> Frankly, I think that there is a very urgent need to implement a working > tensor module enabling the usage of the Einstein summation convention. I had > a look at sympy.tensor, but it looks like that module is still far away from > working (unless I didn't figure out correctly how it works).
sympy.tensor is meant to be used for code generation. It does it well, it is not used a lot, it lacks good integration with rest of sympy. It is not meant to represent nD-arrays of symbolic expressions or tensors. Yes, the name of the module is not great. > Yesterday evening I wrote a draft about a FourVector object and a Dirac > Spinor object. The underlying engine is based on for-loops for the > FourVector and on matrices for the Dirac Spinor. This is somewhat vague. From the description I gather that these are not geometric entities but just the representation of a FourVector or a Dirac Spinor on some basis. If this is to be useful, one would need a natural way to represent them independently of a basis. Both `diffgeom` and `physics.quantum` achieve this through very different ways. > I think that now it would be more appropriate to disregard Geometric > Algebra, for some reasons: > > [...] There is actually a geometrical algebra module in sympy. (and there is a rewrite in progress) Regretfully, it is not used a lot at the moment and I do not know much about it. In any case, given the differences, a small repetition of code in the core of a module is not a problem if it keeps two modules well separated. So work on tensors and on GA can be done independently. > > I suggest the following steps to extend the physics module: > > write some tensor code working with Einstein summation convention. > implement objects with their transformation properties under the Lorentz > group (possibly Poincarè too). > make sure that the end-user interface is different from the internal engine > using tensors. > once this framework is finished, start thinking on how to convert a > tensor-based engine to a GA-based engine. This is too vague to comment, but if you give details (on the wiki for instance) we would be happy to help. > By the way, what's the exact status of sympy.tensor? Is it going to work? I > can't wait for it to finish, I think I'll write my own temporary workaround > to start doing some physics. As I said sympy.tensor is not meant for geometrical tensor objects or for nD-arrays containing symbolical expressions. -- 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.