license choice
Ondrej Certik
> sorry, you're absolutely correct. I'm a long term debian user and I
> tend to say one or the other as a short hand which is a terrible habit
> because they are completely different things.... I'll check out the
> differences between the BSD and GPL (and subversions) and all that red
> tape required to coerce people into being reasonably well mannered,and
> decide when I come to release it. Basically I want a very strong
> license so that anyone who uses it has to be free and respectful about
> it...I don't want people putting a GUI on my masters project and then
> trying to claim it is all theirs and sell it or something, I also
> don't want some cheeky company taking my ideas and patenting them as
> if they were their own (and then possibly suing me for using 'their'
> idea some way down the line -- apparently this is happening)
Feel free to discuss any licensing issues or concerns here.
Basically there are very good arguments for GPL (you stated some) and
there are also very good arguments for BSD, see for example:
http://www.scipy.org/License_Compatibility
The only important thing from my side is that all SymPy developers and
users can agree on some common interests, so that we can produce
something useful together, instead each of us working on our own thing
only.
I can also offer my own personal view (but other sympy developers
doesn't have to agree fully with me here): I generally prefer less
restrictions than more restricitons, so that the code I write can be
used by more people any way they like. I want them to acknowledge that
though and both BSD and GPL require that.
The thing I don't like on GPL is that it forces the user to also use a
GPL. So technically if you use GPL2, the user cannot license his
program as GPL3, because they are not compatible. You can use GPL3,
but then the user cannot use GPL2 programs. You can use GPL2 or later,
but then you are not stating the precise terms to copy/redistribute
your program, but rather giving the future of your program into the
hands of FSF, i.e. whatever license they create in the future could be
applied to your code as well. But see also below.
> it...I don't want people putting a GUI on my masters project and then
> trying to claim it is all theirs and sell it or something, I also
I think both BSD and GPL require them to give credit to you. Otherwise
I don't mind anyone building on top of my code and improving it. See
below about contributing back.
> don't want some cheeky company taking my ideas and patenting them as
> if they were their own (and then possibly suing me for using 'their'
> idea some way down the line -- apparently this is happening)
I am not a lawyer, but it seems to me it is not possible to patent a
code that someone else has written and released as BSD or GPL, is it?
There is also a question which license is the best to encourage people
to contribute the code back to the project and to encourage people to
work together on one thing, instead of forking things and doing things
on their own. Linus says it's GPL2, but I think more important than
the license itself is the way the project is handled and how many
people use it. I think the motivation to work on one thing doesn't
stem from the license, but rather because it is more effective for
everyone involved.
Nevertheless, I am not 100% convinced either way, I can see good
arguments for both ways.
Ondrej
Robert Cimrman
IMHO the things are different depending on what the purpose of your code
is. In the case of a fundamental library (in the sense of basic
research), for example, a general symbolic engine, a finite element (FE)
engine, a plotting library, the less restrictive license the better -
here one wants a large user base, lots of developers/contributors, and
thus also a mutual _compatibility_ of projects built on top of that
library (free or not). This is exactly the case of sympy, by the way. On
the other hand, when developing a special stuff (e.g. applying a general
FE code to some particular deep problem) it is not so clear to me, all
depends on what the author wants of her code to become.
Just a historical remark: I have initially released SfePy (a FE code)
under GPL, but following the discussions on
Numpy/SciPy/Matplotlib/IPython mailing lists I have changed the license
to BSD. Then Ondrej (and others) joined and things started to accelerate
:) This may also put some weight on the balances - it seems to me that
in the world of Python scientific libraries the BSD license is more
usual (just a hunch, I have no numbers.)
r.
Ondrej Certik
BTW, SymPy was also GPL originally, but after discussing it in the issue:
http://code.google.com/p/sympy/issues/detail?id=138
and getting a permission from all developers, we changed it to BSD too.
Ondrej
physnut
The concerns I have about abuse of my code is more related to things
like this:
http://news.zdnet.co.uk/software/0,1000000121,39289505,00.htm
(article about microsoft trying to sue 'Linux' - whoever they think
that is!, most probably rich vendors, even though every single Micro
$oft 'innovation' is either bought up or stolen from open source
code...)
There is clearly much going on in the world of software patenting and
not wanting to get stung I prefer to err on the side of caution...
I would prefer people to be able to play with my code however they
want, that benefits everyone (especially me if the post it on the net
and give me credit...). It would be better if I didn't have to lock
my door and remember my keys but I just can't trust everyone! If BSD
license protects me from legal nasties it sounds good enough for me.
I'm quite happy about people developing what I have done into
something better/customized/evolved and would love to see things
flourish like that...I guess if I see some company making money from
my code i'll just have to ask them if they would be nice enough to
donate some money to keep me out of corporate work and in science.
