Will it make a comeback?

Will it make a comeback?

Attached: Fortran_acs_cover.jpg (610x780, 91K)

No, because literally everything good from it has been ported to newer, less clunky languages. I'm pretty sure the only reason why modern FORTRAN implementations are maintained is because some hyper-autistic supercomputing software absolutely needs it to function.

When did it go away?
t. Python nu-dev who probably thinks his libraries are actually written in his shitty language
FORTRAN is fast as fuck

I write everything in Lisp and Forth actually.

I honestly doubt it, unless your "everything" is just the occasional snippet to jerk off in meme threads.

No, you're welcome

It never went away. People with a brain use it instead of C/C++

Go, Swift, D are all gc languages and so dont count as a replacement to C/C-++

Rust borrow syntax is too difficult for mainstream use

I had to use this when working with particle simulations in college.
For those unfamiliar with Fortran, imagine trying to debug an application that's thousands of lines of code. All this code was written by somebody who barely knows how to code, who broke every single good programming practice you've heard of. Everything is in a single main method. Global variables and side effects run rampant. There is no abstraction. There are GOTO statements and subroutine calls everywhere. You try to make a small change and the program segfaults because you put something in one of the first six columns (which are reserved for punchcards), bad enough that you may need to reboot your computer because it's corrupted some of your memory.

All working with Fortran has taught me is the value of things like pure functions and immutable data structures.

Fortran was once cutting edge but those times are long-gone. The sooner this language dies the better.

Attached: anipepe.png (999x1228, 596K)

It's not really any better compared to C. The only real difference is Fortran's extremely strict aliasing rules which apply even to pointers of the same type. C fixed this optimization gap when it introduced the restrict keyword; it's as fast as Fortran since a long time ago. Or should I say, it's as undefined as Fortran. It's great for arrays and matrices and shit but it makes systems programming literally impossible -- -fno-strict-aliasing is pretty much a standard flag on my C toolchain.

You forgot about multiple function entry and exit points.

SUBROUTINE S(X, Y)
R = SQRT(X*X + Y*Y)
C ALTERNATE ENTRY USED WHEN R IS ALREADY KNOWN
ENTRY S2(R)
...
RETURN
END

C USAGE
CALL S(3,4)
C ALTERNATE USAGE
CALL S2(5)

C SUBROUTINE WITH ALTERNATE RETURN. THE '*' IS A PLACE HOLDER FOR THE ERROR RETURN
SUBROUTINE QSOLVE(A, B, C, X1, X2, *)
DISCR = B*B - 4*A*C
C NO SOLUTIONS, RETURN TO ERROR HANDLING LOCATION
IF DISCR .LT. 0 RETURN 1
SD = SQRT(DISCR)
DENOM = 2*A
X1 = (-B + SD) / DENOM
X2 = (-B - SD) / DENOM
RETURN
END

C USE OF ALTERNATE RETURN
CALL QSOLVE(1, 0, 1, X1, X2, *99)
C SOLUTION FOUND
...
C QSOLVE RETURNS HERE IF NO SOLUTIONS
99 PRINT 'NO SOLUTIONS'

It's still used in research and that's about all it's useful for. Making some complex simulation run faster. It's never going to make a huge comeback because 99% of people don't need the performance you can get from it if you program within its constraints, which are a cunt.

Never left

Can I get a job in Fortran?

COBOL seems to be the better choice. Developers are dying left and right leaving banks and the like without programmers to maintain their ancient software

Never left, plus there's infinite legacy shit thats still being maintained and thus never can leave.

The real problems with rust aren't about its syntax, thats just what makes the brainlets cry. It is completely unusable in 2018 but by version 2 or 3 it'll probly start seeing real use.

what about this is the fault of FORTRAN rather than the person who wrote the program? you can do all of this shit in any other language too, and the punchcard/line printer centric shit doesn't exist anymore in newer standards

90% of the arguments against this language always seem to boil down to "it's OLD! haha look there's punch card stuff why aren't we using python instead??? that does math too" or "but I saw shitty code in it once that I couldn't understand"

>the program segfaults because you put something in one of the first six columns (which are reserved for punchcards
How do you fuck that up?

Attached: 1476634433082.jpg (514x459, 40K)

Holy fizzbuzzing Christ, you NEETs are complete hacks even when you have all the time in the world to actually get good at programming.

