS's Friends/followers. Then you get a bunch college drop outs using these shits at great lengths, ignoring true *nix systems.
So you get a Grad, Linus, that reinvents a *nix system knowing full well it will be more practical than any other OS on the market. GNU only marketed lies, what Merricans only buy, so you get 150+ Linux distros.
IOW, OP, you got sold a lie, and you believed it. RMS is just a prick and delinquent (successful one at too). But again, we can say the same thing about Microsoft, and Apple Inc..
One major factor in the selection of Unix was its portability. At the time, RMS was already mourning the passing of the ITS system. All of the core software that composed ITS was written in PDP-10 assembly (as were most operating systems of its type), so when the PDP-10 died ITS died with it. The same thing would happen to MIT's Lisp machine culture when Symbolics ceased waking life.
At the time, Unix was the only OS prevalent in academia that had been widely ported to multiple computer systems. RMS concluded that if he could use Unix as a base for his system, the wider Unix community would assure its continued survival. This is exactly what happened when the Hurd effort failed - the GNU project simply adopted another Unix-like kernel in its place.
Incidentally, this approach explains why many of the GNU tools have this weird split personality. Things like info or mandatory long options are near useless to people who are accustomed to Unix shells - their real purpose is to be used as primitives in an environment that bears more semblance to a Lisp machine (namely, emacs).
Name:
Anonymous2013-02-13 21:46
>>22
All of this GNU shit was inherited from the Lisp machine. That explains everything.
Name:
Anonymous2013-02-13 21:55
>>23
It's not so much direct inheritance as influence - Lisp machines were programmed in Lisp from the ground up, with only a small core in assembly. The command line interface exposed by the GNU utilities is very similar in presentation, but the implementation (C, with a C kernel) is totally different.
This is significant because a pure Lisp system would otherwise have had terrible performance on early micros like the 386. This allowed RMS and co. to keep running emacs with decent performance by ``shelling out'' to C tools with a similar looking CLI.
Name:
Anonymous2013-02-13 22:12
>>24 with only a small core in assembly.
The assembly itself was Lisp and supported macros.
Name:
Anonymous2013-02-13 22:21
>>25
I wouldn't say it was Lisp. The instruction set provided operations that were expressly designed to implement Lisp, but at a lower level. It wasn't really a Lisp unto itself.
If any of you actually believed in the so-called ``Unix philosophy'' you wouldn't be posting to a web-based BBS.
Name:
Anonymous2013-02-14 2:20
>>11
RMS doesn't hate Unix, he doesn't care about it. Computers with practical functions are more useful than computers with limited functions.
>>14
The kernel is not the OS. The kernel has always existed within the context of a full OS.
Name:
Anonymous2013-02-14 2:24
>>27
Why is telling the truth mean talking bad about Lisp machines? Can you provide good evidence to show that a purely Lisp system would run acceptably on a microprocessor.
>>29
The web exists as a way to share information over the HTTP.
>>28
It's difficult to make absolute statements about the merits of the architecture, given its age. Deeply micro coded architectures can be matched or beaten by a less specialized ISA and a good compiler. The latter approach has the advantage of not being tied to a particular application profile - compilers can evolve; microcode generally does not.
>>31
There is nothing wrong with the truth, and you did well in explaining on how the GNU community grew. I am just talking about the schizophrenic GNU policy, that taints today's industry, esp. the Linux operating system with the GNU utilities. And about the lisp machine, I have provided on the links above.
>>30
We know. But by the time GNU got to adopt Linus' _Kernel_, the _kernel_ had most of the components to be its own operating system. The same thing can be said about Mimix.
Whatever, at least don't teach lies to the children. They deserve better than the cold war. I hate what is happening now. Sad Neo-Cons.
Name:
Anonymous2013-02-14 16:36
>>30
In Unix terminology, the kernel is the operating system. Whether it's sh or chess, these are application programs bundled with the OS. There used to be some name for an operating system with additional user programs, but I forget what it was. UNIX resellers used to use it when they shipped UNIX plus a few of their own programs.
This is all independent of GNU and Microsoft and other groups who have chosen to use the term ``operating system'' differently. Read The Design of the UNIX Operating System by Maurice J. Bach which explains the structure of UNIX very well.
