Stallman on free software, Torvalds and Microsoft (LinuxWorld) [LWN.net]

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 17:56 UTC (Wed) by allesfresser (subscriber, #216) [Link]

Did you miss the two or three other instances where he says "we" published the GPL v3? I took his "I" comment as being the person who *initiated* the process of developing v3 because he was the principal author of v2, and this is entirely proper.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 18:34 UTC (Wed) by rsidd (subscriber, #2582) [Link]

That's like saying I'll lose the freedom to fly by flapping my arms.

It's a bit more than that: it's saying you'll lose the freedom to get anyone other than the plane manufacturer to help you fly your plane.

It's taken for granted (or used to be...) that you can choose any mechanic to tinker with your car. My car mechanic is not an "authorised" centre for my brand, but he does a great job. I don't see why people not only accept sacrificing that sort of freedom in computer software, but actually justify why it's necessary.

It's indeed the "same old same old", but it still needs saying.

The only thing I find jarring is RMS's repeated insistence on GNU/Linux. Let's put it in perspective. If we take his claim of "a GNU system without a kernel" at face value, it took the GNU project 6 years to do that (1984-1990). 17 years later, the Hurd is still not ready, while the Linux kernel is competitive with any proprietary system. Moreover, after Linux emerged, much GNU software development has explicitly been driven by the requirements of Linux systems. So, merely based on timescales, it seems quite fair that "Linux" should get the credit.

Plus, it's a stretch to say that "the GNU system" came entirely from the GNU project. The X window system came from MIT. TeX came from Don Knuth. Various utilities came from BSD. The major contributions of the GNU project are the compilers/toolchain and Emacs (and the latter is really optional). Everything else was either also there in BSD, or could have been rewritten in short order.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 19:24 UTC (Wed) by pfavr (subscriber, #38205) [Link]

IIRC Linux was developed using GCC (GNU Compiler Collection).

Buy a GNU t-shirt the next time you get the chance?

;-)

BR Peter

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 21:40 UTC (Wed) by nix (subscriber, #2304) [Link]

Actually it was developed with gcc (the GNU C Compiler).

It didn't get renamed to the GNU Compiler Collection until many years
later :)

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 18, 2007 1:33 UTC (Tue) by landley (guest, #6789) [Link]

If Netscape was originally developed with the Microsoft C compiler, would
Microsoft own it?

http://fabrice.bellard.free.fr/tcc/tccboot.html

(And this is glossing over the gcc/egcs split. Neither Cygnus nor Code
Sourcery are subsidiaries of the FSF.)

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 18, 2007 19:52 UTC (Tue) by nix (subscriber, #2304) [Link]

Under some of the MSVC++ licenses, yes. (Some of the licenses MSVC++ have
been released under are simply ridiculous, and I can't believe their terms
would fly anywhere in the western world.)

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 18, 2007 23:36 UTC (Tue) by sflintham (subscriber, #47422) [Link]

I'll bite, give us some gross-but-cool examples please!

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 19, 2007 20:04 UTC (Wed) by nix (subscriber, #2304) [Link]

They endeavour to do things like restrict the licenses under which code
generated with those compilers can be released, restrict the operating
systems to which code ever touched by those compilers can be ported, and
that sort of thing.

All quite unenforceable of course.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 19, 2007 22:36 UTC (Wed) by landley (guest, #6789) [Link]

It's worse than that. Everybody forgets that the _original_ cause for
the big Microsoft antitrust trial was Microsoft' license on the client
version of NT, which prevented Netscape from running its web server on
NT. (The server was where Netscape made its money.)

ftp://ftp.ora.com/pub/examples/windows/win95.update/ntwk4...
http://www.infoworld.com/pageone/opinions/petrel/petreltw...
http://www.oreilly.com/news/differences_nt.html

That's why Netscape filed suit in the first place. The suit expanded
into a full-blown Sherman antitrust act along the way, but initially it
was just about Microsoft's licensing terms on its OS preventing people
from running any non-MS servers on NT.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 18, 2007 23:40 UTC (Tue) by landley (guest, #6789) [Link]

Are you saying the FSF is being as insane as Microsoft? If so I can see
what you're getting at, but I personally think that only Stallman's that
out of touch with reality, and the rest of the organization around him
mostly isn't.

