GPL version 2 only for BusyBox 1.3.0.
Rob Landley rob at landley.net
Thu Sep 14 06:43:08 UTC 2006
Erik Andersen wrote:
> On Wed Sep 13, 2006 at 11:07:51PM -0400, Rob Landley wrote:
> > I'm planning to simplify the BusyBox license to GPLv2 only tomorrow, so
> > that BusyBox 1.3.0 will be GPLv2 only when it ships.
> >
> > Are you ok with that?
>
> Sure, works for me.
So, I confirmed that Erik has objection to simplifying the BusyBox license to
GPL version 2 only, dropping the "or later". I also asked the Software
Freedom Law Center, and on Monday they confirmed they're still happy to
represent us as a GPL version 2 project. And I've brought it up several
times on this list over the past few months and nobody's come up with a
compelling argument against it.
So I'm going to do it.
Obviously the license on the already released versions isn't changing, and in
fact the 1.2.2 release I'm in the process of putting together will be "GPLv2
or later". Probably the last version that is. If somebody really wants to
do a GPLv3 fork based on that, I honestly don't care. (Auditing whether or
not you can actually use various bits of it under GPLv3 is your problem, not
mine.)
But BusyBox 1.3.0 and so on will be under the same license as the Linux
kernel: GPL version 2.
This lets us merge the diethotplug code, and ntpclient, and not have to audit
our codebase for things like the fact that our fsck_minix.c was written by
Linus Torvalds (and although the permission statement he put on it back in
1992 is a bit unclear -- "usable under the terms of the GNU Copyleft" -- he
made it very clear back in 2000 that his code has always been GPLv2 only).
It also means we can stop arguing about it. :)
I'll update the boilerplate on the individual files as I get around to them.
Rob
--
Never bet against the cheap plastic solution.
GPL version 2 only for BusyBox 1.3.0.
Glenn L McGrath bug1 at ihug.co.nz
Thu Sep 14 08:08:53 UTC 2006
This is pretty disappointing to me... I had hoped you would have taken
a more conservative approach.
The politics of Free Software is important for me, and the FSF champion
the cause more than any other organisation, as such i think i owe them
my support.
When GPLv3 comes out i expect i will stop releasing code under GPLv2.
I havent contributed any patches for a long time so im sure you
will get along fine without me.
Best of luck
Glenn
GPL version 2 only for BusyBox 1.3.0.
Erik Andersen andersen at codepoet.org
Thu Sep 14 17:28:00 UTC 2006
On Thu Sep 14, 2006 at 08:08:53PM +1200, Glenn L McGrath wrote:
> This is pretty disappointing to me... I had hoped you would
> have taken a more conservative approach.
This is the conservative approach... The non-conservative
approach, which is implicit assumption behind the current 'GPLv2
or later' is that any later iteration of the GPL is going to be
more acceptable than the GPLv2. I have some concerns with the
current GPLv3 drafts (which are well beyond the scope of this
mailing list). For all I know, Richard Stallman may decide he is
tired of being poor and altruistic and at EvilSoft's request may
in his old age draft up a GPLv4 which states "This code now
belongs to EvilSoft". Unlikely? Sure? But it is within the
realm of possibility. Until such time as a final GPLv3 is
actually released upon an unsuspecting world, it is premature to
pass final judgement. The problem implicit in the "or later"
clause is that there is no a priori way to pass judgement on all
possible future license iterations and thus requires that
everybody accept them sight unseen. I don't really like that
idea.
> When GPLv3 comes out i expect i will stop releasing code under GPLv2.
When the GPLv3 comes out, it will be time for many many
organizations (including BusyBox) to evaluate the merits of the
GPLv3, and it may well be time for BusyBox to switch over to
GPLv3. Or not. Regardless, such a change should be just as
feasible then as it is now.
-Erik
--
Erik B. Andersen http://codepoet-consulting.com/
GPL version 2 only for BusyBox 1.3.0.
Rob Landley rob at landley.net
Thu Sep 14 18:53:23 UTC 2006
On Thursday 14 September 2006 1:28 pm, Erik Andersen wrote:
> When the GPLv3 comes out, it will be time for many many
> organizations (including BusyBox) to evaluate the merits of the
> GPLv3, and it may well be time for BusyBox to switch over to
> GPLv3. Or not. Regardless, such a change should be just as
> feasible then as it is now.