Name:
Anonymous2013-02-14 16:52
>>36 Real men don't attack straw men
I love this thread. The OpenBSD team raised some interesting points, like why is providing HTTP links to proprietary software in a textfile unethical because it promotes the use of proprietary software, but thousands of lines of code in Emacs and dedicated binaries to run it on Windows are somehow kosher?
People who think the Linux kernel constitutes the most important part of the operating system should be forced to use Android for real work (and compare it to Debian GNU/kFreeBSD). Jesus fucking Christ.
>>37
You're a fucking idiot. In ``Unix terminology'', the concept of the base system as it exists in, for example, the BSDs (but not Linux) is the operating system. The kernel is the fucking kernel.
Name:
Anonymous2013-02-14 19:04
>>40
Nope. To quote the book I mentioned, The operating system interacts directly with the hardware, providing common services to programs and insulating them from hardware idiosyncrasies. Viewing the system as a set of layers, the operating system is commonly called the system kernel, or just the kernel, emphasizing its isolation from user programs.
This is followed by a nice diagram showing this model.
I'm pretty sure Tanenbaum defined an operating system as the program that manages resources or something like that, but I can't find that definition anywhere (can anyone help me out here?).
Instead, here's a quote from Operating System Essentials (after mentioning the ``operating system = kernel + application programs'' definition): A more common definition, and the one that we usually follow, is that the operating system is the one program running at all times on the computer—usually called the kernel. (Along with the kernel, there are two other types of programs: systems programs, which are associated with the operating system but are not part of the kernel, and application programs, which include all programs not associated with the operation of the system.)
Of course, I'm happy to accept that words change over time. I'm not a prescriptivist, and I'm okay with ``operating system'' meaning what Richard Stallman thinks it means in the same way I'm okay with ``hacker'' meaning computer criminal. Semantic change, and all that.
>>41
In addition to being a fucking idiot, you're also a condescending toolbag and wrong. Kernel-as-operating-system is literally only used by the dimmer parts of the Linux community. It's not a meaningful definition, it's not a useful definition, and it's not a correct one.
Name:
Anonymous2013-02-14 21:39
FACT:
C is faster than Lisp.
Name:
Anonymous2013-02-14 21:51
>>43
Lisp can be just as fast on a system with tagged memory.
Also, just to muddy the waters of this awful thread even further:
The Linux kernel in its most commonly deployed configuration cannot even bootstrap itself fully without the assistance of user programs (init and udev). On any Unix like kernel, death of the init process will induce a kernel panic because a root process must always exist.
If you define "operating system" to mean only that software which is always running and required for all other programs to function, does this definition not necessarily include "user programs" like init? Do combinations of the same kernel with an alternative init daemon then constitute different operating systems?
But what we are arguing is that at the point GNU adopted what many call the the linux kernel, it as a full fledge operating system like Minix, capable of writing asm and C into disk, making TCP connections, et cetera.
>>43,45,47
Depends on the application and architecture. The best way to find out is to test them both, in all optimizations you can apply. On my enterprise level apps, that require LOTS of abstracts between multiple server, Scheme executes much faster than C pointers.
>>49,50
thanks. At the end of the day, GNU adopted anything it could to sell RMS' propaganda, to the point of using hearsay trust to make _recommendations_, than investigating like scholars and _defenders of freedom_. Study the history, it's on their own damn website and logs.
I need some sleep.
Name:
Anonymous2013-02-15 13:31
>>42
It has nothing to do with Linux. Linux didn't exist when The Design of the UNIX Operating System was published, and GNU was nowhere near release yet.
You have things a little backwards--it wasn't until recently that the distinction mattered. Older operating systems were definitely one complete ``kernel'', which was the operating system. Here's First Edition Unix, for example: http://minnie.tuhs.org/cgi-bin/utree.pl?file=V1
Again, I'm happy to accept that things have changed. If you want Linux to be the ``kernel'' and some other set of programs combined with Linux to be the whole ``operating system'', then that's fine. Words change meaning, especially in computer technology. All I ask is that people play fair--if you have the right to name your system ``GNU/Linux'' and consider it an ``operating system'' (and you certainly have that right), then I should equally have the right to use a more traditional definition and call ``Linux'' my ``operating system'' (I don't even use GNU, besides gcc and its dependencies).
Name:
Joke of the Day2013-02-15 22:36
Why did the gnu choose Unix?
Because gnus are smarter than most humans.
AHAHAHAHAHAHAHAHAHAHAHAHAHA