Fortran Notes

Fortran is a programming language that was invented back in 1954 by IBM. Fortran is faster and more efficient than any other programming language, making it well-suited for tasks such as scientific simulations and large-scale numerical modeling.

These notes are based on the book "FORTRAN 77 for Humans" written by Rex L. Page and Richard L. Didday and published in 1980.

General

g95 -o NAME SOURCE.f compile source files into executable named NAME
program code must start on 7th column (for historical reasons)
comments start with C

Arithmetic Expressions

INTEGER standard signed integer, can have +/- in front
REAL signed floating point numbers, can use scientific notation eg. 6.0123E+23
CHARACTER*n string of character enclosed by single quotes must specify n, which is length of string when declaring

Operations

PRINT *, list prints each of the values in the list of a line
READ *, list places data into list from the data line (user input)
END tells the compiler to stop doing statements in program
Operations are +, -, *, /, ** (last is exponentation)
If your line is too long, then put & in 6th column of next line:
```
PRINT *, 'THIS LONG LINE IS',
&'CONTINUED HERE'
```

Loops

IF (e1 rel e2) GOTO s e1 and e2 are expressions, both of the same type rel is relational operator s is statement label (line number) decide if relationship is true, if yes proceed to line s
Relational operators: .LT. less than .LE. less than or equal to .EQ. equal to .NE. not equal to .GE. greater than or equal to .GT. greater than

indent if loop blocks, so starting after IF and going until last GOTO:

100  IF (test for exit) GOTO 200
         fortran operation
         .
         .
         GOTO 100
200 continue program

can do while loops (if statement at beginning) or do-while (if statement end)
pre-test loops are safer and clearer
Do loops are like for loops:
```
      INTEGER TOP = 5
      DO 10 N=1, TOP,1
10    PRINT *, N
```
Code above loops ten times. After DO the first value is line number after which the loop goes to the next iteration. N is the iterator that increases or deacreases, top is ending value and last integer is increment.
CONTINUE statement that does nothing, used as terminal statement in the range of DO-loops

Nested loops are possible. Note that both line statements point to line 10:

    DO 10 I=1, 10, 1
        DO 10 J=1, 10, 1
10          PRINT *, I*J

Logical operators

If-else structure in fortran:

IF (test condition) THEN
    fortran operation 1
ELSE
    fortran operation 2
ENDIF

Can also use multiple else-if with ELSE IF structure. Add THEN after each ELIF
.EQV. equivalence operator, used to compare boolean values
.NEQV. nonequivalence operator
May omit ELSE in IF-THEN statement.
LOGICAL data type to hold booleans
Use INT(x) to truncate real number x or NINT(x) to round
MOD(x, n) is modulus operator

Arrays

INTEGER ARR(n) declare an array of length n
ARR(1) access first element (indexing starts from 1 not 0)
MUL(3,3) to create 2-dimensional array. first row, then column

To loop arrays use the following structure (reads 20 ints from user):

    DO 100 I=1,20
        READ *, ARR(I)
100     CONTINUE

Character Strings

Substrings can be accessed with STR(j:k). STR is string, j and k are indices

multiline code can be written with + sign in between:

IF (PHRASE .EQ. 'HECK'
+ .OR. PHRASE .EQ. 'DARN')

when assigning substrings, can't reference same portions eg. N(1:9) = N(5:15)
INDEX('ABCDEF', 'C') operator returns 3 as a result. Returns 0 if not found. - A = B // C concatenates B with C and stores it in A

Subprograms

Fortran uses subroutines and functions
CALL AVG(a1, a2) calls subroutine AVG with arguments a1 and a2
SUBROUTINE AVG(a1) declaration of subroutine
Subroutines must end with RETURN and `END statements
Function returns one value, subroutines can return any number of values through it's arguments. Choosing which to use is matter of taste.
type FUNCTION f(p1) declares function f that takes in one parameter returns variable with name f at RETURN statement
first write main program, then function/subroutine definitions

Line Spacing

Format PRINT statement by replacing * by line where FORMAT is stated:
```
    PRINT 10, SNGLSP, 'STRING'            
10  FORMAT(A,T25, A)        
```
This will print the text STRING in 25th column.
SNGLSP is carriage control character that can be the following:
- move down the page one line before printing
- 0 move down two lines
- 1 move to the top of next page
- + overprint previous line

Looks like SNGLSP is not needed when printing to terminal, so this is enough:

    PRINT 10, 'STRING'            
10  FORMAT(T25, A)

Different data descriptors: A for chars, I for ints, F for reals (I4 or F6.2)

Files

OPEN(speclist) open files with this command
CLOSE() disconnect file
READ/WRITE() read or write files

igorp/fortran