GPL: either version 3 of the license, or any later version (Announce)
> > I look at GPL3 as the latest version of one of the best open source
> > licenses. I generally agree with the philosophy that if you pick
> > something
>
> We roughly agree, specially on the bit that forcing people to do the
> morally good thing is hard.
Here's what I don't understand: how can you force anybody to do anything? Especially legally, how can you force somebody to share changes? Or force somebody not to reverse engineer? Etc. etc. It seems legalistic people are obsessed with forcing other people to do things. It's an annoyingly confusing idea. I'm all for sharing code, very much so, but "forcing" somebody just seems odd. I just can't imagine that ever working (though it does, rarely, which is weird).
But we all agree that open source is better than buggy, abandoned, proprietary, expensive software.
> And that brings us to the force bit, and the place where we differ. IMHO
> the GPL should only be applied in special, strategic cases, and in most
> common cases the force bit hurts more than it helps, while your position
> could be summoned up as "force, just in case"
GPL is heavily tied to Linux and FSF and *nix-centric philosophy. As such, their code barely works (if even) on non-POSIX systems. And for things like GCC, making modifications is difficult because it's so complex (and mirroring sources is huge). So that's all a pain.
> And I already described one prime example of that pain: getting usage of
> the GPL approved in corporate hierarchies outside the handful of IT
> multinationals. Even if the planned usage would be completely in line with
> the license.
"Want" isn't as important as "need". The only reasons to complain about GPL is if it literally cannot be used, i.e. when it's out of your control. If you want to use DUGL with paq8o8 (I don't, it's just an imaginary example), you can't because it's disallowed. Just preferring something else isn't a real reason.
But yeah, *BSD (or MIT, Apache, ZLIB) is simpler. I mean, GPL is very very verbose and legalistic. And nobody reads the damn thing anyways (else all billion forced-included copies of COPYING would say "51 Franklin St." instead of old obsolete snail mail addresses that nobody writes to anymore anyways).
> The GPLv3 worsens this, it tries to screw an handful vendors out of some
> minor customization and scripts. (and usually these are not the smartest
> ones, since otherwise they wouldn't get caught.
It doesn't actively screw anyone, AFAICT, but it does whine against TiVoization (which I think is pretty rare) as well as patents (losing battle, they're not going anywhere). I think most corporations are honestly afraid that they will have to give up some or even all of their patents (software or otherwise) if they ever redistribute GPLv3 software. I don't see how that could happen, but fear makes people do crazy things.
And of course the biggest open source juggernauts come from FSF / GNU, e.g. Emacs and especially GCC, so they have a lot of political sway just out of practicality because everybody and their brother uses some of their software (except Microsoft).
> On the flip side however, people using the GPLv3 only internally might have
> to rework their infrastructure if something tied to a GPLed part suddenly
> faces outwards. And decide to avoid the GPLv3 all together, IMHO rightfully
> so.
I really don't see much difference between v2 and v3. I mean, there's nothing shocking in there. Patents are the only thing, and it makes perfect sense not to disingenuously let someone use software just to later sue them for patents. I mean, it's really annoying how everybody sues everybody else. Sharing code and fixing bugs, getting things done, etc. is good, but legalities suck.
> It does finely continue the GPL tradition of preparing for the worse
> though, damn the consequences
They have a lot of sway. GPLv2 is by far the most popular open source license, and Linux is by far the most (over)hyped open source OS. It's almost impossible not to deal with it somehow.
> It caused the BSDs to defect (they halted upgrading of GNU software in
> base, and stick to the GPLv2 while they work frantically on deGNUifying
> base), and the Linux kernel to change its base license to GPLv2 only.
The *BSDs are weird. Well, honestly, everybody thinking from a purely political licensing standpoint comes across as weird. It's people like Linus himself who take a more pragmatic approach, which is more reasonable. I mean, it's not like I love proprietary software either, but hey, some free software zealots can really cut off their own nose to spite their face. I mean, when you start rejecting perfectly working software because "our free is more free than your free", you've missed the point, IMHO. That seems to be all too common these days, sadly. (And when your wifi doesn't work because the firmware ain't included and you can't download it, you'll wish somebody included it anyways, free or not.)
> And I really wonder if whatever v3 brings is worth carving up the united
> OSS front.
Note that they didn't intentionally do this. In fact, I would say they need to be more careful next time because of it. *BSD is smaller than Linux but still important. A working consensus is almost always better than a purely idealogical standpoint anyday.
If it were up to me, I'd add a few more ideas to the (already confused pot):
1). You don't have to mirror (big!) sources if they're already mirrored in 10 other places.