Keep in mind that Microsoft was sending shrinkwrap licenses out with its
software back in the DOS days, long before the DMCA. They were totally
unenforceable in court but would _you_ want to go to court against
somebody who could spend $50 million on laywers to make an example of you
and never _notice_ the expense?

There's a certain type of bully that doesn't back down when confronted
with proof of wrongdoing. It's not even the "what are you going to do
about it" type denial, it's more "the black knight always triumphs"
simply ignoring inconvenient reality. This category ranges from
W's "signing statements" through Microsoft's response to its sherman
antitrust conviction, through the entire SCO trial. Lots of people were
amazed by SCO's behavior because they'd never encountered that kind of
thing before, but it's actually pretty common. The railroad robber
barons did this. It's how the early US settlers took the land away from
the natives. The legal system just defers the stabbing, it doesn't make
people sane or nice.

Lots of the law is actually like this: if you can convince somebody else
that you're right to the point they never even challenge you in court,
then for all intents and purposes you _are_ right. If you don't know
what your rights are, you haven't got them.

When you _do_ know your rights the solution is often either to ignore
them right back. Sometimes you have to hound them for years until actual
enforcement actions kick in. (These are the people for which a "writ of
mandamus" was invented; anything short of freezing their bank accounts
and showing up at their door to arrest them is just talk. And if the
consequences only kick in five years down the road, by which time the
issue is moot, and delay means you effectively get away with it for
another day.) But you come to expect the other side to _bluff_ a lot,
and throw in things that simply aren't enforceable as an intimidation
tactic, and the correct response is to publicly state that you don't
believe it to be enforceable (not because you're speaking in ignorance
but because you asked _your_ lawyer what's relevant in your
jurisdiction), and then to ignore the idiot blowing smoke and go about
your business. (I haven't looked up the MSVC thing because I've never
written a piece of software for Windows in my life, and don't plan to
start.)

No, I'm not surprised that Microsoft puts unenforceable stuff in its
licenses. And I'm not surprised the FSF came out with GPLv3 either. I
consider them equally unhelpful to Linux.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 19, 2007 20:08 UTC (Wed) by nix (subscriber, #2304) [Link]

Are you saying the FSF is being as insane as Microsoft? If so I can see what you're getting at, but I personally think that only Stallman's that out of touch with reality, and the rest of the organization around him mostly isn't.

Wow. Pot, kettle, black much?

(RMS, a bully? WTF? Intransigent, yes, but if he's a bully he's perhaps the least competent I've ever heard of, and I've been bullied by experts. I think your prejudices are yet again blinding you. I agree with some of what you say. Your description of that type of bully is probably correct. Your attempt to connect them to the FSF is blatant guilt-by-association at best and outright flaming crackpotism at worst.)

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 20, 2007 6:11 UTC (Thu) by landley (guest, #6789) [Link]

Random early example:

http://www.faifzilla.org/ch11.html

Starting with the bit:
> Even before the LinuxWorld show, however, Stallman was showing an
> increased willingness to alienate his more conciliatory peers.

The problem (as detailed in the above chapter from his biography) is his
explicit rejection of any compromise, ever, AND his rejection of
half-measures. If they come partway to your position, he believes you
have less leverage to bring them the rest of the way, so he doesn't
believe in incremental progress.

This means his position has tended to become more extreme and brittle
over time. As other people move towards his original position, he backs
into a corner.

I wasn't talking about bullying above, I was talking about willfully
ignoring reality to pursue an agenda.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 20, 2007 23:12 UTC (Thu) by jordanb (guest, #45668) [Link]

> This means his position has tended to become more extreme and brittle
> over time. As other people move towards his original position, he backs
> into a corner.

