[Pipet Devel] License

J.W. Bizzaro bizzaro at geoserve.net
Tue Apr 18 16:52:07 EDT 2000

Jean-Marc Valin wrote:
> This is not the part of the Linux license I was referring to (and unfortunatly,
> I cannot find it now).

Well, let me know if you come across it.

> It Linux was using the "pure GPL", then it would mean
> that every part of it would have to be GPL compatible.

You need to distinguish between the Linux kernel and the GNU/Linux OS.  Every
part of the kernel needs to be GPL, true.  But, every part of the OS does NOT
have to be GPL.  This is because the applications that run in the OS are not
'linking' to the kernel by the GPL's definition of 'linking'.  This is what
Linus means when he says that other applications are not a 'derivative work'
of the kernel.

> Also, the (L)GPL states
> that all the code, including what you link (in the sense of library linking) to,
> has to be GPL-compatible.

The ordinary GPL states that, not the LGPL.  That's the whole reason for the
LGPL: to allow linking.

> If Linux was "pure GPL" then it would not allow you to
> "link" Linux with a proprietary driver. I know Linus made an explicit exception
> that allow that "softens" the GPL and allow you to use proprietary drivers, as
> long as they are a module,

Right, they are not 'incorporated' into the kernel.

> and (I'm not sure about this one) the driver doesn't
> require a change in the kernel. BTW, the reason Debian didn't ship with KDE
> until Qt was "open-sourced" was that KDE didn't comply with the GPL by linking
> to Qt.

Actually, if Debian shipped KDE, they would have to ship the commercial Qt
too.  And it was *Qt* that violated the Debian definition of free software
;-)  But I do recall a number of people questioning the legality of linking
the GPL'd KDE with the commercial Qt.  It goes to show that the definition of
'linking' is not very clear in the (L)GPL.

> Anyway, I think we all agree on the important: allow people to develop
> proprietary Overflow modules and distribute them as .so only.

So far, you, Jarl and I have stated we would like some modified version of the
LGPL.  But we haven't heard from Brad and Dominic, both of whom contributed
code.  Every author needs to agree on this.

                      |           J.W. Bizzaro           |
                      |                                  |
                      | http://bioinformatics.org/~jeff/ |
                      |                                  |
                      |        BIOINFORMATICS.ORG        |
                      |           The Open Lab           |
                      |                                  |
                      |    http://bioinformatics.org/    |

More information about the Pipet-Devel mailing list