2). You don't have to include "COPYING" anymore, everybody already has it (not that they care).
3). It's needs to be a lot shorter in length, nobody's going to read 30 kb of text in one sitting.
4). It needs to deal with the whole "or later" issue. The fact that some projects are "v2 only" while others are "v2 or later" is insane, IMHO.
5). It needs to take a stand on whether GPL software built with non-free tools is "free" or not (e.g. FreeDOS kernel). I would say yes, but clearly Debian and Fedora disagree.
6). Patents suck, but they aren't going away. It needs to be more careful to say, "Only relevant patents to this particular software apply here. Just don't sue us over any accidental overlaps here, okay? All others are irrelevant in this instance." In other words, make sure to clarify that it doesn't automatically import all your patent portfolio to /dev/null by just redistributing this.
> P.s.
>
> I found an old document I wrote that discusses the "force" argument a bit,
> as part of a very old BSD vs GPL discussion:
>
> http://www.stack.nl/~marcov/bsdvslgpl.txt
Shared libs? (sigh) I don't know why nobody can agree on whether it counts or not. But my understanding of that is fuzzy.
Yes, BSD is easier, but it's really true that without GPL, a lot of hoarding and subsequent "fork / close / sell" attempts would be made. I've seen it happen.
Honestly, I just hate licensing, it just always gets in the way and impedes real progress. It's just such a pain. People would rather argue for years over licensing rather than actually write or patch real software. I hate licensing! It's so useless and destructive. (And BTW, software patents aren't so great either. How can you sue somebody for "accidentally" violating something? Ridiculous.)
> note that these were quickly written down bulletpoints for what was going
> to be the finishing sheet(s) of a lecture, and it is deliberately put very
> crude and argumentative to provoke discussion after the lecture. Don't
> worry, my own opinions are slightly more nuanced (the biggest flaw is
> heaping LGPL and GPL together)
Open sourcing or abandonware never happens. So much software is literally lost to the ages because it literally dies on the rack because nobody mirrored it, nobody cleared the copyright, or else nobody wanted to share it, even though it's no longer sold. I hate that. That's another advantage of GPL, at least (in theory) the software will live on.
Submitting fixes is fine in theory, but upstream is often VERY picky about what they will accept anyways, and they don't like certain platforms (ahem).
The real enemy of GPL (besides selfishness) seems to be that people are just too lazy to bother. BSD is a bit worse in that regard, they don't feel like it's their responsibility. Hence GPL carries more responsibility due to its demands.
In theory, you would think BSD would be more fruitful or more widely accepted, but it doesn't appear to be, surprisingly. GPL is far more popular. It seems that people genuinely don't want their code to be pilfered to closed source (though I don't think most people would bother trying that anyways), plus GPL is (incorrectly) often assumed to be a good way to keep people from making money off it (heavily reduces the incentive to almost nothing). GPL is often equated with "non-commercial only", which is wrong but still implied.
In reality, like I said, all this licensing talk sometimes puts people off. I'm mostly agnostic (out of pragmatism, you can't defeat the world, you have to sometimes go with the flow), but I'd honestly prefer the simplest possible, or even nothing ("public domain").
I mean, at the end of the day, who cares? If you just want to get the job done, why would licensing matter at all?
Complete thread:
- mTCP Open Source release - mbbrutman, 27.05.2011, 23:47 (Announce)
- mTCP GPL - ecm, 27.05.2011, 23:58
- GPLv3 - mbbrutman, 28.05.2011, 00:20
- GPL: either version 3 of the license, or any later version - ecm, 28.05.2011, 00:55
- GPL: either version 3 of the license, or any later version - mbbrutman, 28.05.2011, 01:09
- GPL: either version 3 of the license, or any later version - ecm, 28.05.2011, 01:18
- GPL: either version 3 of the license, or any later version - mbbrutman, 28.05.2011, 01:25
- download, compile, ... - ecm, 28.05.2011, 01:30