It doesn't follow that not compromising on principle results in that. It sounds like Stallman has a very clear idea of where he stands and is willing to work with anyone iff they are willing to work within the constraints of his principles.

Nobody can accuse Stallman of being nice, or of being friendly, or even having a basic sense of civility or tact, and I'm sure that puts a lot of people off to him. But that doesn't mean that his principles aren't sound.

Anyway, the GPL was created to ensure the four freedoms to all users of covered software. Over time legal realities change and loopholes are discovered allowing unscrupulous distributors to deprive their customers of the four freedoms, so a new version is produced. The GPL was designed to be upgradeable (I think lawyers call it 'renewable') for that reason and it has happened twice now. Neither update has caused the GPL to move away from those guiding principles, so I'm having trouble figuring out what makes the GPL version 3 unpalatable to people who accept the spirit of the GPL version 2, (which is unchanged across the versions, of course, only the legal aspects have been changed slightly). The only possible reason why one would be upset with version 3 is that they were profiting from a loophole in version 2.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 21, 2007 0:23 UTC (Fri) by landley (guest, #6789) [Link]

> Anyway, the GPL was created to ensure the four freedoms to all users of
> covered software.

My previous laptop was created to run Windows. It came with a little
sticker on it and everything. It ran Linux pretty well, though.

The intentions of the people who create a tool, and intentions of the
users of that tool, are not always on speaking terms. The users tend to
win, even if they're using a wrench to pound in a screw.

In the case of the GPLv2, it prevents projects from forking. The code
can fork, but source to each new fork must be promptly made available so
new features can be re-merged before it can drift far enough for this to
become hard to do. You don't get somebody shipping a proprietary version
plus cool new feature X for 5 years (X11 support for a new graphics card,
https support in a web app back when cryptographic export regulations
were more evil) and then _maybe_ opening the code after it no longer has
any commercial value.

Availability of source is _all_ GPLv2 asked for. It didn't place any
other restrictions on you, you must give back changes to the source but
otherwise, do what you like with it. It's simple, minimal, elegant.

GPLv2 explicitly did not cover _use_ of the software, just distribution.
The new one tries to control how the software may be used.

> The only possible reason why one would be upset with version 3 is that
> they were profiting from a loophole in version 2.

This is the kind of reasoning that gets Stallman in trouble. "You
disagree with me, therefore I know how you think." Er, no?

Here's a very fundamental reason for being upset about GPLv3: GPLv2 still
works just fine. The "escape clause" in GPLv3 is there in case something
was found to be wrong with GPLv2, especially something that rendered it
unenforceable, or a major workaround came up for the need to give back
the source. That never happened. Somebody hijacked the escape pod to
pursue a political agenda, and those of us quite happy with GPLv2 were
told that code could be relicensed without our permission, and lots of us
said "oh no you can't, we contributed under GPLv2 and we meant GPLv2
_only_".

GPLv3 was not motivated by any specific lack in GPLv2. GPLv2 continues
to do what it's always done, and it's been _upheld_ in court repeatedly.
GPLv2 is a strong, enforceable license, in the US, Germany, Israel...
The main motivation for GPLv3 seems to be that the FSF thinks it doesn't
go far enough, so the FSF wants to change it to cover what users can do
with the code.

Restricting uses you can put the code to (in the name of freedom) is a
fundamental change, and it's totally unmotivated. I never got a clear
explanation what the difference was between "prevent end users from
changing the code by burning it into ROM" and "prevent end users from
changing the code with a cryptographic signature". The effect's the
same, it's an implementation detail. And "you must let me upgrade any
system that uses my software" is a new requirement that GPLv2 did not
impose. It's actually a chilling requirement: If the World of Warcraft
servers use Busybox, can I demand root access to those servers to upgrade
that version of busybox just because I have a World of Warcraft account
that uses those servers? Where do you draw the line? I don't want to go
there. I actively want to AVOID going there. I don't even want to USE
any code that goes there, because I don't want to give root access on my
laptop out to random third parties, and my root login EXACTLY this sort
of cryptographic mechanism for controlling upgrades. GPLv2 didn't go
there, and I'm glad it didn't.