On Jun 23, 11:33 am, "Ondrej Certik" <ond...@certik.cz> wrote:
Andrew Straw
> I am not a lawyer, but it seems to me it is not possible to patent a
> code that someone else has written and released as BSD or GPL, is it?
>
IANAL either, but I don't think this is correct -- patents and
copyrights (licenses are implemented using copyright law) are
more-or-less orthogonal. Just because something is patented doesn't mean
it cannot be released under an open source license. You can have and
distribute the source code, but you might need to seek a patent license
to actually use the code. One example of this I know about is the
marching cubes algorithm in VTK (although that patent has now expired).
GPLv3 may explicitly deal with this issue, but I think GPLv2, BSD and
MIT licenses don't.
-Andrew
Robert Kern
>
> Ondrej Certik wrote:
>
>> I am not a lawyer, but it seems to me it is not possible to patent a
>> code that someone else has written and released as BSD or GPL, is it?
>
> IANAL either, but I don't think this is correct -- patents and
> copyrights (licenses are implemented using copyright law) are
> more-or-less orthogonal. Just because something is patented doesn't mean
> it cannot be released under an open source license. You can have and
> distribute the source code, but you might need to seek a patent license
> to actually use the code. One example of this I know about is the
> marching cubes algorithm in VTK (although that patent has now expired).
I believe the important part about Ondrej's hypothetical is that the
code was written and published under an open source license *before*
someone else tries to patent it. Under essentially all patent regimes,
prior publication of the invention by someone else is grounds for
preventing or invalidating a patent (with a lot of effort). The VTK
situation is different; the VTK code was written after the algorithm
was patented.
--
Robert Kern
"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
-- Umberto Eco
Ondrej Certik
>
> It's an interesting point you raise about the GPL license terms...
>
> The concerns I have about abuse of my code is more related to things
> like this:
> http://news.zdnet.co.uk/software/0,1000000121,39289505,00.htm
> (article about microsoft trying to sue 'Linux' - whoever they think
> that is!, most probably rich vendors, even though every single Micro
> $oft 'innovation' is either bought up or stolen from open source
> code...)
Well, Microsoft is just a company with good and bad people. They are
for example funding the windows port of Sage, i.e. supporting an
opensource
alternative to Maple/Mathematica, so not everything that MS does is bad.
>
> There is clearly much going on in the world of software patenting and
> not wanting to get stung I prefer to err on the side of caution...
> I would prefer people to be able to play with my code however they
> want, that benefits everyone (especially me if the post it on the net
> and give me credit...). It would be better if I didn't have to lock
> my door and remember my keys but I just can't trust everyone! If BSD
> license protects me from legal nasties it sounds good enough for me.
I think it does, see below.
> I'm quite happy about people developing what I have done into
> something better/customized/evolved and would love to see things
> flourish like that...I guess if I see some company making money from
> my code i'll just have to ask them if they would be nice enough to
> donate some money to keep me out of corporate work and in science.
My own opinion is that as long as there is a good opensource operating
system with enough programs so that I can do my everyday tasks with
opensource (which there is, e.g. Debian and other distributions) and
as long as there are people contributing back (this is important), I
don't mind also companies inventing clever ways to make money, even
with my own code --- they have the same opportunities as I have, e.g.
I can do the same as them and make money too. Fair enough.
On Tue, Jun 24, 2008 at 12:33 AM, Robert Kern <rober...@gmail.com> wrote:
>
> On Mon, Jun 23, 2008 at 17:20, Andrew Straw <ast...@caltech.edu> wrote:
>>
>> Ondrej Certik wrote:
>>
>>> I am not a lawyer, but it seems to me it is not possible to patent a
>>> code that someone else has written and released as BSD or GPL, is it?
>>
>> IANAL either, but I don't think this is correct -- patents and
>> copyrights (licenses are implemented using copyright law) are
>> more-or-less orthogonal. Just because something is patented doesn't mean
>> it cannot be released under an open source license. You can have and
>> distribute the source code, but you might need to seek a patent license
>> to actually use the code. One example of this I know about is the
>> marching cubes algorithm in VTK (although that patent has now expired).
>
> I believe the important part about Ondrej's hypothetical is that the
> code was written and published under an open source license *before*
> someone else tries to patent it. Under essentially all patent regimes,
> prior publication of the invention by someone else is grounds for
> preventing or invalidating a patent (with a lot of effort). The VTK
> situation is different; the VTK code was written after the algorithm
> was patented.
Yes, I meant publishing the code before someone else patents it.
Thanks for the clarification.
Ondrej