Ballmer: Linux users owe Microsoft

[Vogateer]

Some people say that Richard Stallman needs to be "pragmatic" etc. etc.

I don't know, the HURD might have benefited from a more pragmatic approach.

Thinking about Microsoft again, I keep thinking this is more likely to be FUD than anything. Microsoft is out to make money, and dragging an "end user" to court would yield no money, since end users aren't going to have enough money to pay out if they get sued, while taking on the big boys would be too expensive, and probably get them nowhere. There is the thought that they might sue someone just big enough so that everyone knows their name, but not big enough to put up a costly legal fight, but that will be difficult without suing one of their own customers. I bet you it's much more profitable to simply posture as though you might sue, and try to scare people away than actually carrying through with a lawsuit.

[Tsuroerusu]

Some people say that Richard Stallman needs to be "pragmatic" etc. etc.

I don't know, the HURD might have benefited from a more pragmatic approach.

What do mean? The reason it never really got anywhere is because the micro kernel design they chose, apparently, is a pain to develop and maintain (Horrible bugs can come out of nowhere), aka a big design mistake, kind of like the steaming pile of goatshit known as Novell ZENworks.

(Yes, been watching The Angry Nintendo Nerd)

[Vogateer]

That was just meant as a joke, but if you're going to take anything semi-serious from it, you could look at it as example of following a philosophy too far. That micro kernel design was chosen on the basis of philosophical beliefs that it was a better design, though unfortunately it was entirely impractical, and now they've become a joke. The only time I ever hear HURD mentioned is to get a laugh from how pointless it has become.

[Tsuroerusu]

That was just meant as a joke, but if you're going to take anything semi-serious from it, you could look at it as example of following a philosophy too far. That micro kernel design was chosen on the basis of philosophical beliefs that it was a better design, though unfortunately it was entirely impractical, and now they've become a joke. The only time I ever hear HURD mentioned is to get a laugh from how pointless it has become.

Go back and listen to TLLTS episode 79, therein Stallman explains why he chose the micro kernel design, and openly admits that it might have been a mistake, and I think on the recent interview with him he did admit that it's not that relevant of a project.

So the micro kernel design was chosen because, at the time, it was supposedly the simpler and better design, later the opposite has proven true. That's not about taking a philosophy too far or needing to be pragmatic, that's just a matter of "things proving otherwise".

Some people (Microsoft among others) thought that tablet PCs would revolutionize computing, and look, almost no one uses them.
Sure might be good for browsing the web while taking a dump, but when you're not hitting the can it's so easy to break while carrying it around.

[Vogateer]

I've read about why they chose the micro kernel design, and I know it was supposed to be, in theory, the better design. I think the big debate between Torvalds and Tanenbaum gives a good snapshot for the ideas of the time. Whenever we disagree, I wish you wouldn't immediately refer to some reading or episode in an assumption that I haven't done any reading or study on a subject and that my views must be held in ignorance. Once again, I'm sure you don't mean to present this sort of attitude, but it has an air of arrogance about it.

In this case, reading what I already have about the HURD, I would say that they did take the philosophy too far. Once they realized that their philosophy--micro kernels are a better design--didn't conform to reality--they're not, they should have revisited their philosophy on kernel design and changed it. Instead they kept plugging away at it. If they hadn't held so strongly to this mistaken philosophical belief, they might have produced something useful.

I'm not saying any other philosophy held by FSF people are wrong, just that in this specific case the adherence to this specific philosophy produced poor results.

[Tsuroerusu]

I've read about why they chose the micro kernel design, and I know it was supposed to be, in theory, the better design. I think the big debate between Torvalds and Tanenbaum gives a good snapshot for the ideas of the time.

Indeed, I have read Linus' book, so I know what you're talking about.

Whenever we disagree, I wish you wouldn't immediately refer to some reading or episode in an assumption that I haven't done any reading or study on a subject and that my views must be held in ignorance.

I didn't assume that you hadn't researched this matter, but I just referred you to Mr. Stallman himself saying that he made a mistake. The thing that made me refer to that was the use of the word "philosophy". Personally, I think using it in this context presents the wrong picture. I am of course not a 100% native English speaker (Although I think I am doing pretty good), so I may misinterpret a word or two sometimes, which is very unfortunate in some situations, but hey, it's reality.

Once again, I'm sure you don't mean to present this sort of attitude, but it has an air of arrogance about it.

This is what Zack Rusin described about internet communications sometimes being limited, because if I could use a facial expression when I said what I said, I'm quite sure you would have interpreted it differently.

