Modernizing the Fortran Ecosystem
A team of researchers and programming professionals from universities in the USA, the U.K., Germany, Brazil, France, the Netherlands, and Canada are bringing the classic programming language Fortran into the 21st century. Their research not only digs into the strengths and limitations of Fortran but also focuses on making the language a more popular go-to for those who code for scientific and mathematical applications.
The Challenges Fortran Faces
Fortran, which has been around since 1957, has waned in popularity despite being an effective solution for a range of scientific and mathematical challenges.
Fortran Lacks a Standard Library
A standard library (stdlib) is the cornerstone of many programmers’ workflows because it makes it faster to complete common or repeated functions. Even though Fortran has a range of preset modules and procedures, it lacks the kind of stdlib that makes Python, C++, and other popular solutions easier to use.
A Relatively Sparse Online Community and Resources
Fortran lacks a centralized website and a fervent online community of users. With a website, new Fortran users would have access to resources and learning tools, which would pave the way for wider Fortran adoption.
In addition, Fortran lacks a community compiler that programmers can use to build and test new tools.
Resources for Overcoming Fortran’s Challenges
To better position Fortran for widespread usability, the research team has begun to implement a series of solutions. These can aid in both Fortran’s adoption by programmers and make it a more popular focus for university curriculums.
A Fortran Stdlib
By creating a stdlib for Fortran, the team gives the programming community a powerful resource for improving the quality of code and speeding up the programming process.
Also, with a Fortran stdlib, there’s no need for programmers to code app components from scratch if what they need is already in the stdlib. This helps eliminate redundant, time-consuming coding.
A Fortran Package Manager
Currently, a programmer new to Fortran has to learn both the language and at least one build system. But the Fortran package manager the team is developing provides a build system for Fortran and simplifies the compiling process.
Another core benefit of the Fortran package manager is the way it enables you to reuse code from one project to another and manage dependencies that different Fortran projects share. This not only saves developers time but also reduces the chance of errors sneaking into the code.
The Next Steps for the Fortran Community
New and experienced Fortran users, university professors, and programming educators should embrace the new resources this team is developing. This is particularly important in light of the increasing popularity of machine learning (ML) programming. ML programming often uses arrays as a primary data type. This makes Fortran a natural fit for machine learning solutions because it supports multidimensional arrays.
Of course, the potential for Fortran isn’t limited to ML, and these new resources chip away at barriers that have been inhibiting more widespread adoption. If programmers, educators, and data scientists give the team’s resources a shot, they can lead the way in making sure Fortran is top of mind for the professionals it benefits the most.
To dig deeper, you can check out the complete paper here, in IEEE Computer Society’s digital library, or complete the form below for a free download.
