Re: Top 10 IDL Requests

[landsman(at)my-deja.com]

Here are my wishes for math updates in IDL:

1.  My longtime wish is to update the "Numerical Recipes" which call a
user-specified function (e.g. QROMB, LMFIT) and currently require the
user function to accept a scalar argument and return a scalar value.
This goes against the spirit of IDL and makes the routines unacceptably
slow for most real-life applications.     (I wonder if it is a problem
with the licensing agreement with Numerical Recipes that prevents
modifying the routines to make them more IDL compatible.)


2.  Median filtering is a non-linear operation which cannot be easily
vectorized, so I would like to see more capabilities in the MEDIAN
function:

  1.  Allow separate specification of the X and Y filter size (e.g. im1 =
median(im,[15,3]), which would be useful, for example, in removing bad
columns.  2.  For 2 or 3 dimenional images, allow MEDIAN to work over a
specified dimension, e.g. for removing cosmic rays from a stack of images in
a 3 dimensional array.	(This capability would also be useful in MIN and
MAX.)  3.  Allow the user to specify a "kernel" (as with CONVOL) specifying
weights to be assigned to pixels within the filter size.  One such
application of a such a generalized median is described at
http://adsabs.harvard.edu/cgi-bin/nph-bib_query?bibcode=1995PASP..107..496S&d
b_key=AST&high=38aaedbba709215

--Wayne Landsman                           landsman@mpb.gsfc.nasa.gov


Sent via Deja.com http://www.deja.com/
Before you buy.