On 03-18-2011, at 11:49 PM, Aaron S. Meurer wrote:

> On Mar 18, 2011, at 9:36 PM, Tim Lahey wrote:
> 
>> Hi,
>> 
>> I've been thinking about applying for GSoC and working on a Cython core. I'm 
>> assuming that it would be an optional thing? So, there would be two cores, 
>> an optimized one that's done in Cython and a pure Python one for when Cython 
>> isn't available.
>> 
>> Is the idea to update and clean up sympyx and properly integrate it into 
>> sympy?
> 
> Well, the hope is that we can use things like @cythonized decorators in the 
> core, and make it just work either way (either it uses pure Python, or cython 
> if the cython has been made).  This is how the cython in the polys works 
> right now.  We will have to look at it to see if this will work for the core. 
>  Anyway, it would definitely be optional no matter how it is done, because we 
> want people to be able to use SymPy using pure Python as the only dependency.

So I guess looking at the polys branch (which is where exactly?) is a good idea 
to see how they do it.

> 
> Also, it was my understanding that the core needs a lot of cleanup before it 
> can be properly cythonized.  In particular, the old assumptions have to be 
> completely removed and replaced with the new ones.  Others know more about 
> the state of the core and could probably answer your questions better, though.

I kind of guessed that the core would need to be cleaned up. Even if the work 
doesn't get finished as part of GSoC, at least things would be further along. 
So, if all I did was clean up the core and cythonize some of the core, things 
would be better.  

> 
>> 
>> On a development note, is there a good way to have a separate python and 
>> packages (e.g., a sympy fork and cython) so I don't mess up my system 
>> python? I'm working on OS X, but I could install a Linux VM in Virtualbox 
>> (if so, which Linux?).
>> 
> 
> If you are working on just SymPy, you can just work straight from the git 
> repository.  There is no need to install it.  
> 
> If you want to create a virtual environment with Python, you can use 
> virtualenv.  Maybe someone else here can explain how to use it better than I 
> can (I am still new to it), but I think you basically type "virtualenv 
> directory" and then install things using the virtual python installed in that 
> directory (I think I might be missing a step or two here, so look at the docs 
> before trying).
> 
> Also, an alternative to the system Python on Mac OS X is to use fink.  That 
> is what I do.  Everything is installed in /sw/, and if you screw up the 
> installation, you can completely trash /sw/ and start again if you want (it 
> is completely independent of the system stuff).


I'd rather not install Fink since I use Homebrew instead and I don't want 
duplication of system packages (except Python), so I'll look at qsnake.

I realize that this is a non-trivial task, but the last time someone looked at 
this was June 2009 based upon Ondrej's repository.

Someone has already done work on a new assumption system haven't they? I'm 
guessing it just isn't complete or is it complete and just not integrated?

Thanks,

Tim.

---
Tim Lahey
PhD Candidate, Systems Design Engineering
University of Waterloo
http://about.me/tjlahey


-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To post to this group, send email to sympy@googlegroups.com.
To unsubscribe from this group, send email to 
sympy+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/sympy?hl=en.

Reply via email to