Re: It costs how much?!!?!

["J.D. Smith" <jdsmith(at)astro.cornell.edu>]

billybobsuzanne@my-deja.com wrote:
> 
> Wow.  I have been hanging around this newsgroup reading for a little
> while now and I downloaded a demo of IDL a month or two ago to
> determine what IDL could do.  Despite being limited to a 7 minute demo
> now, I am quite impressed.  I finally got around to asking RSI exactly
> how much it cost...  WOW.  I am a meager graduate student and my
> supervisor isn't too keen on paying so much money for software which he
> figures I can live without... "How much??? (cough, cough, weeze) I'm
> sure you can do that in Excel" were his exact words.  I'd buy it
> myself, but there is no way I could afford it at what they quoted me.
> I looked into the 'student version' but it is far too limited to be of
> much use with my *large* datasets.  I guess it is back to
> fortran, 'gri' and 'gnuplot' for me :(
> 
> Will the people at RSI ever come back to reality here?  Their academic
> prices are simply too high (a few hundred dollars perhaps) but what
> they are asking is crazy.  I would have thought that us
> engineering/physics graduate students are the people that RSI would
> target sales for!  In a few years, I'll be in industry making decisions
> about software purchases...  too bad they don't realize this...

We can all sympathize with you there, Bill.  If you have a large user
group of IDL, you can substantially reduce the cost per license by
pursuing a site license agreement (with some restrictions).  I did this
in the department here, and now we have 100 licenses to play with (there
are other versions ... down to 25, or up to 500 licenses).  It's
actually a tremendous amount of work, getting dozens of faculty and
staff to agree on fair pricing policies, etc., so make sure you know
what you're in for if you go this route.

Otherwise, you might consider giving the student version a try.  I was
once considering rewriting all of my code to be compatible with the
student version -- and to deal with arrays all of which were smaller
than 65k (the student version limit).  How?  Use arrays of pointers to
smaller arrays.  E.g., you have a 512x512 image.  You can break it up as
4 256x256 images, all sitting in a pointer array im=ptrarr(2,2).  Of
course you lose much of the convenience of the vector language, but at
least all of the plotting, visualization, etc. is there.  With only one
layer of pointers you can get up to 4.3e9 elements.  And if you have
16GB of memory, you can afford to buy IDL.   Most of the I/O
restrictions are trivially overcome by using FITS or someother custom
format (though you might have to modify them to deal with the array
partitioning silliness). One concern is the student version still seems
to be based on v5.0, but surely that must be out of date. I don't
understand why they wouldn't bring the S.V. forward to.

You might also consider free IDL alternatives (no free replacements,
sadly), some of which were mentioned a month ago or so in the group. 
Matlab also has a cheap fully functional version for students I
believe.  Not that I advocate using it;)

Good luck,

JD