Re: IDL or PV-WAVE? [message #100 is a reply to message #99] |
Mon, 17 June 1991 08:25   |
landsman
Messages: 93 Registered: August 1991
|
Member |
|
|
jvb7u@fermi.clas.Virginia.EDU (Jon Brinkmann) wrote favorably about PAW;
a package for reducing high energy physics data. He mentioned it having
superior memory management, graphics and least squares capability to
interactive languages such as IDL or PV-WAVE.
I do not know PAW so I cannot comment directly. But for me the main
selling point of IDL is not as a plotting package, or image processing
system or math library, but rather as a programming language. (In fact,
I sometimes begrudge IDL for its fancy 3-D graphics which I never use, but
which add bulk to its ever growing manual.) IDL lets me spend more time
being a scientist and thinking about my data, and less as a programmer
thinking about DO loops and data types. For example, suppose I have
wavelength and flux vectors, and I want to know the wavelengths where
the fluxes are negative. The IDL statement
IDL>print,wave(where (flux LT 0) )
is not only more concise than coding this in FORTRAN or C but also closer
to what I am thinking as a scientist. Of course, some people are better
able to switch between scientist and programmer modes of thought. But
when I programmed in FORTRAN, I was always frustrated by the long time
lag between thinking about what I wanted done with my data, and actually
doing it.
Some day in the future, computers may understand human speech and we
will be able to directly tell them to "Normalize these two spectra and
then plot their difference" or "Set all the negative values of this image
to 0 and display using a logarithmic color table". Meanwhile, I'll
have to settle for IDL...
Wayne Landsman
MOUSSE::LANDSMAN (SPAN) landsman@trifle.gsfc.nasa.gov
ST Sytems Co. NASA/GSFC (301)-286-3625
|
|
|