As I mentioned above, it is possible for any of these languages to process a command line argument like "file name".
YourFORTRAN -f name for example.

But hearing about how much this file format will be "shuffled around", I even more strongly recommend the "filter" approach.

Update:

Another way to think about your input file, is that is a database. Well, what kind of operations can we do? 1) Transform it in some way by additions, deletions, or changes to existing records. 2) Make reports from it based upon some subset of the data, perhaps using some calculations from the data(or subset) and/or sorting of the data. All of these things can be done with the filter model.

Let's go through some simple tasks:
-generate some kind of report A of the whole dataset - ok, make a genreport filter. cat data | genreport.pl
-generate the same report but only between dates X -Y, - ok, filter out data not between those dates with a date filter and pipe result to the genreport filter.
cat data | date.pl -s startdate -e enddate | genreport.pl
this date gizmo might just be a fancy grep command..
-add records, just use cat (or type on Windows), cat new >>data
-delete records between dates, well that is just inverse of the date filter
-change in some way, use transform filter, I don't know what this would do but I hope you are getting the idea..

We are beyond Perl here, but the way to set up something like I think you need is to have a bunch of little "filter" programs that can be combined and used in different orders. These programs are "stitched together" via a shell script, not with Perl. Each of these programs takes to the extent possible input from STDIN and output goes to STDOUT.