Open Source Initiative loses corporate status
We are concerned about the impact that the suspension of the Open Source Initiative could have on open source developers, users, projects, and associated investors and vendors. The 451 Group has clients in all of the above categories so we believe it is appropriate to inform them of the suspension of the Open Source Initiative's legal status and how it might impact them. We are in the process of creating a formal analysis of the situation for 451 Group clients. We also believe that the potential impact is significant enough that, while the bare facts are already public, the issue deserves to be brought to the attention of the wider open source community. We will let the members of that community come to their own conclusions about what it means to them." Your editor is still struggling to figure out whether there will be any "impact" at all.
Posted Oct 6, 2009 18:39 UTC (Tue)
by ncm (guest, #165)
[Link] (16 responses)
Posted Oct 6, 2009 21:44 UTC (Tue)
by dmarti (subscriber, #11625)
[Link]
Posted Oct 7, 2009 1:10 UTC (Wed)
by jamesh (guest, #1159)
[Link] (14 responses)
Posted Oct 7, 2009 5:41 UTC (Wed)
by rahulsundaram (subscriber, #21946)
[Link] (13 responses)
Fedora Project has the most comprehensive list of Free and open source licenses in the world that I am aware of, far more than either FSF or OSI.
http://fedoraproject.org/wiki/Licensing:Main
If you do a quick check on the RPM metadata which embeds the license tags, GPL, LGPL, MIT, BSD and Apache remain the popular ones. With GPLv3, they are all compatible with each other as well.
OSI could have done several useful things but I am not sure they are making enough of a impact as they should be at this point.
Posted Oct 7, 2009 9:20 UTC (Wed)
by rsidd (subscriber, #2582)
[Link] (12 responses)
3-clause BSD and Apache 1.0 are not compatible with GPL (the advertising clause). The renaming clause of Apache (derived products may not use the word Apache) is also GPL-incompatible I believe. Sun's CDDL is incompatible with GPL (so, no kernel-space ZFS for Linux). Even GPL2-only software (such as the Linux kernel) is not compatible with GPL3. And so on.
Basically, in most cases, GPL is only compatible with itself.
Posted Oct 7, 2009 10:03 UTC (Wed)
by rahulsundaram (subscriber, #21946)
[Link] (11 responses)
http://www.gnu.org/philosophy/license-list.html#GPLCompat...
Since you mentioned Sun CDDL, you might also be interested in reading about that among other things at
Posted Oct 8, 2009 1:18 UTC (Thu)
by rsidd (subscriber, #2582)
[Link] (10 responses)
I also think the article you link to is unfair to the impact of the BSD and MIT licenses. Historically they gave us the networking stack, the X window system and a huge set of utilities. If the original TCP/IP stack had been GPL'd, all the commercial guys would have rolled their own, and we wouldn't have the internet. If X had been GPL'd, the Unix world would have developed a bunch of proprietary and incompatible interfaces. There are GPL'd implementations of SSH, but OpenSSH is the standard. I can't think of a single piece of GPL software that has become an industry standard the way these things have. RMS himself recommended the BSD licence for Ogg Vorbis, if I remember right, because the goal was to try and make it a standard. Of course, it's all "GPL compatible" which means it can be embraced and extended -- in one direction only. The FLOSS community is well known for protesting loudly when a certain commercial company does that.
Posted Oct 8, 2009 1:23 UTC (Thu)
by foom (subscriber, #14868)
[Link] (1 responses)
That's not just "in practice"! The GPL itself contains that restriction:
Posted Oct 8, 2009 1:29 UTC (Thu)
by rsidd (subscriber, #2582)
[Link]
Posted Oct 8, 2009 14:16 UTC (Thu)
by dwheeler (guest, #1216)
[Link] (7 responses)
I'm the author of the referenced article. I completely agree that X-windows, TCP, and so on are important works, and we should thank the folks who developed them (thanks!).
But it's not true that all important FLOSS works are BSD/MIT licensed. There are also GPL'ed works that are very widely used and mostly "own" their niche. GCC and MySQL come to mind immediately. Also, while there are many *BSD users, they are completely dwarfed by the number of Linux kernel users; Linux has far broader device support and far greater CPU scaleability.
That is not the primary point I'm trying to make, anyway.
At no time do I claim that everyone should choose the GPL. Heck, I've released different software under the MIT, LGPL, and GPL licenses myself. I do claim that you should choose a GPL-compatible license, which includes the BSD-new and MIT licenses.
