HN Explorer

Do you know why rms didn't start with a kernel?

6 points by access_denied ↗ HN
Does anyone here knows, why Richard Stallman choose the order for developing the several GNU programs he did? Why didn't he start with the kernel? Also, why the name GNU? Shouldn't be the kernel the only part of GNU named GNU? After all, all the other programs could be Unix.

23 comments

[ 4.2 ms ] story [ 16.4 ms ] thread
markup
UNIX is a proprietary operating system (not a kernel) and GNU aims to be quite the opposite, hence that acronym (GNU is Not UNIX). As for the decision to leave the kernel at last, well just give a look at how long is it taking for HURD to be completed... also I am guessing he needed other stuff to be created first (compiler, etc)
access_denied
What is the difference between an operating system and a kernel? (HaHa) Well, if you can point me out to some ressource regarding my questions about RMS' decisions, that would be cool. Regarding the GNU name, as far as I am informed RMS wanted to point out that the GNU is a UNIX-like OS as opposed to a non-UNIX like VMS. He decided to go with a unix because it had proven itself to be portable (unlike ITS for example).
markup
A kernel is part of the operating system, but the operating system is not composed by the kernel only. Take MacOSX: it is an operating system, its kernel is XNU. Linux itself is a kernel, not an operating system, etc.
gaius
Don't confuse unix (which doesn't really mean anything, it's just a general set of philosophies of how an operating system should be structured) with Unix(tm) which is a wholly owned trademark and a set of technical standards to which a unix must conform if it is to be a Unix(tm). Linux for example is not a Unix(tm) (OSX is).

The simple answer is that the hardware available at the time wasn't up to getting a useful level of performance out of the HURD microkernel architecture and so he gave up. HURD still isn't any closer to being a production kernel.

access_denied
Do you can point me out to some reference re "and so he gave up"? Thanks!
cchooper
To look at the question the other way around: is there any reason he should have started with the kernel?
access_denied
The name GNU suggests this.
cchooper
How?
access_denied
The name refers to the kernel/OS part of the project. Like if I would call my project DIND (DIND is not DOS). Not TINT (TINT is not TurboPascal). If somebody says to me: let's make a free | punk | whatever operating system, I would think of a kernel and maybe a shell or windowing environment. Not of a compiler, an editor or some 35 utilities. (I realize the UNIX distribution contained more than the OS).
cchooper
As other people have said, Unix was the name of an operating system, not a kernel, and the name has always referred to more than just a kernel and a shell.
lgriffith
Stallman objected to the fact that software could be sold so he made software that couldn't be sold. I am sure he understood its true value and priced it accordingly.

Create and deliver value. Demand to be paid for it according to its value to the customer. If the customer refuses to pay, find another customer.

Free is as successful as poverty.

chris_l
You can deliver value without being paid for it. For example by raising a child.
lgriffith
Who is the customer? You or the child or both?

Perhaps you are paid for the value you deliver. If not, why did you do it? However, it is a payment that cannot be transformed by exchanging it for other things you need or want. Its primary benefit resides within you and your child.

Money is not the only form of payment but it is an extremely important form. This is because of its use as a medium of exchange in a division of labor economy. The existence of both supports and sustains a far higher standard of living than could otherwise exist.

Stillman's philosophy is to reduce the production of intellectual values to the state of a barter economy. That kind of economy rises only slightly above subsistence. This is why I say "Free is as successful as poverty."

chris_l
Just because you don't "extract" any value, to use the MBA term, does not mean you did not deliver it.
access_denied
You have the historical facts wrong re Stallman's motivation.
lgriffith
I disagree. What he says is totally and absolutly irrelevant compared to what he actually does accomplish. The purpose of a system is what it does. Reduction to the barter system is what application of Stallman's philosophy accomplishes. THAT is his purpose.

According to Stallman, one should only be able to trade software artifacts for other software artifacts. This is even worse than a barter economy. Its as if an egg producer could only trade his eggs for other eggs or a potato farmer could only trade his potatoes for other potatoes. This in one simple step eliminates a division of labor economy with all of its attendant benefits and multiplication of productivity.

access_denied
But the result of GPLed software is also that it can be freely modified by anyone who cares. I think your stance is to one-dimensional, while remaining interesting. I also cannot prove you wrong. I will think about it. My guess at the moment is, that RMS changed his views over time.
lgriffith
Again, I don't care what RMS says, what he does is what is important. Simply read his re-re-revised EULA. His focus is to prohibit useful and valuable combinations of proprietary and OSS software by requiring the free distribution of the proprietary intellectual property content and abandonment of any applicable patents.

Which result has the most fundamental impact? The restriction to trading software artifact for software artifact no matter what or allowing modification by anyone who cares. For the former, there is no choice. For the latter, there is the choice to modify or not. I suggest the former has a far more profound effect on the final result.

Perhaps our disagreement is that I think the results are mostly detrimental and you think they are mostly good. The question is good for whom and for what reason?

I don't think that I am the one who is being one dimensional here.

markup
> The purpose of a system is what it does

Therefore the purpose of the internal combustion engine is to waste the 82% of energy? :-)

lgriffith
The result of an internal combustion engine is the transformation of heat energy into kinetic energy. Thermodynamics says this conversion cannot be 100%. Hence some of the heat is not converted. The "wasted" heat is not lost. It is simply not used.

To say the purpose of an internal combustion engine is not to use heat energy distorts the meaning of the words to the vanishing point by implying that the purpose of everything is not to use heat to the degree they don't do it.

The statement is "what it does". Its not "what it doesn't do".

hapless
- GCC was a necessary prerequisite to starting in on a kernel. What good would a free OS have been if it wasn't self-hosting on free software?

- Emacs was a politically important project because there had been a serious free-software kerfuffle among emacs-like editors. The most popular such editor for unix, Gosling emacs, had recently gone from open source to closed source without warning.