Ross wrote:

What's your opinion of "GNU/Linux"?

So, OK, you deserve a serious answer -- even though I was semi-serious, in that you sometimes just enjoy ranting and trying to get a rise out of people, and heavens knows it's a harmless hobby and sometimes almost an art form. (Nor am I in the least offended by people merely holding strong views, especially if they express themselves in an interesting and engaging manner. Example: Google for the phrase "considered harmful". Lots of good reading, there.)

What you posted encompasses several related questions. Here's the entire omnibus discussion, FAQed and suitable for framing:


Q: Are you bugged by people saying "GNU/Linux"?

A: No.


Q: Are you bugged by people not saying "GNU/Linux"?

A: No.


Q: Are you bugged by people insisting that others {say|not say} GNU/Linux?

A: A little, but not enough to make an issue of.


Q: Do you feel that Stallman's claim that people should say "GNU/Linux" on grounds of granting credit to the GNU Project has merit?

A: {sigh} Yes, somewhat: The effort to build a complex OS from scratch using non-encumbered code, of necessity, has to start with the less-flashy but utterly crucial developer tools and libraries first, which unfortunately becomes something of an iceberg proposition, where fame and public credit is concerned (nine tenths underwater and invisible). E.g., the GNU Project painstakingly created and debugged the following utterly crucial codebases: glibc, gcc/g++, gdb, cpp, bison, autoconf, make, automake, emacs, cvs, and the binutils. The crucial importance of those can be seen from how many of them have been instrumental in the development of {Free|Net|Open}BSD and even (in many cases) the proprietary 'nixes, as well.


Q: But aren't the lines of code from the GNU Project utterly dwarfed by the lines of code from other sources?

A: The question makes the foolish assumption that all lines of code are of equal importance, and therefore cannot be usefully answered as posed.


Q: But if the GNU Project deserves credit, don't others equally, such that you're really asking us to say "GNU/BSD/Perl/Apache/MIT/Python/Linux"?

A: First, I'm not asking you to say anything. The question you asked, and are in effect now re-asking with elaborations, was whether Stallman's request strikes me as having merit. Second, you're ignoring the point about which code has been most crucial during development (which includes maintenence and enhancement, not just initial creation) of everything you've alluded to. Besides, some of those items aren't even essential to running a 'nix system in the first place, such as the X Window System and the Apache Foundation pieces.


Q: But how can you say that their claim of credit has merit when they're not even competent to create a usable kernel within the current geological era?

A: Because that question rests on a non-sequitur assumption, in part. Yes, they screwed up in being led down the garden path of microkernels that was in vogue among OS theorists of the day. (Their design may yet be vindicated, but I wouldn't hold my breath.) But the state of the HURD kernel is not relevant to the point I've made.

Additionally, the GNU Project people acknowledge that HURD still sucks, which is why they run Linux instead, and is part of why their request is that people say "GNU/Linux" instead of just GNU.


Q: But isn't it possible to edit and debug code without emacs, maintain code repositories without cvs, use other implementations of make, use other compilers besides gcc, use other lexical analysers other than bison, use other C libraries other than glibc, and so on?

A: Yes, but that's ignoring the point. The point is that, during the development history of practically all the software comprising a modern Linux distribution, those key GNU development tools have functioned as keystone pieces that made everything else possible. The reasons why those were used, and whether alternatives have been available at various points, are entirely beside the point. The point is that those pieces have been and remain of crucial importances as developer tools in crafting just about everything else. That's a fact. I'm simply pointing out the fact.


Q: But given that you can use a Linux system all day long, and never use any of those pieces other than glibc, how can you say they're that crucial?

A: Because they are. The fact that you don't happen to use them directly and personally is irrelevant to the point. The important fact is that those have been and are key tools relied upon by those who produced the software that you do use on a Linux distribution, by and large.


Q: But isn't it more logical to call the system "Linux", since Linux is what starts the system and runs everything?

A: No, one might argue with no more or less sense that the system should be called "Phoenix" or "Award" because the BIOS ROMs start it. Besides, it's /sbin/init that fork-and-execs to launch processes, not the kernel.


Q: But isn't the kernel more fundamental to the system?

