Thanks Bill. I am at the stage of convincing Astronomers to use Pharo to do 
their soft, but this is not easy

Cheers,
Alexandre


On 5 Nov 2010, at 02:57, Schwab,Wilhelm K wrote:

> Alexandre,
> 
> I don't have any direct experience with astronomy, but we no doubt share some 
> requirements.  I started purging Numerical Recipes from my arsenal long ago, 
> in part over licensing and mostly because so much of the code suffers from 
> Fortran-confusion-syndrome.  What do I mean by that?  Referring to their C 
> book, look at their theoretical development of the Fourier Transform (kinda 
> nicely done), its very natural use of zero based offsets in the theory; note 
> the unit offset code.  Take in the early section in which they attempt to 
> "free" the reader "from the zero based thinking that C encourages, but does 
> not require."  Who did they think they were going to fool with that nonsense? 
>  They had Fortran code and did not want to change it - nothing more.
> 
> GSL is not too much better, but they have a working Levenberg-Marquardt fit.  
> Squeak's lack of callbacks haunts us here.  So far, I have gotten it to run 
> by writing C functions for the model functions of interest to me.  GSL seems 
> to almost randomly require or ignore (often at the wrong times) use of their 
> vectors and matrices.  They have good Fourier and Wavelet transforms.  
> Squeak/Pharo do not have a Double Array class.  I have gotten around that, 
> but so far at some cost to robustness; I have described many double* 
> arguments void* to allow FFI to make the calls.
> 
> The biggest problem with GSL is the license.  Do you have a strategy for 
> releasing code that links to it?  I am perfectly willing to have others use 
> my code.  I would prefer that they be able to make money if they wish.  I 
> will not be told that I cannot make money from it.  A recent thread contains 
> a suggestion for using a "GPL-compatible" license.  Any ideas?
> 
> Didier Bessert has been kind enough to release his code under MIT.  To be 
> blunt, I didn't buy the book originally because I was not enamored with the 
> idea of doing numerics in Smalltalk (most/some of it belongs in C for speed), 
> and his selection of topics is not a very good match to my interests.  
> However, it might be worth considering as a way to avoid GPL code.
> 
> For plotting, I have been using PLplot.  You will find many astronomers 
> hanging around PLplot.
> 
> Bill
> 
> 
> 
> ________________________________________
> From: [email protected] 
> [[email protected]] On Behalf Of Alexandre Bergel 
> [[email protected]]
> Sent: Thursday, November 04, 2010 3:14 PM
> To: Pharo Development
> Subject: [Pharo-project] Pharo and Astronomy?
> 
> Hi!
> 
> I am looking for any experience in using Pharo to develop software for 
> Astronomy. Especially regarding interoperability with Python, using 
> Mathematical libraries (www.nr.com, www.gnu.org/software/gsl, 
> matplotlib.sourceforge.net/gallery.html)
> 
> Cheers,
> Alexandre
> --
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel  http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
> 
> 
> 
> 
> 
> 
> 

-- 
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






Reply via email to