We do have the source control history and know which patches are going in
since the switch, and it's not like we haven't cleaned out code we weren't
sure of the license of before. (Speaking of which I still plan to replace
the "with advertising clause" BSD code that's currently in the tree, although
that's partly because a fresh implementation could significantly improve upon
it.)
But just the quick check of the BusyBox codebase I did came up with instances
of GPLv2-only code, some of the code I'd like to add today is GPLv2 only, and
I really don't want to try to police this issue much more than I already
have. I really like having one license, and that license being the same one
that's on the Linux kernel.
But yeah, if the Linux kernel switched to GPLv3, I'd put in the effort to
follow suit.
Rob
--
Never bet against the cheap plastic solution.
GPL version 2 only for BusyBox 1.3.0.
Bruce Perens bruce at perens.com
Thu Sep 14 21:18:31 UTC 2006
Rob Landley wrote:
> So I'm going to do it.
>
Oh thanks tremendously much for not even waiting for a reply to your
email before making this decision. It does feel spiteful from here.
Bruce
GPL version 2 only for BusyBox 1.3.0.
Rob Landley rob at landley.net
Fri Sep 15 07:06:34 UTC 2006
On Thursday 14 September 2006 5:18 pm, Bruce Perens wrote:
> Rob Landley wrote:
> > So I'm going to do it.
> >
> Oh thanks tremendously much for not even waiting for a reply to your
> email before making this decision. It does feel spiteful from here.
We've been talking about this topic on and off for almost NINE MONTHS:
http://busybox.net/lists/busybox/2006-January/018001.html
We've had lots of discussions about here in lots of threads since then, to the
point that I think most of the people in this list are outright _SICK_ of the
topic.
You showed up at the last minute, after I already _made_ my decision, and all
but announced it earlier that same day (eight and a half hours before you
posted to the list), and I quote:
http://busybox.net/lists/busybox/2006-September/024397.html
> P.S. The SFLC says they have no problem doing license enforcement for a
> GPLv2 only project, Larry Doolittle's ntpclient (which I'm looking at as a
> candidate for inclusion although that's not the same as actually deciding
> to add it) is GPLv2 only, and so is the version of BBSH I'm currently
> working on. That last one might just be a hint how I'm likely to resolve
> this...
Notice the end there, "The new code I am writing is GPLv2 only, and I'm going
to add it to the tree." What exactly do you think that means?
This was after doing lots of research and checking with lots of people and
posting and reading an awful lot of messages on the list. Why was the above
not yet an official announcement? Because I had one (and only one) step
left: confirming that Erik was ok with it. I had bounced the idea off him
before, but wanted his explicit approval before making a formal announcement.
Then you showed up and asked us to hold everything for you after we hadn't
heard a thing out of you in seven years. From here, that felt amazingly
arogant.
Despite that, I carefully read your message, thought about each point you made
in case you had something new that nobody had brought up before (you hadn't)
or in case I could think of something I'd overlooked (I did, and it supported
my decision), and replied to your message. I stopped what I was doing and
re-evaluated the whole darn mess again, just to be _darn_ sure. I spent
hours writing a reply (which involved research), saved it in the drafts
folder rather than sending it immediately, took the bus home, fired up my
laptop and read it _again_ (and toned bits of it down to be more polite and
conciliatory than the first draft), and then sent it.
Then I sent an email to Erik to get his final confirmation, pointing out in
the message that you were arguing with me about it in case that made a
difference. He did not object, so I went through with it.
There are a number of things I haven't been impoliticly explicit about, and
was trying to _avoid_ bringing up. To be honest, I don't trust the Free
Software Foundation after they gratuitously picked a fight with Mepis. Mepis
had an official partnership with Ubuntu that involved press releases which
quoted Ubuntu's founder stating his support for the partnership, yet Mepis
couldn't then rely on Ubuntu's repository to distribute software Mepis was
using unmodified from Ubuntu? That struck me as INSANE. The FSF attacked
someone who hadn't done anything wrong, trying to make an example out of them
for no good reason I could see except possibly to draw attention to itself as
custodian of the GPL, and possibly trying to make people view GPLv2 as
dangerous. (Instead I viewed the FSF as dangerous.)
You're asked me to trust these people to change the terms of the GPL. Not
likely.
