Not having the "linux" confusion helps.
Not having the "linux" confusion helps.
Posted Apr 14, 2003 21:23 UTC (Mon) by Peter (guest, #1127)In reply to: Not having the "linux" confusion helps. by ber
Parent article: One-on-one with Richard Stallman (SearchEnterpriseLinux)
Check the fraction of tools and the availability of alternative Free Software components apard from the kernel. Mostly the kernel as main component was missing in the early 90s. It is depressing that most people dont't know this.
Hey, it's all a game of semantics, just like the "GNU/" prefix. Specifically:
- What is considered an integral part of the operating system (which, it is claimed, was mostly finished by the time Linux showed up)?
- Of this, how much can the GNU project legitimately take credit for?
- And of this, how much was in place and mature by, say, 1993, when Linux as a whole was starting to be considered production-ready?
As to the first, obviously one must have a kernel, and the GNU coreutils (formerly known as shellutils, fileutils, and textutils) surely belong as well. But what about Emacs - yeah, it's a lot of lines of code, but is it really part of the operating system? (Some, of course, could argue that Emacs is its own VM OS, and the rest of the system is merely a hypervisor.) What about GNOME? I wouldn't consider GNOME part of the OS, although I know there are those who disagree.
As to the second, consider the X Window System. It was developed quite independenly of the GNU Project, even though the Project claims X as part of the GNU OS. Likewise for TeX. I don't know which implementation of vi is officially "part of GNU" but I don't think the GNU Project developed any of them. And I believe the only MTA developed by GNU is smail, which nobody uses anymore.
As to the third point, quite a lot of software has been developed in the last 10 years or so since Linux hit the big time. GNOME, for instance, was not even born in 1993. Ditto many of the "desktop" applications most people (though not me!) would certainly think of as part of the OS. Apache, that other flagship brand name for free software, was still struggling to emerge from its NCSA cocoon.
This is why I call into question the assertion that the GNU OS was "almost done" by the time the Linux kernel came along. And the attendent assertion that because of this, the GNU Project deserves prominent billing in the very name of the product. (Oh, and it's also fun to troll the FSF acolytes from time to time.)
Posted Apr 15, 2003 15:53 UTC (Tue)
by coriordan (guest, #7544)
[Link] (2 responses)
> What is considered an integral part of the operating system one also needs[1] a C library, and a dynamic linker. Every distro [1] without these, you could call LiLo or GNU Grub an OS, GNU/Linux OSes used to come with compilers but this ended with the Windows era. > how much can the GNU project legitimately take credit for? (I think this is the more important point) These companies (like Micrsoftware/IBM/etc.) usually talk about "Linux" crap, I'm late for something, as usual.
Posted Apr 16, 2003 21:49 UTC (Wed)
by Peter (guest, #1127)
[Link]
The debate ended by the new extension to Godwin's Law? But of course. (: Once again, anyone can play semantic games about what an OS comprises. Many people, myself not included, would consider a full GUI environment (including at least one web browser, various media players, a "friendly" mail/news app, and GUI control panels for various system settings) to be part of the OS. After all, this stuff does come with Windows XP. I tend to the minimalist view that the OS consists of a kernel, boot loader, program loader, shell and associated utilities. And indeed, many of these components are supplied by the GNU Project. Then of course there is the semantic point of what constitutes "done". (I was once told that the word means "well-browned on both sides", but that probably doesn't help.) RMS says the GNU Project was "almost done" with the GNU OS by - say - 1993. But if you take the 1993 versions of all the GNU OS software, how many people would consider the result a "finished" OS? I might; my needs in a Unix-like OS are pretty modest. But most people probably would not. Definitely a matter of perspective. If you (like rms) view proprietary software as inherently evil, then it is a Bad Thing that such software is able to leverage the sweat of the Linux and glibc developers. Myself, I just can't see it that way. I see proprietary software not as a moral ill, but as a market inefficiency that will, with luck, eventually go away. Free / Open Source software is a better methodology that will in the long term produce better quality products. In the short term, though, you can't just ask automotive engineers to stop using HyperMesh and LS-DYNA for their crash simulations: there is simply no free software equivalent, and that sort of thing can't be done without the proper software (except by building physical prototypes and crashing them into walls, which is a great deal more expensive). Anyone can ignore the matter of liberty. I understand rms's concern about users not understanding the freedoms they have been given, and being willing to settle for less. But honestly, I think it's more important for computer professionals to understand these things, and that job seems basically done. Sure, there are lots of IT people with a lot of training and emotional investment in proprietary software, but I don't think they prefer it simply out of ignorance of their freedoms. Or, at least, those who are ignorant of these things nowadays are willfully ignorant and there's not much you can do about that.
Posted Apr 17, 2003 3:41 UTC (Thu)
by Ross (guest, #4065)
[Link]
Hi Peter (same Peter from the *other* debate?)Not having the "linux" confusion helps.
> [...]one must have a kernel, and the GNU coreutils
uses GNU Libc and GNU binutils for these. One probably needs
a shell. Many people use GNU Bash, some don't but every distro
uses GNU Bash for it's scripts (because it's a bourne compatible
POSIX shell). Every individual doesn't need a C compiler but the
OS couldn't exist without a C compiler. The GNU project wasn't
about writing as much software as possible, it's about writing the
necessary software to have an OS. As Stallman says: "it's not very
exciting writing a new implementation of Unix Tar, but we did it,
LD isn't exciting either, but we wrote one of those too..."
on the other hand is a "POSIX compliant" OS.
(I suspect they'd rather you bought software than wrote software.)
It's not actually about taking credit, if the system was complete and all
that was left to worry about was the credit then it would be better to
sit back and simply bask in the existence of the OS, but this is not the
case. The OS is under constant threats from proprietary software houses
trying to entice people to run their software on GNU/Linux. And MegaCorps
pushing for stricter copyright and patent laws.
and "OpenSource". This allows them to ignore the issue of liberty since
the celebs from those camps remain quite (probably not wanting to be seen
as zealots).
Ciaran O'Riordan
Not having the "linux" confusion helps.
Hi Peter (same Peter from the *other* debate?)
The GNU project wasn't about writing as much software as possible, it's about writing the necessary software to have an OS.
if the system was complete and all that was left to worry about was the credit then it would be better to sit back and simply bask in the existence of the OS, but this is not the case. The OS is under constant threats from proprietary software houses trying to entice people to run their software on GNU/Linux.
These companies (like Micrsoftware/IBM/etc.) usually talk about "Linux" and "OpenSource". This allows them to ignore the issue of liberty since the celebs from those camps remain quite (probably not wanting to be seen as zealots).
Actually bash is a really heavy-weight program if you just want something to interpret Bourne-shell scripts. Some distributions use pdksh or ash.
Shells