I was also discussing this Nishant in real life, SymEngine **is not** a C++ clone of SymPy, rather SymEngine aims to be very fast core for SymPy and hopefully other Computer Algebra Systems. A major reason for the success of SymPy is that it is written in Python and which makes it is easy for someone to understand and extend even if she/he is not an expert programmer. The issue with python is that it is slow. Programs also follow something like Pareto principle, 10% of the code takes 90% of the time and it makes sense to **only** optimize that 10% of code. You should move that 10% of code in SymEngine and leave the rest in SymPy. Another test of determining if some code belongs to SymEngine will be asking the following questions:
* Is this code settled? In the sense that there are no major bug reports for code from a long time and we aren't expecting any api or structural changes. * Does this code do a lot of heavy lifting? * Is there are lot of code which depends on this part of the code? If the answer is yes for all three questions only then it is strong candidate for SymEngine. Though I'm involved in the efforts with SymEngine, hence I request Ondrej to express his thoughts on my proposed test. As far as imageset is concerned, though the solveset depends on it, it doesn't do a lot of heavy lifting and it mainly serves as a representation for the solutions. Unless you are planning to implement solveset in SymEngine, having imageset there won't help much. On 19 March 2016 at 21:09, Akash Trehan <[email protected]> wrote: > > > On Friday, 26 February 2016 22:58:07 UTC+5:30, Harsh Gupta wrote: >> >> >Hi Akash, >> >> >1.) Sets module is clearly is a dependence for the solveset module, the >> > output of solveset of set and the module also make heavy use of a lot of >> > >functionalities of Sets including intersection, union and especially >> > imageset. SymPy's set are **very** flexible and we aim to have the >> > capability to > > I was looking at the code of `imageset` in sympy to see how it computer the > function sent to it. So it further calls a function `_eval_imageset` which > itself imports `solveset` for solving some expressions. I wanted to ask if > some functions of solveset would have to be implemented before starting with > the implementation of `ImageSet` in SymEngine? Can you provide me some > references for this? >> >> >represent every set possible. My limited knowledge in C++ tells me that >> > it might be hard to achieve the flexibility set module in it, you should >> > >discuss the details with Ondrej and other people involved with Symengine. >> >> >2.) We had to go through a lot of discussion to designing the right input >> > and output API, we have tried to document those decisions here >> > >https://github.com/sympy/sympy/blob/master/doc/src/modules/solvers/solveset.rst >> >> >BTW I don't think solveset is mature enough to be ported to symengine, >> > but porting the sets module to symengine definitely seems like a good >> > >idea. >> >> On 26 February 2016 at 13:53, Akash Trehan <[email protected]> wrote: >>> >>> Hi Harsh >>> >>> I'm Akash Trehan. >>> I'm proposing to implement the solveset module in Symengine this GSoC. I >>> found out that you started solveset in sympy. I wanted some help on certain >>> things, >>> 1) How much of the Sets module is being used for solveset? >>> 2) What were the problems faces in starting a new module? >>> >>> Thanks, >>> Akash >> >> >> >> >> -- >> Harsh >> Sent from a GNU/Linux > > -- > 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 [email protected]. > To post to this group, send email to [email protected]. > 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/31bf572d-d301-40da-93b7-5831f9552566%40googlegroups.com. > > For more options, visit https://groups.google.com/d/optout. -- Harsh Sent from a GNU/Linux -- 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 [email protected]. To post to this group, send email to [email protected]. 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/CADN8iuqny26QY8bn9b6UmgBdCtU8F-W1wr0n1c_Lj6YE%3Deji7g%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
