Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
dsolve: expm/jordan solver #286
Good news, thank you for reply.
Your code was sitting around in the pr queue for years without review... So, good luck.
Are you about this project? Yeah, but you know my history, at least partially. That wasn't my idea to fork the SymPy, but now it is what it is. For good or bad - now there is a different project with different goals and priorities.
Ok, now I think it's ready to be merged (to replace old horrible LODE solver). Rewriting to sin/cos was reverted back, perhaps I'll add this later.
Solutions sometimes seems (with and without sin/cos rewriting) to be more complex (see tests changes), but I believe it worth, as new solver is much more simple.
Please tell me if you would like to review changes.
Jun 20, 2016
referenced this pull request
Mar 1, 2018
I think the form of answer is coming different.
Yes! Math is correct I'am just wondered why
dsolve can be very sensitive to the form of solutions. That's why the dsolve tests were all written with checkodesol, to verify that the solution being tested is correct (it seems this practice was not extended to the systems tests). Generally, if the output form changes a little bit, you should just update the test (after verifying that the expression is still mathematically correct of course), unless it is significantly worse for some reason.
In the case of