This application analyses a sequence of Fortran 77 source code files, divides their contents into program units, and produces statistics about the number and distribution of code and comment lines in each unit. These statistics are compared with typical values expected for well-crafted Fortran 77 code and obvious deviations from the expected values are flagged.

It is intended as a quick-look tool for assessing the size and likely quality of a Fortran 77 software system and is based on the observation that adequate commenting is one of the first things to be sacrificed if corners are cut during code development. A relatively simple analysis of comment and code lines can therefore reveal if this has occurred and can highlight potential trouble spots for visual inspection.

IN() = LITERAL (Read)

KEY = _LOGICAL (Read)

OUT = FILE (Write)

FORSTATS $\ast$.FOR

Analyses the source code held in the files $\ast$.FOR and writes a summary of the derived statistics to the default output file FORSTATS.LIS. This file may then be printed.

FORSTATS SOFTWARE.FOR SOFTWARE.LIS NOKEY

Analyses the source code held in the file SOFTWARE.FOR and writes the derived statistics to the file SOFTWARE.LIS. No explanatory key is written to the output file on this or subsequent invocations of FORSTATS until a TRUE value is given for the KEY parameter on the command line.

FORSTATS IN=["A$\ast$.FOR","B$\ast$.FOR"] OUT=ANALYSIS.LIS

In this example, a sequence of input file specifications is given and each will be analysed in turn. The combined statistics will be written to the file ANALYSIS.LIS.

This is a general-purpose tool and may be used on any Fortran 77 software system, since it does not depend on any particular layout or format. However, the analysis performed does include a test for the existence of prologue information. For this purpose, prologues are taken to be delimited by ’$+$’ or ’-’ characters appearing in the second column of a comment line. Note that unlike some other SST applications, both these characters are considered as equivalent by FORSTATS; i.e. the first occurrence begins a prologue and the second occurrence ends it, regardless of which character is used.

The execution time is approximately proportional to the total number of source code lines supplied for analysis. The time will be increased somewhat if the code resides in a large number of separate files, due to the need to open and close each file.