![]() When it comes to online tutorials, be aware that there are many obsolete ones. You can get a lot of help at Stack Overflow in the Fortran tag. Modern Fortran: Style and Usage by Norman and Spector is for more advanced programmers who want to use good style but know the basic language. There are several other books that I did not have in my hand. Some other books are oriented more practically, like the Fortran for Scientists and Engineers series by Chapman or Modern Fortran in Practice by Markus. ![]() It is written in a university textbook style and I prefer to use it as the recommended reference for my course. However, it is rather terse and can also be used as a reference book. ![]() I always preferred the Fortran Explained (Fortran 90 explained, Fortran 2003 explained, Modern Fortran Explained) by Metcalf, Reid and Cohen. Fortran 2008 added smaller enhancement and parallel programming using coarrays.įortran can be learnt from book and there is a choice different styles of introductory books. The amount of features that appear in the successive revisions increases strongly, the biggest steps being Fortran 90 and Fortran 2003.įortran 90 was completely superseded by Fortran 95 which fixed several defects or inconsistencies.įortran 77 and earlier are very old and the programs often use techniques that are strongly discouraged today (and even deleted from the standard or marked obsolescent).įortran 2003 brought object orientation, portable interfacing to C and much more. Python was not originally designed for number crunching, and neither was C or C++ or Java: these languages were designed to be more general, and are used for writing everything from operating systems, to GUIs, to web applications, to computer games, etc., whereas Fortran was literally designed for number crunching (a bit like MATLAB and Julia, the latter which was discussed here at MMSE) so once you spend some time familiarizing yourself with it, I'm confident that you'll find it much more natural for scientific computing than other languages you've used in the past.įortran is a language that evolved strongly and a programmer just maintain very old legacy codes will have very different needs from a programmer maintaining codes written in Fortran 95 and different needs will have a programmer who wants to employ modern techniques and develop new program that use Fortran 2003, 20 (and future revisions). I find Fortran to actually be in many respects, simpler and easier to master than Python, especially since most Fortran code is not object-oriented, and you'll see that it's really designed for scientific computing (the way you can output numerical values for example, will probably convince you of this). Changing the number of digits for a numerical output in QE is an easy thing that I think you can try doing right now! Next you can add more output (more words, more numbers, etc.) You could also try to add new input flags for the feature you want to implement, and at every step of the way, see if it compiles, and follow the error messages at compile time: I learned so much Fortran that way! Very soon you'll find yourself feeling quite comfortable and you'll gain your confidence this way! I recommend playing around with existing Fortran code, and adding simple features with gradually increasing complexity. It turns out though that I learned much faster "by doing" rather than by reading. I went to the library and picked up a bunch of books (online tutorials and youtube videos were unheard of at the time), thinking that I'd learn based on strategy #1 above. When I began Fortran as an undergrad summer student, I was told ahead of time that I'd be coding in Fortran, and I dreaded it. (2) Practicing: As the comment by user tmph says the best way is often just to practice.
0 Comments
Leave a Reply. |