Re: Woes of Software Distribution, Re: Something else

[Craig Markwardt <craigmnet(at)cow.physics.wisc.edu>]

davidf@dfanning.com (David Fanning) writes:
>
> I have the same problem with my library. If you make
> programs that are *designed* to work together, then
> you have to get all of them. If you try to put a small
> utility routine (CMCongrid is a good example) in all
> the dependent routines, then it is virtually impossible
> to make a change to the utility routine, because you 
> can't find all the copies.

I agree with most of your sentiments.  This is kind of the same
philosophical difference between Microsoft and Unix.  Microsoft tries
to abide by the KISS principle (Keep it Stupid, Simpleton!), whereas
Unix users tend to be forced to figure everything out themselves.  I
try to take the middle road.  I reckon that a lot of people don't
really understand about having your own IDL path with your own custom
programs.  Why burden people too much more?

I also had the same problem of modifying utility routines in my
published programs.  I solved the problem this way: in my personal IDL
directory, the utility programs are kept in separate files.  When I
put them on my web page I run a little script which inserts the
appropriate utility routines directly into the code.  If you look at
PLOTIMAGE or PLOTBIN, you will see something like "insert HERE," which
is a remnant of the script's cue to insert the utility routines.  Kind
of ugly, but it keeps things organized.

Craig

P.S.  And yes, I'm a Unixer.

-- 
--------------------------------------------------------------------------
Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
--------------------------------------------------------------------------