My original thinking was to make a general routine that would handle all extended GEDCOM structures of any form. To ensure I could do this, I had to take an inventory of what structures were being used that I currently wasn’t handling.
I looked through the 54 test GEDCOM files that I have as produced from 23 different genealogy programs. I was surprised by what I found!
Only 17 used SOUR (sources), 9 used REPO (repositories), 8 used NOTE (note references), 4 used SCHEMA (to define new GEDCOM tags), 1 had a date from another calendar, and 6 used cross reference links on tags such as ALIA (alias), ASSO (associated), _WITN (witness), etc. Also, only 5 of the 54 had OBJE tags that referred to JPEG pictures.
So it seems that after I handle the NOTE tag and then do the sources and the repositories, there’s not that much left that needs to be handled. Seems like I can just forge ahead.