I didn't ask the FSF to do this. During the review process I asked them
_not_ to. I like GPLv2, I prefer GPLv2 over GPLv3, GPLv2 it's simple,
well understood, and there's nothing wrong with it. I see all this
unmotivated change the FSF is pushing and what runs through my mind is
Darth Vader going "I am altering the bargain, pray I don't alter it any
futher". I have no interest in going along with that.

And when you talk about the only possible reason for me, the AUTHOR OF
THE CODE, to want to stay with GPLv2 when I am licensing MY COPYRIGHT out
to other people for free, is because I must be exploiting a LOOPHOLE...
And logically the same applies to Linus Torvalds and Linux, and everybody
else who objects...

How dare you?

Now do you know why I consider Stallman so blinded by his own viewpoint
that he can't see anybody else's? Why I've written the man off as
someone whose path I do not choose to follow?

But you can't disagree with 10% of what Stallman says and agree with the
other 90%, because he considers you 100% wrong. And thus I must consider
him 100% wrong, and go my own way.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 21, 2007 3:21 UTC (Fri) by bronson (subscriber, #4806) [Link]

Great summary Rob. I'd just like to expand a bit on why a fair number of people (me included) feel that v3 actually is a major departure from v2.

We all agree that the GPLv2 does not restrict where software can run. It *should* have contained this sort of restriction, however, because the Four Freedoms have said this all along. That's the FSF's position as I understand it.

I suppose, if you squint really hard, you might be able to convince yourself that this restriction would come from either freedom 0 or freedom 3. But, really, isn't that quite a stretch? The Four Freedoms require the distributor to make source code available, yes, but I sure don't see them mandating working hardware as well. And they definitely don't say anything about requiring the user to be able to change the code if the distributor can! (This is the case that Rob mentions where the FSF somehow considers mask/OTP ROMs acceptable but somehow Flash ROMs require you to turn over your hardware keys. To me, this seems like a contorted argument based on some awfully specialized reading of FSF documents.)

However you feel about the previous paragraph, at least we can agree that this restriction was never even remotely a part of the GPLv2 itself.

Either way, this is why so many people feel like the GPLv3 is such a radical departure from v2. I hoped that the the FSF would not substantially change the license but, by my reading, they most certainly have. It's mostly GPLish, but with one significant, bizarre, additional restriction. It saddens me.

As for the last sentence... By your reasoning, jordanb, all GPLv2 authors who are unhappy with v3 must be exploiting loopholes? That's "the only possible reason"? Wow. Just wow. Do you really believe this?

As for the future, I hope somebody will take the time to truly understand v3 and use that knowledge to write a GPLv2.1. It would be substantially the same as the GPLv2, yet it would include all of v3's excellent improvements. (In fact, it probably would be exactly v3, just a little shorter and a lot less confusing. Think of it as a small but very important refactoring.)

Hey, I can dream!

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 22, 2007 11:27 UTC (Sat) by nix (subscriber, #2304) [Link]

So jordanb's attitude is *Stallman*'s fault?!

They're not the same person, you know.

(And failing to note that the authors of the GPL *always* had explicitly
political goals in mind is quite strange as well. Of course they did; this
was explicit before the GPL even existed. Complaining about it is sort of
pointless. Use the GPLv2 if you wish, but don't complain that the GPLv3 is
all nasty and political, because the GPLv2 is as well. You just choose not
to see it, and now you criticise RMS for doing exactly the same thing. And
*that* is hypocrisy.)

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 22, 2007 19:52 UTC (Sat) by landley (guest, #6789) [Link]

> So jordanb's attitude is *Stallman*'s fault?!

No. Inspired by perhaps, but what I was trying to point out is that they
show similar failure modes.

> And failing to note that the authors of the GPL *always* had explicitly
> political goals in mind is quite strange as well.