Fortran is nice for physics. It has good array syntax and it goes really fast. You can always wrap it in some python if you are feeling gay (I usually am).

Fortran is used in practically every research field and isn't going anywhere. Most academics don't have the time, or the need, to port their software or libraries to another language because it's the hot new language which compiles a bit faster. The Fortran code is still going to run faster and at least it has proper/powerful array syntax compared to a language like C :^).

Fortran gets a bad reputation because academics generally have a lot of terrible programming habits or still use Fortran 77, but that doesn't mean Fortran is bad. The new Fortran standards have a bunch of features which younger languages do -- even if they've basically only been implemented just to keep the language 'current'.

Don't make fun of my unemployment, user. I'm trying my hardest :

As someone who used fortran despite barely knowing how to code, I understand it. Physicists and chemists (probably) wrote that code, they didn't give a shit about good programming practice and all they wanted was something that worked. Reading fortran is piss easy anyway.

>or still use Fortran 77
There is literally nothing with fortran 77.

This

It is still heavily used in anything that requires a lot of number crunching. Computational fluid dynamics, finite element structures, multiphysics solvers etc all are heavily Fortran+MPI based simply because it’s really easy to write very fast number crunching code with it, whereas you’d have to muck around with a bunch of pointer bullshit if you wanted to do the same with C/C++. This is something physicists/mechanical engineers etc don’t want to bother with because these numerical methods are extremely complicated to begin with and further obfuscation due to having to deal with C’s intricacies is counterproductive.

This is why even rather modern stuff like NumPy rely heavily on Fortran routines in the background. Whenever you want to find the eigenvalues of a matrix etc, Python/Matlab will typically rely on LAPACK due to how fast it is compared to native Python code.

>wrap FORTRAN in Python
If you're that desperate just use numpy.

This. If you want to make money using an ancient language, COBOL is the one.

Generally there is nothing wrong with Fortan 77 but I've found that academics who still use Fortran 77 usually write the least readable codes.

it is still relevant, just use the modern standard (at least 90 or 95, preferably 2008)

>All this code was written by somebody who barely knows how to code
same goes for java, C (legacy or not), or pretty much any language
>Everything is in a single main method
there is a tendency for larger methods that do more things but anyone who puts everything in main without using modules is a retard or the program was written in the 60s or 70s
modern fortran has all the modularity features that say C has (actually, fortran has more)
>here are GOTO statements and subroutine calls everywhere.
so is everything in the main method or is it in subroutines? Make up your mind.
gotos are perfectly fine as long as they are not used obscurely (like assigned gotos, or in place of other control flows like loops)
they are great for breaking out of nested loops (even though fortran allows labeled loops + exit (that is break for you C-babs) with labels)
Subroutines are just functions that instead of returning a value performs operations that usually are impure (e.g heavily modify an array)
>You try to make a small change and the program segfaults because you put something in one of the first six columns (which are reserved for punchcards)
why are you using fixed format unless you are working on legacy code?
Use free format. It allows longer lines that fixed format too (up to 110-120 characters until you have to add breaks iirc, rather than the 6+72(?) of fixed)
>All working with Fortran has taught me is the value of things like pure functions and immutable data structures.
so you have never actually used it for array processing?

the most AIDS thing about fortran is reading input from unformated (or poorly formated) files, and perhaps formating output

Attached: 1517901332573.png (497x473, 241K)

>so is everything in the main method or is it in subroutines? Make up your mind.
With Fortran subroutines can be in the main file, you just tack them on at the end. I'm assuming that's what he meant.

I suppose that can be pretty messy if you throw global/shared variables into the mix

definitely.

PHP is the only language the world needs. Forget your Fortran, C, Java, Go, Rust... All trash compared to the superiority and beautiful syntax and useability/scalability of PHP.

I still have to grab the compiler for numpy

this. i have written my own kernel in php and it's better than ganoo plas leeenox

If only it would die, fucking weather model written in Fortran.

>scalability
Buzzword bingo!

Tfw as I wrote that with full knowledge that scalability comes more down to the platform used to run/host the server, nginx Vs Apache and such.

Fortran is now object oriented too, retard.

No, it's still the favorite among Physicists writing HPC because it's a much nicer language to use than C in that regard. A lot of legacy FORTRAN code is being ported to C++ though/