- GPL: either version 3 of the license, or any later version - mbbrutman, 28.05.2011, 01:25
- GPL: either version 3 of the license, or any later version - marcov, 30.05.2011, 09:37
- GPL: either version 3 of the license, or any later version - mbbrutman, 30.05.2011, 15:32
- What does the GPL allow? - ecm, 30.05.2011, 16:08
- What does the GPL allow? - mbbrutman, 30.05.2011, 16:30
- GPL something - ecm, 30.05.2011, 17:08
- GPL something - mbbrutman, 30.05.2011, 18:05
- theoretical licensing problems and such - ecm, 30.05.2011, 18:28
- licensing again... executables - ecm, 30.05.2011, 22:34
- licensing again... executables - mbbrutman, 31.05.2011, 03:33
- licensing again... executables - ecm, 31.05.2011, 13:59
- licensing again... executables - mbbrutman, 01.06.2011, 01:16
- licensing again... executables - ecm, 01.06.2011, 01:53
- licensing again... executables - mbbrutman, 01.06.2011, 02:17
- licensing again... executables - ecm, 01.06.2011, 15:49
- licensing again... executables - mbbrutman, 01.06.2011, 02:17
- licensing again... executables - ecm, 01.06.2011, 01:53
- licensing again... executables - mbbrutman, 01.06.2011, 01:16
- licensing again... executables - ecm, 31.05.2011, 13:59
- licensing again... executables - mbbrutman, 31.05.2011, 03:33
- GPL something - mbbrutman, 30.05.2011, 18:05
- What does the GPL allow? - Rugxulo, 30.05.2011, 18:14
- mTCP in DOSEMU; Linux kernel developers' GPLv2-only reasons - ecm, 30.05.2011, 22:38
- What does the GPL allow? - marcov, 30.05.2011, 22:49
- What does the GPL allow? - mbbrutman, 31.05.2011, 03:41
- What does the GPL allow? - mbbrutman, 31.05.2011, 04:18
- What does the GPL allow? - Japheth, 30.05.2011, 18:47
- Gift trolls and other curiosities - ecm, 30.05.2011, 19:20
- Gift trolls and other curiosities - mbbrutman, 30.05.2011, 19:54
- editing - ecm, 30.05.2011, 20:03
- Gift trolls and other curiosities - mbbrutman, 30.05.2011, 19:54
- Gift trolls and other curiosities - ecm, 30.05.2011, 19:20
- What does the GPL allow? - marcov, 30.05.2011, 22:38
- GPL something - ecm, 30.05.2011, 17:08
- What does the GPL allow? - marcov, 30.05.2011, 22:37
- What does the GPL allow? - mbbrutman, 30.05.2011, 16:30
- GPL: either version 3 of the license, or any later version - marcov, 30.05.2011, 22:35
- GPL: either version 3 of the license, or any later version - mbbrutman, 30.05.2011, 22:56
- GPL: either version 3 of the license, or any later version - marcov, 02.06.2011, 16:10
- GPL: either version 3 of the license, or any later version - mbbrutman, 03.06.2011, 04:12
- GPL: either version 3 of the license, or any later version - marcov, 03.06.2011, 13:36
- GPL: either version 3 of the license, or any later version - Rugxulo, 03.06.2011, 16:45
- GPL: either version 3 of the license, or any later version - marcov, 03.06.2011, 13:36
- GPL: either version 3 of the license, or any later version - mbbrutman, 03.06.2011, 04:12
- GPL: either version 3 of the license, or any later version - marcov, 02.06.2011, 16:10
- GPL: either version 3 of the license, or any later version - mbbrutman, 30.05.2011, 22:56
- What does the GPL allow? - ecm, 30.05.2011, 16:08
- GPL: either version 3 of the license, or any later version - mbbrutman, 30.05.2011, 15:32
- GPL: either version 3 of the license, or any later version - ecm, 28.05.2011, 01:18
- GPL: either version 3 of the license, or any later version - mbbrutman, 28.05.2011, 01:09
- GPL: either version 3 of the license, or any later version - ecm, 28.05.2011, 00:55
- GPLv3 - mbbrutman, 28.05.2011, 00:20
- mTCP and SwsVPkt - Japheth, 31.05.2011, 09:03
- mTCP and SwsVPkt - mbbrutman, 01.06.2011, 03:11
- mTCP and SwsVPkt - Japheth, 01.06.2011, 13:14
- mTCP and SwsVPkt - mbbrutman, 01.06.2011, 15:04
- polling - ecm, 01.06.2011, 16:07
- polling - Japheth, 01.06.2011, 17:12
- polling - ecm, 01.06.2011, 17:27
- polling - Japheth, 01.06.2011, 17:50
- polling - ecm, 01.06.2011, 17:55
- polling - Japheth, 01.06.2011, 18:39
- polling - ecm, 01.06.2011, 18:49
- polling - Japheth, 01.06.2011, 18:39
- polling - ecm, 01.06.2011, 17:55
- polling - Japheth, 01.06.2011, 17:50
- polling - ecm, 01.06.2011, 17:27
- polling - Japheth, 01.06.2011, 17:12
- mTCP and SwsVPkt - Japheth, 01.06.2011, 13:14
- mTCP and SwsVPkt - mbbrutman, 01.06.2011, 03:11
- mTCP GPL - ecm, 27.05.2011, 23:58