Re: [Gramps-users] tools to compare two Family Trees

[Enno B. <enn...@gm...>]

Hello Deborah,

> I found gedcom diff here: https://github.com/elliotchance/gedcom and the releases (built binaries, no packages) for mac, windows and linux here: https://github.com/elliotchance/gedcom/releases.

I tried it, but am not very fond of it. Reason is that when I let it compare GEDCOM files exported from Gramps with a little over a month between them, it does not give me a proper view of the changes that I made in that month, which is what I expect a diff to do. Loading the files into a program like WinMerge works way better for this.

For a proper compare, I expect that, especially when files come from the same program, the program can quickly find out which persons were removed, which ones were added, and for which ones details were changed, and that's not what this program do.

When you compare trees like these, you know that persons have the same ID, _UID, or _FSFTID values, and the 1st run a diff program can do is to check which of these have been remover or added. Then, for all entries with identical IDs, the program can check for changes in person details, and sources, and the same can be done for families.

When GEDCOMs come from completely different sources, they will not have common __UIDs, and common IDs have no value, but in that case, you often know that they have groups of 'identical' persons in them, meaning that if you compare your tree with one independently made by a cousin, they will both have your common ancestors in them, and the ancestors' ancestors, etc. etc., and a smart program should be able to find such clusters of persons, because of the close similarity in these persons vitals, and the vitals of their relatives. And then, the program should also be able to quickly find out which ones have connections that their counterpart in the other tree don't have and vice versa, and show you that some common person has an ancestor that you didn't know about, or a child, a spouse, or a sibling, etc. etc.

For an expert, this is like walking both trees, starting from a person with high similarity, and comparing the paths, and that's something that I haven't really found in a GEDCOM diff, even though the principle itself is very simple, and taught in every programming class.

So much for now.

Regards,

Enno