Of course they did. This is common. I eat at chick-fil-a, doesn't mean
I'm a born-again christian. I bought bathroom tile at the Habitat for
Humanity re-store, doesn't mean I wanted Austin's triangle park paved
over and covered with condominiums (not that habitat was involved in
that). I got a cat from a no-kill animal shelter; doesn't mean I'm a
vegetarian.

There are political agendas everywhere, but often they're implicit and
easily ignored. If the use of hops in beer was invented by medieval
european monks, do you convert to their religion to drink budweiser?

GPLv2 has been a useful tool for me for 15 years, and still is. I don't
care if my hammer was manufactured by "hammers for a better tomorrow" if
it doesn't affect how I can use it. GPLv2 didn't cover any use of the
licensed software, just distribution. GPLv3 imposes new requirements
that conflict with what I want to do, not because of a change in me but
because of a change in the license terms. And not because of any
impairment of GPLv2, the original explicit justification for the "upgrade
clause", but because the FSF now wants more. How am _I_ being
hypocritical for objecting to this, and making it very very clear that no
software I've ever written is licensed under the new terms?

The FSF is free to come out with any license it wants to, but in doing so
it has shown not to represent my interests and made more work for me
clarifying that my code is NOT covered by v3, and as such I oppose its
actions on practical grounds. I am free to do so, with both a big and
small F. :)

The GFDL I can simply ignore (especially since it's largely a failure and
everybody uses creative commons anyway). GPLv3 I have to go out of my
way to actively exclude, or people will try to apply it to my code
against my wishes. That's annoying.

Rob

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 19:25 UTC (Wed) by Zack (guest, #37335) [Link]

>>So what freedom would I lose if I "follow [Torvalds]"? The freedom to run modified Linux on a Tivo?

Hardly any freedom. I don't understand why poeple insist on harping on the actual Tivo device, reducing the argument mostly to a burn the strawman exercise.
From what I can understand it can already be "worked around" by using mythtv on a small computer.

As I see it, the point of Stallman (forever "blessed" with Cassandra syndrome, it seems) is that computer devices are already getting smaller, more specialized and more numerous. In 10 (100, 1000 etc) years, the desktop as we know it might be a niche market, most computing needs tended to by ubiquitous computing devices.
It's good to be critical of restricting the use of linux on devices that may at one point constitute the lion's share of all computers.

*Right now* everything is going swimmingly, and the loss of access to a single niche device isn't very much of a problem. But to blindly assume that simply because it is this way, that it is always going to be this way
is shortsighted.

Corporate CEOs get big salaries for trying to see into the future. The free software community has access to one with a proven track record at no cost. Maybe people should listen to him more often.

Stallman on free software, Torvalds and Microsoft (LinuxWorld)

Posted Sep 12, 2007 20:02 UTC (Wed) by atai (subscriber, #10977) [Link]

>>Corporate CEOs get big salaries for trying to see into the future. The free software community has access to one with a proven track record at no cost. Maybe people should listen to him more often.

This is very well said.

Self-aggrandizement?

Posted Sep 12, 2007 20:27 UTC (Wed) by man_ls (subscriber, #15091) [Link]

Looks like the same old same old.

Some things have to be repeated again and again, and sadly people still don't get them.

So what freedom would I lose if I "follow [Torvalds]"?

The freedom to run modified software on your devices. In this particular instance, Linus says that it is OK to take the kernel and TPM it to death, since you can look at the code and that is good enough for you.

The fact that people find ways to limit this freedom now doesn't mean you have to put it easy for them. Not buying from them is an excellent way to preserve that freedom, another one is not to develop software with licenses fit for their use.

Stallman's self-aggrandizement ("I developed version 3" -- all by yourself, eh?) is a bit irritating too.

Stallman was the driving force behind GPLv3, it is hard to deny him that. On the other hand, when asked if the free software community can win the fight against Microsoft he says:

Nobody knows who will win this fight, because the outcome depends on you and the readers.

I like that. We as a whole may as well declare victory over proprietary Unix, and that is no small feat in itself. But we cannot rest on our laurels, there is much more to do, and many of us (me included) cannot do much.