A: No. You're not a developer, are you? What Stallman would have us call the GNU system can run just fine on kernels descended from 4.4BSD Lite, the Minix kernel, or any number of others, and those kernels perform services that are essential but not as key a resource to (overwhelmingly userspace) development efforts as the other pieces. (Besides, again, Stallman wants people to credit the kernel effort.)


Q: So, do you say "GNU/Linux", yourself?

A: Most of the time, no.


Q: But isn't that inconsistent with your answer?

A: It's an imperfect world: I'm lazy, "GNU/Linux" is awkward both to say and to write, and one gets tired of being buttonholed by people wanting to argue about it.


Rick Moen
rick@linuxmafia.com

Brandioch wrote:

...Others deserve as much credit as GNU does....

Transparent question-begging that completely ducks the issue, noted without further comment.

Fundamentally, I rather suspect you just didn't read what I wrote very carefully. Oh well.

Rick Moen
rick@linuxmafia.com

Ross wrote:

See, this wasn't really a FAQ, but a AASOPQ - artfully arranged set of pseudo-questions - like Galileo's "dialogue".

Well, it was amusing, but somebody (presumably that other guy, "Brandioch") seems to have established the bizarre notion that I had posted a bunch of questions as a forum-wide survey. Obviously, my conceit, rather, was that some hypothetical person had somehow conned me into having an interview-style dialogue on the subject, whereby the follow-up questions were prompted by the contents of my antecedent answers.

If you substitute radically different answers, then most of the questions no longer make the least bit of sense.

But if you just need a backdrop to rant against, without actually addressing the subject, I suppose they'll do as well as anything else.

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

Should I start calling my comic strip "CORELDRAW/Help Desk" simply because I rely on CorelDraw to create it every day?

Does your comic strip include a copy of CorelDraw inside every copy of the newspaper, and not have a prayer of functioning without that embedded copy? Your parallel would seem to require that, for starters.

Looks to me like that parallel's pretty non-Euclidian, to say the least.

Let's play this game some more:

If I created a *proprietary* program using emacs and named it "GNU/ProprietaryApp"....

When you say "created using", do you mean "linked to and incorporating for key portions of its functionality"? You could certainly create such a thing and nobody could reasonably object. (Distributing it would be copyright violation, on account of licence conflict.) But that's not what you mean, is it?

See the category error?

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

You can argue that every one of my comic strips "contains" the PNG format.

You could. But that would be rather silly. We're talkin' 'bout inclusion of big honking codebases -- code -- not the platonic ideal of a graphics format.

There's a difference between your incorporating the idea of a sandwich into your baloney on rye and breaking into the Kilpatrick's bakery to grab some loaves so you can include them. Either way, though, I say what you have there is most definitely baloney. ;->

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

I'm not breaking into the bakery. Kilpatric has placed all the sandwich parts on a table, and has expressly given me permission to use them.

In which case, slathering them with mayo and giving them to the public as "cwbrenn's 100% pure slab o'mayo" wouldn't be illegal. But it would be a rather tacky way to behave towards Kilpatrick Bakeries. (Not to mention greasy.) Which of course was the point.

Sorry, I just don't think you can creatively analogise your way out of the ethical issue. Though you're hardly the first person I've seen try that.

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

I don't see it as an ethical issue. Sorry.

Well, as I said up front, I'm not trying to convince anyone to use or not use particular names for things. I have no stake in doing so, and no special reason to care. I was just trying to answer Ross's question.

My main point to you has been that pretty much 100% of your arguments have suffered gross defects of logic. And, after demolition of those chains of logic, you're now reduced to...

But it's a political act and has nothing to do with ethical obligations.

...question-begging. {shrug} Oh well.

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

I've yet to see you actually PROVE any of that.

Delighted to say, I can live with your having not gotten the point. So, we're both happy.

Rick Moen
rick@linuxmafia.com

cwbrenn wrote:

I have no idea what "non-Euclidean" means, other than "not Euclidean". :)

I was making a witticism about situations where parallel fails to be parallel.

When mathematicians were attempting to see if any of Euclid's postulates were absolutely necessary to an axiomatic system, they found to their surprise that negating one of them (the parallel postulate) created self-consistent, functional alternative systems of geometry, which were thus dubbed non-Euclidean geometries.

If none of that rings a bell, well, read some history of mathematics and science.

Rick Moen
rick@linuxmafia.com