Afew months ago, as I neared completion of the GEDCOM loading module within RetroGen, I stopped developing while I pondered over a dilemna. My initial goal, somewhat noble, was to develop a genealogy program using old programming language and database support to support the entire breath of the GEDCOM 5.5 standard. As my coding neared to an end, with about 95% completion of this goal, I had been realising that many of these features outlined in the standard would hanrdly get used. The design outlined in the standard was beautiful in how it cover all possible aspects of data collection, but the use of the Berkeley indexed sequential files system in conjunction with COBOL, allowed for only one variable length field to be stored in a file.
This meant that a simple GEDCOM file would take alot of space on a disk, not alot by today’s standards, but alot compared to the original GEDCOM file size itself. The limitation was the fixed lengths of data were not space compressed and it seemed to be alot of space being alloted to fields in the standard which would not be utilised.
I pondered on this for awhile. The development of the implementing the standard was near complete but simplification was required.
I decided to focus on implementing GEDCOM files sourced from Ancestry, and this enabled a vast reduction in the file schemas and also, enabled me to adequately use the last field for a variable length file. The removal of the GEDCOM tags no longer supported and recoding of variable length has reduced the size of the database considerably.
For me, the vision is clearer as I commence the coding of editing data.