Hi Imam, On Mon, Feb 8, 2016 at 9:08 AM, Imam Tashdid ul Alam <imam.a...@gmail.com> wrote: > Hi Ondrej, > > Thanks for your reply. An interesting discussion > (http://opensource.stackexchange.com/questions/304/can-i-take-bsd-licensed-code-and-distribute-it-under-gpl) > also pointed me to the linux kernel case. I looked at the source tree on > Github. The top-level COPYING says nothing about the BSD license issue but > files inside the tree contain it. > > So I am back to square one. Seems like the right thing to do is to mark the > code that I take from sympy with the BSD license. Only if I modify it enough > should I be able to call it mine and dispense with the BSD license. Reminds > me of Sorites paradox (https://en.wikipedia.org/wiki/Sorites_paradox). Who > knew contributing to libre software has its own riddles.
I think it's actually pretty simple. In the case of SymPy, people who contributed the individual patches (see the git history) own the copyright to them, and they give SymPy a copyright license to use the code under the SymPy's BSD license. That BSD license allows you to take the code and distribute under your own terms, say, under the GPL license. You copy the SymPy's BSD license somewhere in the code or documentation. So people who you distribute the code to must follow the GPL license, and if those people distribute the code further, they need to follow GPL, and they still need to keep the copy of the SymPy's BSD license somewhere in the documentation or code. Regarding when you can call it yours, it depends on what you mean by "yours". If by "yours" you mean when you'll have the rights to license the code to the people you distribute it to under your own terms, then you have this right already (as given by the BSD license), so it is already "yours". If by "yours" you mean when you'll own the copyright of the code, then the answer is never, unless you get the copyright transfer from all the authors of SymPy. The rule of thumb is that you should never transfer the copyright of code that you contribute to open source, you should keep it. (You must give the project a copyright license of course to use it under their license.) One exception is the FSF, e.g. if you contribute a patch to gcc, you transfer your copyright to FSF, and they give back to you a copyright license to use the code --- but then I think legally you can't take the code that you contributed, and give somebody else a copyright license, since you don't own the copyright to your own code. I could be wrong on any of the above, but that's my understanding how licensing works. > > But since I have your attention for the moment, do you have any suggestions > about rewriting the core? Is there a set of design choices you wish someone > would explore? I like it that the expression-tree is immutable. Would you I have some ideas regarding speed, and I essentially implemented most of them into SymEngine (https://github.com/symengine/symengine), we are still exploring a few more ideas. But that design proved very fast already. > recommend following a lens-based approach > (https://www.schoolofhaskell.com/school/to-infinity-and-beyond/pick-of-the-week/a-little-lens-starter-tutorial) > to manipulate it? My main goal is to make both the expression-tree and the I quickly read through the tutorial, but I don't understand what the "lense" are good for, I would need to spend more time with it. So I don't have any opinion on that. > allowed transformations serializable so that people can share math bits over > the internet. Pointers and caveat about potential dangers will be very much > appreciated. Yes, being able to serialize an expression is essential. In SymPy, you can print it to a file and load it, or you can use Python's pickle, which is a lot faster, but then the file is not human readable. > > Again, thanks for sympy. I wish it was more popular among physicists. As an > aside, I am also currently enjoying your ongoing theoretical physics book. > Forgive me if I am mistaken and that's not you. If you are, I will be very If you are talking about this book: http://www.theoretical-physics.net/, then yes, that's me. I am glad you found it useful. > happy to contribute to the book. And of course, I intend to contribute to > sympy if I can as well. Definitely, looking forward to both. Ondrej > > > > > On Monday, February 8, 2016 at 5:44:17 PM UTC+11, Ondřej Čertík wrote: >> >> Hi Imam, >> >> On Sun, Feb 7, 2016 at 11:28 PM, Imam Tashdid ul Alam >> <imam...@gmail.com> wrote: >> > Hello all, >> > >> > Recent sympy user here. Need advice on license issues. I am thinking of >> > writing a library with much in common with sympy but the plan is nowhere >> > near complete at the moment. >> > What I would like to know more about is the implications of sympy >> > license. I >> > am thinking of indiscriminately copying the sympy codebase when and >> > where I >> > need it. If I read the license correctly, that should be fine if I >> > license >> > my code under BSD as well. But the problem is, I think I like GPL more >> > for >> > my purpose. I do not understand the subtlety involved in deciding >> > between >> > GPLv2 and GPLv3, but let's say I go for GLPv3. >> > >> > Wikipedia >> > >> > (https://en.wikipedia.org/wiki/GNU_General_Public_License#Compatibility_and_multi-licensing) >> > seems to suggest that it is okay for me to release my code under GPLv3 >> > (since sympy is BSD-3), but I would appreciate guidance regarding where >> > not >> > to go overboard with this. For the sake of argument, can I just fork >> > sympy >> > and just change the license to GPLv3? Does not seem quite right to me. >> >> Yes, my understanding is that you can do that, i.e. that you can take >> parts of SymPy, optionally add your own stuff, package it, and use any >> license you want to distribute it. Let's say you only sell binaries >> using some commercial license. Then all you have to do is to put the >> SymPy's BSD license somewhere in documentation. If you want to use GPL >> and distribute the sources, then you can do that as well (as long as >> you copy the BSD license somewhere), at least that is my >> understanding. >> >> You can look at the Linux kernel how they do that, since the kernel as >> a whole is GPL2 licensed, but they also include some code that was >> originally BSD licensed. >> >> > >> > Also, acknowledgements. Let's say I appropriate a sizable part of sympy >> > (as >> > I intend to). And maybe I have a file called AUTHORS. I think BSD-3 says >> > that I do not put the names in AUTHORS in the sympy tree in it. I will >> > of >> > course acknowledge the heavy reliance on sympy in the documentation and >> > elsewhere, but is there anything more I should do to give credit to >> > individuals? >> >> You only have to copy the BSD license. You do not need to copy the AUTHORS >> file. >> >> > >> > End of questions. Just a bit more on what I am trying to do just in >> > case. I >> > am not currently interested in all the sympy modules but just the core. >> > My >> > primary goal I think is to experiment with a more simplified >> > representation >> > of expressions at the expense of usability so that I can provide more >> > control over the generation of LaTeX code. >> > >> > Thanks in advance and thanks for this awesome program. >> >> I am glad you found SymPy useful. If you implement some new things >> that you think would be useful for SymPy itself, we are always >> interested in getting a pull request. >> >> Ondrej > > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sympy+unsubscr...@googlegroups.com. > To post to this group, send email to sympy@googlegroups.com. > Visit this group at https://groups.google.com/group/sympy. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sympy/a6b05082-fa66-4733-bbec-09eb78d1eddf%40googlegroups.com. > > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to sympy+unsubscr...@googlegroups.com. To post to this group, send email to sympy@googlegroups.com. Visit this group at https://groups.google.com/group/sympy. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CADDwiVB0h2O0pQpm_rd-OMCubt-yPDuTTJBMyLGv9WgP%3DqYyjA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
