Well, regardless, it sounds like the new module can't quite yet replicate all the features of the old one. So just deprecate the old module.
By the way, are you planning on the new module being included in the standard namespace (from sympy import *)? If so, you might want to choose better names for some things (e.g., E will clash with exp(1)). You might consider it regardless, actually. It looks like the old module is not imported by default. So I would recommend choosing names that seem best, without concern over name clashes with the old module. Things like NormalRV might sound OK now, but it's a name that you will have to live with well after the old module is gone. If someone using the old module wants to use the new module, he will either have to handle the name clashing himself, or switch his old code over to the new code (we want to encourage the latter). You'll have to have different file names (this is why we have "poly" by the way, because Mateusz wrote it when "polynomial" already existed). But I see you've already handled this. Finally, if the integrals aren't powerful enough, you should keep up with Tom's work (look at his most recent pull request). Definite integration in SymPy is going to get a power boost really soon. Aaron Meurer On Fri, Jul 22, 2011 at 7:54 PM, Matthew Rocklin <[email protected]> wrote: > Most of the functionality exists in the new code but how you would write it > is different > old: > X.mean > X.stddev > X.pdf(z) > X.cdf(z) > new: > E(X) > std(X) > Density(X) > P(X<z) > The new code solves the problems with integrals though rather than having > the solution hard-coded in. This causes the new code to take longer and fail > more easily. Also, it doesn't (yet) provide sampling while the current code > does. (btw, can anyone point me to a good general sampling reference?) > Almost all functionality is replicated in the new version and easily > re-expressed. Mostly my concern is for whoever out there might be using the > old code. When they update their SymPy suddenly their code will break. > On Fri, Jul 22, 2011 at 3:38 PM, Aaron Meurer <[email protected]> wrote: >> >> How easy is it to replicate the old code using your new code? >> >> Aaron Meurer >> >> On Fri, Jul 22, 2011 at 7:33 AM, Matthew Rocklin <[email protected]> >> wrote: >> > I'm rewriting the statistics module and would like to take over some of >> > the >> > names in the previous version. I'm wondering who would mind or what our >> > policy is on backwards compatibility. >> > Previously there was a distributions.py file with two Distribution >> > objects, >> > Normal and Uniform. I'd like to take over these two words in the >> > namespace >> > to create my own Normal and Uniform random variable objects. This would >> > break anyone's code who uses them currently. Does anyone use them >> > currently? >> > The way I see it there are two decent options >> > 1) Take over these names, delete the old distributions.py file (all old >> > functionality exists in the new version, just with new syntax) >> > 2) Leave distributions.py intact, use names like NormalRV and >> > ExponentialRV >> > for all of my random variable creation functions in my code. >> > -Matt >> > >> > -- >> > You received this message because you are subscribed to the Google >> > Groups >> > "sympy" group. >> > To post to this group, send email to [email protected]. >> > To unsubscribe from this group, send email to >> > [email protected]. >> > For more options, visit this group at >> > http://groups.google.com/group/sympy?hl=en. >> > >> >> -- >> You received this message because you are subscribed to the Google Groups >> "sympy" group. >> To post to this group, send email to [email protected]. >> To unsubscribe from this group, send email to >> [email protected]. >> For more options, visit this group at >> http://groups.google.com/group/sympy?hl=en. >> > > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/sympy?hl=en. > -- You received this message because you are subscribed to the Google Groups "sympy" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/sympy?hl=en.
