| Re: IDL as programming language? [message #2721 is a reply to message #2715] |
Mon, 29 August 1994 13:50   |
shapiro
Messages: 4
Registered: February 1994
|
Junior Member |
|
|
Geoff.Sobering@nih.gov (Geoff Sobering) writes:
> In article <1994Aug28.122441.32497@waikato.ac.nz>, abz@waikato.ac.nz wrote:
>> OK. One the lecturers in our department recently went to
>> Hawaii & America for a few weeks (lucky him) and came back
>> with the idea of using IDL as a complete programming
>> language rather than as just a plotter for FORTRAN generated data.
>> ...
> I don't want to use bandwidth, time, etc. repeating the comments previous
> posters have made to this thread (which I mostly agree with).
> I just wanted note that I use IDL for almost all of my programing, both
> scientific and otherwise.
> For example, I have a network monitoring program that polls our network-hub
> (via 'spawn'ed telnet) and keep track of network traffic. It may also be
> interesting to note that RSI wrote the GUI-builder that is part of 3.6.12
> *in* IDL.
RSI has a problem with quality control. I have been developing software
in IDL for the past three years. IDL has been no better or worse than
any other language. Initially IDL seems like a wonderfull language, it
has all sorts of canned routines and a fairly easy GUI builder. However
every time a new version of the language is released I spend at least a
week patching up old software that the new version breaks. If you are
only writing little programs with a limited lifespan then IDL is
perfect. If you are working in a larger project (ours is 6000+ lines)
the instability in the language can be a real pain.
RSI seems to have little vision of where there product is headed. So
features are added haphazardly and with little forthought. As a result
IDL is ripe with kludge and nuance. They also have a problem with
documenting changes when they happen. Once they included a whole
paragraph in the release notes arguing that some bug was a feature due
to backward compatability (to version 1.1) while the same release notes
failed to mention that widget_info had changed from a procedure to a
function. More recently RSI has added pointers to IDL. They never
mentioned this in the release notes, only that something about pointers
was broken and has now been fixed.
I would never recomend IDL 'as just a plotter for FORTRAN generated
data.' There are far to many cheep or free ploting packages already
available. As a language it seems better suited to little programs and
'one off' projects. Large scale and long term projects are best
attempted in a more stable language (like C).
-Andrew T. Shapiro
CSES/CIRES University of Colorado
shapiro@cses.colorado.edu Campus Box 216
(303) 492-5539 Boulder, CO 80309-0216, USA
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
