Why do astronomers stick with old languages like IDL and FORTRAN?

[Simfish]

I've heard that there's some movement towards newer languages like Python. But is there some resistance to this movement? Is there a similar problem with physics departments?

Do astronomers prefer their grad students to be fluent with FORTRAN or with Python? I know one professor who doesn't know much Python yet, but who recognizes the future of Python, so he let's his students do work in Python.

 

[Vanadium 50]

What problem would it solve to have a mass migration away from FORTRAN?

 

[Simfish]

Good point. I just have the impression that other departments are migrating away from FORTRAN first, and I'm wondering why those departments are doing it before astronomy (though that may be more relevant to the interests of those departments).

 

[NobodySpecial]

and I'm wondering why those departments are doing it before astronomy.

They don't have the same base of tested working complex code going back 50years?

 

[turbo]

FORTRAN is trusted and well-understood. Why change? For the same reason, why transition from SAS if you can get what you want out of that package? They are really simple basic tools that run like a screaming monkey on even a base-level PC, so you can get a lot of work done in a short period of time, and run multiple trials with variables to compare them easily. Those reasons are good enough, IMO

 

[harcel]

In astronomy the use of Fortran actually is going down. Most serious large scale software is written in C or C++. Interactive data analysis indeed is very often done in IDL, but Python is definitely gaining some user base in astronomy now too. Good code, written in a language however old, will of course stay in use. Translating it to something 'modern' is a waste of time.