Most people who prefer the GPL or LGPL licenses will aid projects with BSD and MIT licenses; even Stallman recommends helping such projects, and using their license.
In contrast, if you choose a GPL-incompatible license, many people will not help, and maybe even create a competing project.
And that is my point - choosing a GPL-incompatible license is a bad idea.
XFree86 is a case in point: the project lead, of the project
which was the X implementation of the time,
tried to switch from the
MIT license to an new license that was GPL-incompatible.
That "minor" license change had little effect, except that it was GPL-incompatible. But that's like saying "other than that, Mrs. Lincoln, how did you enjoy the play?" GPL-incompatibility has a massive impact. Changing XFree86 to a GPL-incompatible license was so massive that it caused a rare successful project fork.
Which is pretty strong evidence that you should choose a GPL-compatible license.
The parent post complained that "in practice, 'GPL compatible' means 'compatible only if the resulting combined work is licensed under the GPL' ... it's a one-way street". Well, yes, it's a "one-way street" if you combine an MIT-licensed work with a GPL'ed work; it essentially becomes GPL'ed. But it's also a one-way street if you combine an MIT-licensed work with a proprietary work; it becomes proprietary.
BSD-style licenses are specifically designed to allow the software to be "captured" by another project and not returned to the original project. If you don't like the "capturing" aspect of BSD-style licenses when you release an OSS project, use a different license instead, such as the LGPL or GPL.
Posted Oct 8, 2009 15:26 UTC (Thu)
by rsidd (subscriber, #2582)
[Link] (6 responses)
In theory, mixing BSD with proprietary code is as much a one-way street as mixing BSD with GPL; but in practice that is not always so. Most proprietary code has well-defined copyright holders who can, and do, choose to relicense their code under the BSD licence (perhaps after recouping their initial investment). The most recent example I can think of is the adding of journalling features to UFS in NetBSD. Commercial companies also contribute regularly to several BSD/MIT/similarly-licensed projects, including X.org, LLVM, etc. Because of the BSD licence, they can continue to use their contributions in their proprietary software; if it were GPL'd, they would find it hard to re-use their own code without GPL-ing everything they have (at best, they would need to be very vigilant in keeping it "clean" and free of others' copyrights).
GPL copyright holders are seldom willing to relicense their work under the BSD licence -- in fact I can't think of any examples.
The point about GPL incompatibility causing problems is perhaps true, but unfortunate. In the XFree86 case, however, there were many other problems even before the licence change, and the split would likely have happened anyway.
Posted Oct 8, 2009 18:04 UTC (Thu)
by johill (subscriber, #25196)
[Link]
I think this is only true inasmuch BSD people don't actually ask, but go whine about the BSD incompatibility instead. I've heard from many people that they'd willingly open their code to the BSDs if they'd bother to ask.
Posted Oct 8, 2009 22:05 UTC (Thu)
by paulj (subscriber, #341)
[Link] (2 responses)
if it were GPL'd, they would find it hard to re-use their own code without
GPL-ing everything they have (at best, they would need to be very vigilant in
keeping it "clean" and free of others' copyrights).
This statement is rubbish. They can easily demand that contributors sign some
kind of agreement to give them whatever rights they need - they can ignore
whatever contributors don't comply. There's no difference here to the BSD
licence - except that with their BSD licenced product they need not even get
to see what other people are doing with their code.
Posted Oct 9, 2009 1:47 UTC (Fri)
by rsidd (subscriber, #2582)
[Link] (1 responses)
Posted Oct 9, 2009 14:34 UTC (Fri)
by paulj (subscriber, #341)
[Link]
If company X puts their code out there and are the 'upstream' for that code
a) Be inclined to co-operate with company X, in which case they'll be happy
b) Not be inclined to co-operate, and so not get their code putback to the
I'm not trying to bash the BSD, but it's baffling to find there are people who
Posted Oct 8, 2009 23:32 UTC (Thu)
by xoddam (guest, #2322)
[Link]
That's because, if they wanted their work to be BSD-licensed, they
Now if something were originally written as a contribution to a
But since the GPL places no restrictions at all on your use of
It's free software, after all!
Posted Oct 12, 2009 6:31 UTC (Mon)
by joib (subscriber, #8541)
[Link]
So NetBSD finally gets a journalled file system in 2009; I think you need to try harder than that if you wish to impress the GPL aficionados. After all, the GPL licensed Linux got a journalled file system in 2001, and by now there is a plethora of them (ext3, ext4, xfs, jfs, probably others as well including btrfs which isn't journalling but accomplishes the same thing by COW). Some of these (xfs, jfs) were contributed by commercial companies, and ones developed from scratch (ext3/4, btrfs) have been developed mostly by people working for commercial companies.