And lest this seem like a snap judgement based on an isolated incident, I have
a long history with the FSF. I first started studying the GPL when it was
the license on EMX, a port of GCC to OS/2 which I used (and even wrote
documentation for: http://www.os2bbs.com/Download/gnu.html) back in the early
90's.
An email exchange I had with RMS in 1998 seems to have precipitated his
big "call it GNU/Linux dammit" campaign, which I apologized for to the
Linux-kernel mailing list:
http://www.kernel-traffic.org/kernel-traffic/kt20020805_178.html#1
(And yes I know he was already calling stuff GNU/Linux. I told him if he
wanted to position Hurd as a successor to Linux then he needed to attach the
GNU name to Linux and then position GNU Hurd as an upgrade to GNU Linux. He
mentioned that he'd been calling it GNU/Linux and promised to use that phrase
more. Sigh. At least he stopped complaining about people using the
word "Linux" at all, which he _was_ doing at the time. I'd point you to the
cached copy of the gnu.org web page from 1998 and their "why you shouldn't
call it Linux" page, but the wayback machine appears to be down again.)
Later when I drove up to Boston to interview him once (in 2000 I think), he
borrowed my car and left it unlocked when he parked it, and I still have the
sweatshirt that someone gave him which he immediately gave away because it
had the word "snow" written on it over a little snowflake. (He apparently
refuses to wear anything that has writing on it, and yes I asked: even the
FSF's own shirts.) Why bring this up? Because I have firsthand experience
that even in little things, he seems to think in absolutes to the point of
being brittle.
After that friends of mine hosted Richard Stallman at a barbecue in Austin
Texas (while I was out of town) but they told me he hit on the hostess's 14
year old daughter, insulted the people who had prepared the food, and
basically got everybody there unwilling to ever speak to him again. I
sometimes hang out with Eric Raymond who occasionally wonders what happened
to the friend he met at a science fiction convention in the 1970's;
apparently he wasn't always like this. (My friend Stu Green in Austin knew
Richard back at MIT too. He's one of the people Richard pissed off at the
barbecue.)
Mepis is by no means the first time Richard has attacked his allies. I
remember when Richard accused Tim O'Reilley of being a parasite for producing
good books, and the rant he wrote justifying his this is still part of the
official FSF literature:
http://www.gnu.org/software/gsl/manual/html_node/Free-Software-Needs-Free-Documentation.html
And you wonder why I trust the judgement of Linus Torvalds and the people
around him more than I trust the judgement of Richard Stallman and the people
around _him_?
Can the FSF put out a license that fails to gain traction? How about the GNU
Free Documentation License which isn't nearly as widely used as Creative
Commons? The FSF keeps insisting that nobody but them ever thought about
freedom, yet Project Gutenberg was founded by Michael Hart in 1971, a dozen
years before the FSF.
As for implying I owe the FSF any sort of alleigance: There _was_ no
proprietary software industry to speak of before the Apple vs Franklin
decision in 1983 extended copyright to cover binary ROM images; before that
you had to distribute source code to get copyright protection. (Go get a
copy of "Legal battles that shaped the computer industry" by Lawrence Graham,
it has a chapter on this. A young Bill Gates whined about this very topic in
a 1980 interview here: http://www.thesync.com/geeks/gates1980.html -- note
that he was already lobbying congress to change the law, but didn't have any
clout to speak of back then.)
And the setup for Apple vs Franklin was the rise of the microcomputer. The
PDP-8 FAQ says:
> By late 1973, the PDP-8 family was the best selling computer in the world,
> and it is likely that it was only displaced from this honor by the Apple II
> (which was displaced by the IBM PC).
The same FAQ puts the PDP-8's entire production run at 50,000 machines total.
According to the Innovator's Dilemma, the Apple II sold 43,000 machines in
its first two years (1977-1978), and in 1982 the IBM PC sold out its 250,000
machine initial production run in three months. Before about 1980, the
industry couldn't support a proprietary software industry at anything other
than the Mom 'n pop level. Micro-soft survived on contracts from MITS, Radio
Shack, and then IBM for its first five years, despite being a tiny hole in
the wall enterprise. The FSF tries to take credit for the idea of free
software but before Gates' 1976 "letter to hobbyists" there wasn't even the
_idea_ of anything else, and the Univac shipped in 1950.
