Tom` said:
There are a couple of arguments in favor of the "GNU/Linux" idea. The first is ideological. The Linux kernel, the argument goes, would not be where it is today without the GNU project and the GPL. It would have remained known only to hobbyists and OS developers (like Minix is today, I guess). It wasn't until the Linux kernel was licensed under the GPL and became the official kernel of the GNU project that it started to really gain in popularity.
IMO almost everything RMS talks about is ideological (and by extension, everything he claims the collective that is the FSF believes in and represents). Which is fine, but I can't stand how much manipulative language and propaganda he uses. Although he does have some sense of tact, he still comes off as completely degrading any opponent he has. In other words, he really has no concept of allowing others to pursue ideology differing from his own, and applies very far impacting notions of "freedom" and "ethics" to issues that don't matter nearly as much to most people as they do to him. He also has this strange notion that it's better to do nothing at all than to not do things his way. IE, it's better to do no software than proprietary software. It's not merely that he's ideological, but he's willing to sacrifice many real things for some intangible goal. And that, to me, is just selfishness.
I think it's pointless to argue what things relied on what. We could talk about how if Linus's grandfather had the flu the day he first met his future wife then maybe none of this would have happened. But then other circumstances may have caused the same - or better - events to transpire. This is just a game of credit, but the thing that FSF doesn't understand is that having people acknowledge their contribution does not mean that their purpose is being acknowledged, nor furthered. All it means is that people recognize that some guys have done good work, not that free software is "right" and proprietary software is "wrong."
It's really rather striking to me that a movement like free software, that relies on the contribution from so many people that they'd may as well be anonymous, would have a strong fixation on receiving proper credit. In fact, I'd go so far as to say that a good percentage of the people who have contributed to GNU software don't even really care about anything FSF stands for, and just wanted to help others or themselves by improving software that matters to them.
Tom` said:
The other major argument that I mostly agree with is the one that's already been addressed here: that the GNU system is a fundamental part of the Linux software environment. Exophase argues that this isn't necessarily the case, because it's not a significant component of the end-user experience. This is generally true, but not necessarily relevant - it may be the case that a lot of the software he mentions uses GNU libraries internally (he notes that glibc CAN be replaced with uClibc or whatever, but this is only common on embedded systems, something Stallman explicitly excludes from his argument).
I listed glibc as an exception to everything I've said, and as far as I'm concerned it's the sole one.
When you think about it, what kind of a role does a standard C library really play? Much of it is a thin wrapper to the kernel itself, which probably provides more user space services beyond necessary kernel functions than even libc does. The most commonly used portions are not that big. You can actually get very far with only a few functions.
And this is regarding C programs. Programs in other languages rely on libc even less, taking more from their own standard libraries. I don't see us giving this kind of status to a much bigger language support library like Boost.
Tom` said:
Furthermore, the basic "generic" Linux enviroment consists of the kernel plus the GNU system - sure, you can (and usually do) add X, some desktop environment, or whatever on top, but underneath (and most Linux users do make heavy use of the console, because it's more powerful and less difficult than most Windows users generally claim) you still have and use the GNU tools. It's a more fundamental part of the system than X or Gnome or whatever, whether or not it actually represents a larger part of the system in terms of code size.
This, to me, is more of a historical issue than anything. I disagree that most Linux users make heavy use of the console, or at least, in its role as anything remotely mainstream they don't (I've installed Linux for family members who would never touch it). I think it's really only more core in a historical sense, which shouldn't matter.
All I know is that a lot of the things that RMS harps on about in the "why GNU/Linux" FAQ, such as emacs, GNUChess (come on Stallman), even gcc.. really aren't what's core to Linux being Linux. The shell and "Core utils" a bit more so, but it's just hard to take these things as so worthy, especially when they were mostly settled years ago and don't see the kind of revisions and improvement the kernel does.
In fact, when you stop and think about it, quite a lot of both libc and even more so GNU Core Utilities are really just thin wrappers for kernel calls. So how can anyone make the claim that the the kernel is a small piece of the puzzle (like RMS really has) but that Core Utils are system defining?
