[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

IDL Survey Comments

RSI sent me a survey today and asked for my opinion on a variety of
IDL related topics.  This got me to thinking, and I composed the
comments appended below.  I wonder how many other IDL users share some of these
same thoughts?

----------Additional Comments sent to RSI as part of their yearly survey: ---

To: 		survey@rsinc.com  
Subject:	Additional Comments

I have some comments that you probably don't want to hear but I feel strongly
about.  I feel that RSI is spending too much time introducing new features
while slowly letting the core of IDL become out of date.  This has been going
on for some years now and I think it will eventually ruin IDL if something 
isn't done about it soon.  I will give some specific examples below, but lack 
of time prevents me from compiling an exhaustive list.

1.  Provide better tools for large projects:  There are several problems that
are beginning to hurt us.  As Ray Sterner's library (the IDL guru at
JHU/APL) has grown to hundreds of routines, none of us, Sterner included, can
remember all of the function names anymore.  The fact that there is no
notational difference between a function and an array is causing problems.  
(For instance, WORK(T) might refer to an array named WORK or a function.) 
Another problem is that it is difficult (impossible?) to tell just where on 
your path a routine is being picked up from.  This is frequently encountered 
when one is experimenting with modifying a local copy of a library routine 
and later the local copy is forgotten.  At the very least, when I ".run" 
something it ought to print out the path to that routine.

2.  Provide modern support for presentation quality graphics:
Ditch the Hershey caligraphy set and support on screen characters properly!  
Any characters written to the screen or to a Tiff file look like they are
straight out of the 70's.  IDL has not kept up in this area and it is 
gradually forcing us away from IDL for all applications involving presentation 
graphics. In all cases where I need a Tiff file to send to our $25K Kodak dye 
sub color printer I have to resort to Adobe Photoshop to get decent lettering.

Along the same lines: provide some decent tools to make 
presentation/publication quality graphs (which include things like legends, 
arrows to features, multiple plots).  True it can be done in IDL, but I have
to give coordinates to everything rather than just moving it around 
interactively on screen.  This has led me to abandon IDL in favor of IGOR Pro 
for 100% of my presentations and publications.

	In conclusion then, the picture from my perspective is that I use IDL
mostly as a very convenient high level analysis language.  I used to use it to
prepare slides for talks and illustrations for publication, but there are so
much better tools on the market that I have since mostly abandoned this.  
While RSI seems to be always looking for ways to expand IDL, they have 
neglected the core features that attracted me to IDL in the first place.  I 
really don't much care about new point and click applications.  RSI is never 
going to be able to anticipate my research analysis needs.  What I do want is 
for the core of IDL to be brought up to modern standards in at least the 
areas discussed above.

		Regards, Bruce Gotwols
Bruce L. Gotwols
Johns Hopkins University, Applied Physics Lab., Laurel MD 20723
Internet:  gotwols@tesla.jhuapl.edu   (