The reason I may seem a bit, if not more than that, serious when discussing Richard Stallman, or the FSF, is because sometimes people will just rag on them because of something they recently heard, or base conclusions on incorrect information. I have read Linus and the kernel developers' arguments against the GPLv3, and I largely disagree with what they're saying. First of all, to me, it seems like they base the opinions on "today's world", version of the GPLv2 was written over 10 years ago, and the world of computing was radically different back then. If GPLv3 lasts that long, or longer, you really do want to be forward looking. When Linus say that blocking "TiVoization" (Which is not a very optimal term, I'll agree to that) is not something a software license should do, I think he's failing to realize that it's not just something like TiVo the FSF wants to prevent, it's selling computers with free software, on which the users do not have the four freedoms. Picture this, I go into Circuit City sometime in the future, and buy a PC with GNU/Linux preloaded on it. Since the guys that make them wanna be able to play BluRay or HD-DVD it has HDCP implemented in the kernel, and maybe alongside some other DRM schemes. With the computer, I do get a CD with the source code for everything, so in theory they are GPL compliant. HOWEVER, the machine will only run a signed kernel, so I can't go in and change the kernel, and take out the copy protection or alter the way that it works. I have the source code, but I can't do shit with it, because the machine is DRM nightmare, and more than likely, do not have vulnerabilities to exploit like the original Xbox has. In practice, my freedoms are not being respected, and sure you could argue that I could just not buy that computer, but remember the GPL, and idea behind the FSF's definition of free software (The BSD guys have a somewhat different take on "free software"), is to ensure freedom for ALL the users of the software and that you should be able to receive the software in whatever way and have the four freedoms respected, both in theory and in practice. This next suggestion might be far out there in the future, but imagine that the RIAA/MPAA manages to make it illegal to make, sell or distribute personal computer equipment that does not have DRM bullt-in, then it's suddenly illegal to buy a computer that does not have TPM/DRM in it, and I can't build one myself either, because the motherboards for a free (as in liberty) machine would also be illegal. Sure this may seem very unlikely to Linus and some other folks, but with folks like the RIAA, MPAA, Microsoft, Apple and other bastards out there, I think this fear is well founded.

Some people also argue that the GPLv3 is not needed at all because the GPLv2 is good enough, and wow all of a sudden come this "Microvell-deal" and suddenly the GPLv3 sounds like a damn good idea even to some of the critics.

It seems like FSF, with the help of Eben Moglen, will add a clause that will make Microsoft's promise of of not suing Novell's customers be a promise not to sue any user of the software. I'm sure Linus would argue that that is not a thing for a software license to do either, but I call bullsh!t on such a claim, because Novell could gain such a FUD-based monopoly that they could really put a hurt on Red Hat or Mandriva, and then the even playing field of free and open source software suddenly turned to shreds. To me, the GPL is not just about saying what people can and can't do with the software, it's also about ensuring certain freedoms for everyone. If you couldn't distribute a piece of software because some dude claimed to have a patent on stuff, which he may not have in reality, it would be more than enough for the majority of people to not distribute that piece of software because the fear of a patent lawsuit is very big, and nobody has the reosurces (aka cash!) to look into patent claims or defend against a patent lawsuit.

Why do you think Windows can do as much as it can do today (Yes, it sure can crash, I'm sure the patent of that one is owned by Mr. Gates personally) ? When Microsoft built Windows and started to build their future monopoly, they didn't have the issues of software patents that we have today, or at least the issue wasn't as big as it is now. GNU/Linux and other OSes have to both struggle with technical issues of features and compatibility while also dealing with the patents, Microsoft only had to deal with the technical issues that they faced/faces (Yes, I'm sure Microsoft also has a patent on not running as root, Linspire better be careful it ).

In this case, reading what I already have about the HURD, I would say that they did take the philosophy too far.

Once they realized that their philosophy--micro kernels are a better design--didn't conform to reality--they're not, they should have revisited their philosophy on kernel design and changed it. Instead they kept plugging away at it. If they hadn't held so strongly to this mistaken philosophical belief, they might have produced something useful.

That would essentially mean a rewrite of the kernel, and in that scenario no one was gonna bother anyway because right there was the Linux kernel!

Remember, RMS wanted a totally free system as fast as possible, and Linux got stable quite quickly, so people who might initially have been motivated to work on the Hurd, just used Linux because it was already out there.

I seriously doubt that even if the GNU folks had started over and used the monolithic kernel design, once they ran into all the difficulties of the micro kernel design (And Apple has been so kind to show it's performance weaknesses), that the Hurd would be just as usable as Linux is today.

Also, an idea that I think is interesting, maybe not so feasible in practice, but I just think it's interesting, is that if somebody really do want a GPLv3'ed kernel, they could take the kernel of FreeBSD or OpenBSD (FreeBSD would be the best choice, at least in my opinion, as it's hardware support is superior to OpenBSD's) and just GPL it, and bam, there you go!
I'm not saying anybody should do this, I just think it's interesting.

I'm not saying any other philosophy held by FSF people are wrong, just that in this specific case the adherence to this specific philosophy produced poor results.

The Hurd certainly is not usable, I will agree fully with that. Just a few years ago it had some file system bugs that would destroy files because of timing errors.