Wrt X, I think you are correct that a GPL licensed (or rather copyleft-like licensed, as GPL wasn't around then) would never have been accepted by the unix vendors at the time. Of course, then the vendors tried to outcompete each others by adding proprietary extensions to their own versions (which a copyleft license, had it been accepted, would have prevented), while letting the common X code stagnate, and the "unix wars" were ultimately won by Microsoft. How's that for karma?
But X.org seems alive now again since the XFree86/X.org fork, and while the license hasn't changed, the culture towards code sharing certainly has changed since the unix wars.
Posted Oct 6, 2009 18:58 UTC (Tue)
by jhoger (guest, #33302)
[Link]
Did they just not respond to that? If so I guess they just need to file a couple other forms now and never do that again...
Not much of a story.
Posted Oct 6, 2009 19:01 UTC (Tue)
by jcm (subscriber, #18262)
[Link] (2 responses)
Posted Oct 6, 2009 19:51 UTC (Tue)
by jhoger (guest, #33302)
[Link] (1 responses)
Maybe the answer will give us some insight into what motivates the "451 Group."
Posted Oct 6, 2009 20:46 UTC (Tue)
by jcm (subscriber, #18262)
[Link]
Posted Oct 6, 2009 19:14 UTC (Tue)
by atai (subscriber, #10977)
[Link]
Posted Oct 6, 2009 19:25 UTC (Tue)
by flewellyn (subscriber, #5047)
[Link]
Posted Oct 7, 2009 6:18 UTC (Wed)
by gowen (guest, #23914)
[Link] (1 responses)
Posted Oct 9, 2009 18:45 UTC (Fri)
by BrucePerens (guest, #2510)
[Link]
This reader is not struggling. It will exactly as much impact as when it
existed.
Open Source Initiative loses corporate status
Whether or not a package made it into Debian and Fedora is a much better real-world test of opensourceitude than whether the license passed OSI, anyway.
Where licensing really matters
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
> combined work is licensed under the GPL".
"You may not impose any further restrictions on the recipients' exercise of the rights granted
herein"
Open Source Initiative loses corporate status
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
(i.e. the place where contributors prefer to get their changes put back to)
then any people who contribute to that code will either:
to give X whatever licence they desire (via a contributor agreement or by
licensing their contribution under the BSD licence, etc)
'upstream' project. In the GPL case, they'll not agree to the CA. In the BSD
case it means they won't make their code available at all.
would argue the BSD licence is encourages modifications to be made
available to the original coder and prevents forking (proprietary and
otherwise). It surely flies in the face of all historical evidence to claim this.
You may just use the code, it's free!
> under the BSD licence -- in fact I can't think of any examples.
would have used the BSD licence in the first place. And indeed they
often do, and indeed the same people and businesses often contribute
their work to various codebases under different licences, co-operating
fully with the upstream maintainers, even making copyright assignments
to the projects' owners where that is the done thing.
GPL-licensed program (for instance, a device driver for the Linux
kernel), and a BSD-licensed project could make good use of that
code, there is nothing stopping someone who wishes to port the
code from asking the original author(s) of the code in question
to relicense it under BSD terms. I know it doesn't happen much
(either the asking nicely or the actual relicensing) but there
is probably a wide range of reasons for both sides of that,
from corporate inertia to OS favouratism and NIH syndrome to
being true copyleft believers.
the code, only on the way you pass it on to others, there is
nothing stopping the would-be porter from doing the technical
work anyway and simply distributing the resulting whole under
the terms of the GPL. If, for whatever bizarre reason, someone
desperately needed to run a GPL device driver in a kernel of
BSD origin and the device driver author was unwilling to
relicense it.
Most proprietary code has well-defined copyright holders who can, and do, choose to relicense their code under the BSD licence (perhaps after recouping their initial investment). The most recent example I can think of is the adding of journalling features to UFS in NetBSD.
GPL-compatible != GPL, BSD-style licenses DESIGNED to be 1-way street
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Open Source Initiative loses corporate status
Has Bruce lost his golden touch?
I haven't been associated with the organization in 10 years. I didn't even do their corporate formation. But they'll find a way to blame me anyway.
Open Source Initiative loses corporate status