Another One Bites the Dust: GRAM Atmosphere Model

It looks like the Fortran community has lost another venerable NASA Fortran library to C++. I recently noticed that the latest release of NASA’s Global Reference Atmosphere Model (GRAM) is now C++. From the release link:

Earth-GRAM 2016 is now available as an open-source C++ computer code that can run on a variety of platforms including PCs and UNIX stations. The software provides a model that offers values for atmospheric parameters such as density, temperature, winds, and constituents for any month and at any altitude and location within the Earth’s atmosphere. Earth-GRAM 2010 is available in FORTRAN. Similar computer models of Mars, Venus, Titan, and Neptune are available for diverse mission applications through NASA’s Space Environments and Effects (SEE) Program, a partnership with industry, academia, and other government agencies that seeks to develop more reliable, more effective spacecraft.

The Earth-GRAM model, originally written in Fortran 77, is produced by MSFC and has been around for a while, with many revisions (i.e., GRAM-86, GRAM-88, GRAM-90, GRAM-95, GRAM-99, GRAM 2007, and GRAM 2010). In the 2010 revision, they converted the old Fortran 77 code to Fortran 90, which I thought was a step in the right direction (although maybe 20 years too late). It seems like they could have used Fortran 2003 and provided a C interface using the standard Fortran/C Interoperability feature, which would have enabled calling it from a variety of languages (including C++).

I remain utterly unconvinced that C++ is the programming language that engineers and scientists should be using. Are CS people taking over programming jobs once done by engineers? Is this a failure of university engineering departments? Although it seems like even CS people don’t want to be using C++ anymore. Every time I turn around a new programming language (Java, Rust, Go, D, C#, Swift, …) is created by CS people fed up with the C++ dumpster fire. So I don’t know who is telling engineers that C++ is the future. And the perception that Fortran is an obsolete programming language is very strong, even within the engineering/technical disciplines that it was designed for. Sometimes this is due to sheer ignorance, but sometimes it is due to ancient Fortran 77 legacy code that people have to deal with. I sympathize with them, Fortran 77 code is terrible and needs to go away, but it should be modernized, not thrown out and replaced with C++. Think of the children! (and the memory leaks!)

Parting words of wisdom

See also


Tagged with: , , , , , ,
2 comments on “Another One Bites the Dust: GRAM Atmosphere Model
  1. Pierre says:

    Another possibility is that the scientists “in charge” today were taught in the 90’s and that at that time there was a big urge to modernize code. “Modern Fortran” was not ready or known enough at the time and C++ appeared like a solution.

    So what you observe today might just be the “lag” of science in terms of computer programming. Until proper recognition of scientific computing in higher education and research, choices will be motivated by 20 (or more) years old tech.

Leave a Reply

Your email address will not be published. Required fields are marked *