I've been doing programming in science for 3 years now, and Fortran is a huge, huge problem. Fortran is too much like C, and think about how good C is for writing differential equations. There are tons of potential bugs in the programs I see, and very few opportunities to check for correctness. Development time takes an order of magnitude longer than it should. I've argued with physicists many times over this, and the issues they bring up are always:
- Alternatives are too slow
- I have a huge code base that I don't want to port over.
- I don't need (eg.) closures. If I ever designed a program that needed them, I wouldn't choose Fortran, but that has never happened
- (most popular) Programming is just a tool. I'm interested in the physic, not programming, and Fortran is Good Enough
I don't know if Fortress will succeed. I hope it does, but unless they can show that Fortress programs are both shorter and faster, I don't think it will happen. Mathematica, Matlab and IDL already have too much of a lock on the market for 'inefficient' computing.
My money is on a Mathematica clone eventually becoming faster than Fortran. It should happen.
Believe it or not, Fortran is still used extensively in scientific research circles.. Although badass Java libraries such as JScience are starting to take over.
"As a result, we are able to support features in Fortress such as transactions, specification of locality, and implicit parallel computation, as integral features built into the core of the language."
"Moreover, Fortress has been designed with the intent that it be a ``growable'' language, gracefully supporting the addition of future language features. In fact, much of the Fortress language itself (even the definition of arrays and other basic types) is encoded in libraries atop a relatively small core language."
I haven't looked at the details of Fortress yet, but I don't read 'Java' in the above quotes.
- Alternatives are too slow
- I have a huge code base that I don't want to port over.
- I don't need (eg.) closures. If I ever designed a program that needed them, I wouldn't choose Fortran, but that has never happened
- (most popular) Programming is just a tool. I'm interested in the physic, not programming, and Fortran is Good Enough
I don't know if Fortress will succeed. I hope it does, but unless they can show that Fortress programs are both shorter and faster, I don't think it will happen. Mathematica, Matlab and IDL already have too much of a lock on the market for 'inefficient' computing.
My money is on a Mathematica clone eventually becoming faster than Fortran. It should happen.