| [4773] | 1 | The Fortran conventions that the author of the code in this directory |
|---|
| 2 | aspired to are as follows: |
|---|
| 3 | |
|---|
| 4 | - Module names are prefixed by the static library name, in this case |
|---|
| 5 | "radiation" (no need to suffix by "_mod" since this is always |
|---|
| 6 | obvious from the context). |
|---|
| 7 | |
|---|
| 8 | - Names of derived types are suffixed by "_type". |
|---|
| 9 | |
|---|
| 10 | - Implicit none everywhere. |
|---|
| 11 | |
|---|
| 12 | - Logicals are prefixed by "do_", "use_" or "is_", and should have |
|---|
| 13 | clear descriptive names. |
|---|
| 14 | |
|---|
| 15 | - All variables in SI units by default. |
|---|
| 16 | |
|---|
| 17 | - Variables not in SI units have the units in the variable name, |
|---|
| 18 | e.g. pressure_hPa. |
|---|
| 19 | |
|---|
| 20 | - Integers either prefixed by "j" to indicate loop counter, "n" to |
|---|
| 21 | indicate the size of an array dimension, or "i" otherwise. |
|---|
| 22 | |
|---|
| 23 | - Loop counters should be brief and contain no underscores. |
|---|
| 24 | |
|---|
| 25 | - No variable names should be only one letter, since then they are |
|---|
| 26 | difficult to search for in an editor. |
|---|
| 27 | |
|---|
| 28 | - Integers variables beginning with "n" or "i" should either contain |
|---|
| 29 | no underscores (e.g. "naerosoltypes"), or should have an underscore |
|---|
| 30 | immediately after the "n" or "i" (e.g. "n_aerosol_types"). Thus, a |
|---|
| 31 | variable beginning with "n" or "i" containing an underscore but not |
|---|
| 32 | as its second character need not be an integer (would normally then |
|---|
| 33 | be a real or a derived type), in order to enable real variables |
|---|
| 34 | such as "ice_water_content". |
|---|
| 35 | |
|---|
| 36 | - All variables and procedure names in lower case using descriptive |
|---|
| 37 | names, and if not too long these will be composed of complete words |
|---|
| 38 | separated by underscores. |
|---|
| 39 | |
|---|
| 40 | - All parameters are in CamelCase, and integer parameters are |
|---|
| 41 | prefixed by "I" or "N". |
|---|
| 42 | |
|---|
| 43 | - C-style enumerations (using "enum, bind(c)" and "enumerator") |
|---|
| 44 | should be preferred over integer parameters when they refer to a |
|---|
| 45 | specific setting, because "if (solver_sw==ISolverTripleclouds)" is |
|---|
| 46 | clearer than "if (solver_sw==5)". |
|---|
| 47 | |
|---|
| 48 | - Variable character strings are suffixed by "_name" or "_str", while |
|---|
| 49 | parameter character strings (in CamelCase) are suffixed by "Name" |
|---|
| 50 | or "Str". |
|---|
| 51 | |
|---|
| 52 | - Global data included in modules must be constant; any variable data |
|---|
| 53 | must be passed to procedures via dummy arguments. |
|---|
| 54 | |
|---|
| 55 | - Long argument lists avoided by grouping variables thematically into |
|---|
| 56 | derived types. |
|---|
| 57 | |
|---|
| 58 | - Any variable used in a source file should be defined in a comment |
|---|
| 59 | earlier in the source file (if this is not obvious from the |
|---|
| 60 | variable name) including its units if it has them (even if they are |
|---|
| 61 | SI units). |
|---|
| 62 | |
|---|
| 63 | - All functions and subroutines to be defined in modules and their |
|---|
| 64 | explicit interfaces to be revealed via the "use" command, which |
|---|
| 65 | should use "only". |
|---|
| 66 | |
|---|
| 67 | - Data are read in at run-time from NetCDF files rather than being |
|---|
| 68 | embedded in the code (e.g. using data statements). |
|---|
| 69 | |
|---|
| 70 | - Fortran "statement functions" should not be used. |
|---|
| 71 | |
|---|
| 72 | - "if" statements should always be followed by "then" with the action |
|---|
| 73 | given on the next line, rather than one-line if statements. |
|---|
| 74 | |
|---|
| 75 | - "end" should always be followed by a space, i.e. "end do" ,"end if" |
|---|
| 76 | rather than "enddo", "endif" |
|---|
| 77 | |
|---|
| 78 | - Lots of comments! |
|---|
