It took some time but I now have a fair idea on how the output program should look like:
both existing of a number of routines, matching the ‘members’ of that code. For the header, that means creation of lines for the <meta>, <link>, <script> and other tags in there; the body will contain modules for the header, left, middle and right panels, and the footer. First, these may be pretty hard- coded to see if the constructed output is what I expect it to be (in general, there will definitely parts missing that currently show up)may well be extracted form a text library and adapted to need, similar to the reporting routines that are used in the current WGPAGE program. I will use the same basic structure, I guess, at least the library- and sort routines I had created two decades ago, even the basics of the print routines in that code set – or create new ones.
Since these routines – written win Fortran77 but built using Fortran90 – use VMS’s Virtual memory access methods in their kernel routines, I will use Fortran to create the program; the existing program as a base.
But when looking into this older (Fortran) code, I found I had started already – in 2008 – in a program to work on blogs, and in general, it matches current idea, most is still very basic, in some cases just the routine. Nor does the code match the current naming convention I used for the RMS IO routines. However, it does contain a number of requirements I may have to decide upon: Getting the parameters to locate the files and the action to be performed. So that might be a far better solution: the code also connects for library access and building the output, as used in the current WGPAGE program that creates the current homepage.
There is still a lot to work upon, but after some changes, most of the code compiles.