(Sorry about the duplicate mail, @Prasoon and @Stefan) And I feel, we should atleast have something like a group chat over the weekend before we open a wiki page for the Vector module. It will help us get on the same page, and discuss how Prasoon and I can make our timelines gel together. Once we are clear with the technical details, we can start discussing how Prasoon is going to implement his functionality and how I will put it into my code. As of now, our proposals are a good indicator of the kind of API we are talking about. About the things that are common, I feel a real-time communication will help before finalizing.
On Fri, May 31, 2013 at 10:13 AM, Sachin Joglekar <srjoglekar...@gmail.com>wrote: > I went through Prasoon's proposal in detail, and I think it won't be too > difficult coordinating our work so that his code fits well into mine. > However, I have some queries that I would like cleared from Prasoon and > Stefan to ensure we are all on the same page- > > 1.) Why define every Vector's coordinate variables differently? What > advantage does this method have over defining the variables with respect to > a RefFrame? If we are going to define the time variable with respect to a > frame, I felt it was intuitive to do so for the other variables also. This > will ensure uniformity in the calculation of different quantities like > divergence, curl etc. of vectors defined in a CoordSystem quite easily( > vector.curl(RefFrame)), and using the something like the current > 'vector.express(ReferenceFrame)' system, the Vector could be defined with > respect to variables of different RefFrames. > Say I have two fields (subclassed from Vector), F1 and F2 such that I > define their vars differently. If I add them vectorially, and try to find > out the divergence of the resultant, how will it be evaluated? Or will it > give an error? It would be helpful if you could provide me with a sample > session for this. > > 2.) I like Prasoon's idea of using a class like ParamRegion to facilitate > vector integrals. I will most probably be using that framework in the > application of Helmholtz theorem to calculate fields using their curl, > divergence and boundary conditions (@Stefan, would like your comments on > this). For this reason, have you decided on a method by which you will > represent the entire 3D flat space? This may also be useful for usage of > DiracDelta function in vector integration. > > @Stefan, I am still working on how I am going to represent ChargeDensities > in space, and I will take a few days to get the idea totally cleared. Till > that time, to illustrate the problem you have suggested (Q1-5), I would > like to use a special case of a scalar field, that is the ScalarPotential > of an ElectricField. I can extend the idea to Charge densities once I am > clear myself. > I am sorry for the delay in the setting up of the wiki page, I had a few > technical problems with the internet at home. > > > -- 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 http://groups.google.com/group/sympy?hl=en-US. For more options, visit https://groups.google.com/groups/opt_out.