commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Phil Steitz <>
Subject Re: [math][all] Mentoring for Google's Summer of Code program
Date Sat, 04 Jun 2005 16:22:08 GMT

This is looking good.  Welcome to Jakarta Commons!  

I think the core things below will be a good start and I am sure we
can accomodate the other things as well (possibly moving out of j-c
when we get there).  I would personally pull the special functions
optimization up into core and prioritize this, since this is an
enhancement to what we already have and will boost performance of the
probablity and stats applications that we use them for.  More special
functions would also be welcome.

Given the general positive response, and others' willingness to help
mentor, I will post a general description to the SoC Wiki and/or work
with Ryan to get a full proposal submitted  so we can try to get this
onto the list of potential ASF-sponsored projects (I have inquiries
out to find out what the right next steps are).

One more thing.  There is a lot of work to do on this.  I am sure we
could use more volunteers - mentors, contributors, students -  so
anyone in the community who may be interested, please feel free to
join us!

> Core:
> - polynomial & rational function interpolation
> - univariate minimization
> - univariate integration
> - performance tuning for matrix computation and more decomposition
> methods: mainly adapt code from other sources (Colt) assuming there is
> no license issue
> Additional (subject to discussion):
> - generic (derivative-free) multidimensional optimization
> - performance tuning for special functions
> - FFT if it's not been worked on (I saw it on the wish list)
> - initial value ODE package (RK & Multi-step)
> Elsewhere (maybe merged back if there is enough interest):
> - B-spline package (a generalisation of cubic spline which can handle
> arbitrary order)
> - optimization package
> - PDE solver package
> I'm going away for a few days due to exams but will be following the
> discussion whenever possible.
> Thanks,
> Ryan

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message