The FSF was never a grand visionary anything, it was a conservative
reactionary movement founded to return to the glory days of the 1970s, and it
started by cloning existing Unix technology and Unix culture to do it. This
was a good thing, but it was not remotely innovative.
The FSF rose to prominence in the 1980's because MIT gave it a high-bandwidth
FTP site on athena.ai.mit.edu. Before the dismantling of the NSFNet's
Acceptable Use Policy (which started to have alternatives in 1991 and went
away completely in 1994, see
http://landley.net/history/mirror/nsfnetaup.txt), free FTP sites to host code
on were almost impossible to come by, and thus people like Larry Wall were
willing to sign over their code to the FSF and the GNU project in order to
get hosting for it. When alternatives like metalab/sunsite/ibiblio and
cdrom.com started showing up (heck, _geocities_), that advantage went away.
These days with cable modems and sourceforge and OSL, most people don't even
remember a time when hosting was painful to get.
And yet despite having this huge advantage for years, the GNU project had
already stalled without producing a bootable system by the time the Linux
kernel was created in 1991 (or Linus would likely have used that instead of
Minix). The FSF's slow slide to irrelevance may have started when its FTP
site stopped being enough for it to dictate terms to contributors, but the
fundamental problems are that it's no longer needed, it's not contributing
anything useful, and it's actively undermining itself on a regular basis.
Yes, once upon a time the FSF created a C compiler (something Fabrice Bellard
did for the obfuscated C code project). And yes, the GPLv2 is quite nice,
and what I consider to be the one truly vital contribution that Stallman
made. I don't see it as quite the work of isolated genious some people do
(he developed it in response to forks of Emacs that weren't giving code back,
and between the Emacs license, GPLv2, and GPLv2 it took several tries to
settle on the final form), but I do very much like GPLv2. I've praised it in
the past, and been highly critical of BSD licensing:
http://www.linuxtoday.com/news_story.php3?ltsn=2001-05-10-002-20-PS
The best thing I can say about the FSF these days is that they sometimes let
go gracefully, such as when gcc forked away from the FSF, the new fork (egcs
run by Cygnus) rendered the original irrelevant, and the FSF shut down their
in-house project and handed over the name. A similar thing happened with
glibc->glibc2. Sometimes, they're graceful enough to get out of the way of
people doing real work.
I also appreciate that I have the choice _not_ to buy in to their latest
project. You're complaining that I am exercising that choice in what I
choose to maintain.
My first exposure to collaborative distributed software development wasn't
Linux, it was the WWIV bulletin board. We didn't even have "patch", we had
*.mod files that had instructions how to modify the source code.
Instructions written in english. Applying these mod files to my friend's
bulletin board (and debugging conflicts, and eventually coming up with new
modifications and sending out my own files into the world) is how I learned
C. (If you ever come across the "boxed color-coded onliners with bar graphs
of comparative usage" mod for WWIV... I can only say I was young, and it
seemed like a good idea at the time, I.E 1990.)
Anyway, I realize you do not like the way I'm maintaining BusyBox. It's
obvious from the way you're attempting to undermine me publicly on my own
list. I'm sorry about that, but I'm the one making the decisions because I'm
the one doing the _work_. If you would like to maintain a GPLv3 fork of
BusyBox, feel free. If not, please go do something useful.
> Bruce
Rob
P.S. In case it isn't clear by now, one of my other hobbies is computer
historian. I like to be informed about the context in which decisions are
made. What exasperates me about this exchange is that you're ignoring the
context in the mailing list archive, expecting us to drop everything and
reiterate it all for you, and that you're doing this solely because of who
you are. Correction: because of who you were seven years ago. And in fact,
when I DID stop and take your concerns seriously, ignoring the insulting
title of your message and the general condescending tone, you're now throwing
a hissy-fit that I still disagreed with you. That's uncool, and I'm three
hours late for bed.
P.P.S. Is this enough discussion for you, or would you like more? I haven't
actually changed the license on any of the individual files yet, and the only
people to commit to the tree since I made the announcement are me, Denis, and
Bernhard. However, after reading this thread so far (I know you didn't read
the previous threads but you are reading _this_ one, right?), do you really
think your chances of convincing me to change my mind are very good? I
DISAGREE WITH YOU. This happens. Deal with it.
--
Never bet against the cheap plastic solution.
Copyright 2006 http://lists.busybox.net/pipermail/busybox/